Merge pull request #7475 from open-webui/dev

0.4.7

.dockerignore

@@ -16,4 +16,5 @@ _old
 uploads
 .ipynb_checkpoints
 **/*.db
-_test
+_test
+backend/data/*

.github/ISSUE_TEMPLATE/bug_report.md

@@ -8,6 +8,20 @@ assignees: ''
 
 # Bug Report
 
+## Important Notes
+
+- **Before submitting a bug report**: Please check the Issues or Discussions section to see if a similar issue or feature request has already been posted. It's likely we're already tracking it! If you’re unsure, start a discussion post first. This will help us efficiently focus on improving the project.
+
+- **Collaborate respectfully**: We value a constructive attitude, so please be mindful of your communication. If negativity is part of your approach, our capacity to engage may be limited. We’re here to help if you’re open to learning and communicating positively. Remember, Open WebUI is a volunteer-driven project managed by a single maintainer and supported by contributors who also have full-time jobs. We appreciate your time and ask that you respect ours.
+
+- **Contributing**: If you encounter an issue, we highly encourage you to submit a pull request or fork the project. We actively work to prevent contributor burnout to maintain the quality and continuity of Open WebUI.
+
+- **Bug reproducibility**: If a bug cannot be reproduced with a `:main` or `:dev` Docker setup, or a pip install with Python 3.11, it may require additional help from the community. In such cases, we will move it to the "issues" Discussions section due to our limited resources. We encourage the community to assist with these issues. Remember, it’s not that the issue doesn’t exist; we need your help!
+
+Note: Please remove the notes above when submitting your post. Thank you for your understanding and support!
+
+---
+
 ## Installation Method
 
 [Describe the method you used to install the project, e.g., git clone, Docker, pip, etc.]

.github/ISSUE_TEMPLATE/feature_request.md

@@ -6,6 +6,22 @@ labels: ''
 assignees: ''
 ---
 
+# Feature Request
+
+## Important Notes
+
+- **Before submitting a report**: Please check the Issues or Discussions section to see if a similar issue or feature request has already been posted. It's likely we're already tracking it! If you’re unsure, start a discussion post first. This will help us efficiently focus on improving the project.
+
+- **Collaborate respectfully**: We value a constructive attitude, so please be mindful of your communication. If negativity is part of your approach, our capacity to engage may be limited. We’re here to help if you’re open to learning and communicating positively. Remember, Open WebUI is a volunteer-driven project managed by a single maintainer and supported by contributors who also have full-time jobs. We appreciate your time and ask that you respect ours.
+
+- **Contributing**: If you encounter an issue, we highly encourage you to submit a pull request or fork the project. We actively work to prevent contributor burnout to maintain the quality and continuity of Open WebUI.
+
+- **Bug reproducibility**: If a bug cannot be reproduced with a `:main` or `:dev` Docker setup, or a pip install with Python 3.11, it may require additional help from the community. In such cases, we will move it to the "issues" Discussions section due to our limited resources. We encourage the community to assist with these issues. Remember, it’s not that the issue doesn’t exist; we need your help!
+
+Note: Please remove the notes above when submitting your post. Thank you for your understanding and support!
+
+---
+
 **Is your feature request related to a problem? Please describe.**
 A clear and concise description of what the problem is. Ex. I'm always frustrated when [...]
 

.github/dependabot.yml

@@ -3,10 +3,10 @@ updates:
   - package-ecosystem: pip
     directory: '/backend'
     schedule:
-      interval: weekly
+      interval: monthly
     target-branch: 'dev'
   - package-ecosystem: 'github-actions'
     directory: '/'
     schedule:
       # Check for updates to GitHub Actions every week
-      interval: 'weekly'
+      interval: monthly

.github/workflows/build-release.yml

@@ -10,61 +10,63 @@ jobs:
     runs-on: ubuntu-latest
 
     steps:
-    - name: Checkout repository
-      uses: actions/checkout@v4
+      - name: Checkout repository
+        uses: actions/checkout@v4
 
-    - name: Check for changes in package.json
-      run: |
-        git diff --cached --diff-filter=d package.json || {
-          echo "No changes to package.json"
-          exit 1
-        }
-    
-    - name: Get version number from package.json
-      id: get_version
-      run: |
-        VERSION=$(jq -r '.version' package.json)
-        echo "::set-output name=version::$VERSION"
+      - name: Check for changes in package.json
+        run: |
+          git diff --cached --diff-filter=d package.json || {
+            echo "No changes to package.json"
+            exit 1
+          }
 
-    - name: Extract latest CHANGELOG entry
-      id: changelog
-      run: |
-        CHANGELOG_CONTENT=$(awk 'BEGIN {print_section=0;} /^## \[/ {if (print_section == 0) {print_section=1;} else {exit;}} print_section {print;}' CHANGELOG.md)
-        CHANGELOG_ESCAPED=$(echo "$CHANGELOG_CONTENT" | sed ':a;N;$!ba;s/\n/%0A/g')
-        echo "Extracted latest release notes from CHANGELOG.md:" 
-        echo -e "$CHANGELOG_CONTENT" 
-        echo "::set-output name=content::$CHANGELOG_ESCAPED"
+      - name: Get version number from package.json
+        id: get_version
+        run: |
+          VERSION=$(jq -r '.version' package.json)
+          echo "::set-output name=version::$VERSION"
 
-    - name: Create GitHub release
-      uses: actions/github-script@v7
-      with:
-        github-token: ${{ secrets.GITHUB_TOKEN }}
-        script: |
-          const changelog = `${{ steps.changelog.outputs.content }}`;
-          const release = await github.rest.repos.createRelease({
-            owner: context.repo.owner,
-            repo: context.repo.repo,
-            tag_name: `v${{ steps.get_version.outputs.version }}`,
-            name: `v${{ steps.get_version.outputs.version }}`,
-            body: changelog,
-          })
-          console.log(`Created release ${release.data.html_url}`)
+      - name: Extract latest CHANGELOG entry
+        id: changelog
+        run: |
+          CHANGELOG_CONTENT=$(awk 'BEGIN {print_section=0;} /^## \[/ {if (print_section == 0) {print_section=1;} else {exit;}} print_section {print;}' CHANGELOG.md)
+          CHANGELOG_ESCAPED=$(echo "$CHANGELOG_CONTENT" | sed ':a;N;$!ba;s/\n/%0A/g')
+          echo "Extracted latest release notes from CHANGELOG.md:" 
+          echo -e "$CHANGELOG_CONTENT" 
+          echo "::set-output name=content::$CHANGELOG_ESCAPED"
 
-    - name: Upload package to GitHub release
-      uses: actions/upload-artifact@v4
-      with:
-        name: package
-        path: .
-      env:
-        GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+      - name: Create GitHub release
+        uses: actions/github-script@v7
+        with:
+          github-token: ${{ secrets.GITHUB_TOKEN }}
+          script: |
+            const changelog = `${{ steps.changelog.outputs.content }}`;
+            const release = await github.rest.repos.createRelease({
+              owner: context.repo.owner,
+              repo: context.repo.repo,
+              tag_name: `v${{ steps.get_version.outputs.version }}`,
+              name: `v${{ steps.get_version.outputs.version }}`,
+              body: changelog,
+            })
+            console.log(`Created release ${release.data.html_url}`)
 
-    - name: Trigger Docker build workflow
-      uses: actions/github-script@v7
-      with:
-        script: |
-          github.rest.actions.createWorkflowDispatch({
-            owner: context.repo.owner,
-            repo: context.repo.repo,
-            workflow_id: 'docker-build.yaml',
-            ref: 'v${{ steps.get_version.outputs.version }}',
-          })
+      - name: Upload package to GitHub release
+        uses: actions/upload-artifact@v4
+        with:
+          name: package
+          path: |
+            .
+            !.git
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Trigger Docker build workflow
+        uses: actions/github-script@v7
+        with:
+          script: |
+            github.rest.actions.createWorkflowDispatch({
+              owner: context.repo.owner,
+              repo: context.repo.repo,
+              workflow_id: 'docker-build.yaml',
+              ref: 'v${{ steps.get_version.outputs.version }}',
+            })

.github/workflows/deploy-to-hf-spaces.yml

@@ -28,6 +28,8 @@ jobs:
     steps:
       - name: Checkout repository
         uses: actions/checkout@v4
+        with:
+          lfs: true
 
       - name: Remove git history
         run: rm -rf .git
@@ -52,7 +54,9 @@ jobs:
       - name: Set up Git and push to Space
         run: |
           git init --initial-branch=main
+          git lfs install
           git lfs track "*.ttf"
+          git lfs track "*.jpg"
           rm demo.gif
           git add .
           git commit -m "GitHub deploy: ${{ github.sha }}"

.github/workflows/docker-build.yaml

@@ -312,7 +312,7 @@ jobs:
 
   merge-main-images:
     runs-on: ubuntu-latest
-    needs: [ build-main-image ]
+    needs: [build-main-image]
     steps:
       # GitHub Packages requires the entire repository name to be in lowercase
       # although the repository owner has a lowercase username, this prevents some people from running actions after forking
@@ -364,10 +364,9 @@ jobs:
         run: |
           docker buildx imagetools inspect ${{ env.FULL_IMAGE_NAME }}:${{ steps.meta.outputs.version }}
 
-
   merge-cuda-images:
     runs-on: ubuntu-latest
-    needs: [ build-cuda-image ]
+    needs: [build-cuda-image]
     steps:
       # GitHub Packages requires the entire repository name to be in lowercase
       # although the repository owner has a lowercase username, this prevents some people from running actions after forking
@@ -423,7 +422,7 @@ jobs:
 
   merge-ollama-images:
     runs-on: ubuntu-latest
-    needs: [ build-ollama-image ]
+    needs: [build-ollama-image]
     steps:
       # GitHub Packages requires the entire repository name to be in lowercase
       # although the repository owner has a lowercase username, this prevents some people from running actions after forking

.github/workflows/format-backend.yaml

@@ -23,7 +23,7 @@ jobs:
       - uses: actions/checkout@v4
 
       - name: Set up Python
-        uses: actions/setup-python@v4
+        uses: actions/setup-python@v5
         with:
           python-version: ${{ matrix.python-version }}
 

.github/workflows/format-build-frontend.yaml

@@ -21,7 +21,7 @@ jobs:
       - name: Setup Node.js
         uses: actions/setup-node@v4
         with:
-          node-version: '20' # Or specify any other version you want to use
+          node-version: '22' # Or specify any other version you want to use
 
       - name: Install Dependencies
         run: npm install
@@ -48,7 +48,7 @@ jobs:
       - name: Setup Node.js
         uses: actions/setup-node@v4
         with:
-          node-version: '20'
+          node-version: '22'
 
       - name: Install Dependencies
         run: npm ci

.github/workflows/integration-test.yml

@@ -15,6 +15,13 @@ jobs:
     name: Run Cypress Integration Tests
     runs-on: ubuntu-latest
     steps:
+      - name: Maximize build space
+        uses: AdityaGarg8/remove-unwanted-software@v4.1
+        with:
+          remove-android: 'true'
+          remove-haskell: 'true'
+          remove-codeql: 'true'
+
       - name: Checkout Repository
         uses: actions/checkout@v4
 
@@ -78,7 +85,7 @@ jobs:
   #     - uses: actions/checkout@v4
 
   #     - name: Set up Python
-  #       uses: actions/setup-python@v4
+  #       uses: actions/setup-python@v5
   #       with:
   #         python-version: ${{ matrix.python-version }}
 
@@ -150,7 +157,7 @@ jobs:
           GLOBAL_LOG_LEVEL: debug
         run: |
           cd backend
-          uvicorn main:app --port "8080" --forwarded-allow-ips '*' &
+          uvicorn open_webui.main:app --port "8080" --forwarded-allow-ips '*' &
           UVICORN_PID=$!
           # Wait up to 40 seconds for the server to start
           for i in {1..40}; do
@@ -175,9 +182,12 @@ jobs:
           WEBUI_SECRET_KEY: secret-key
           GLOBAL_LOG_LEVEL: debug
           DATABASE_URL: postgresql://postgres:postgres@localhost:5432/postgres
+          DATABASE_POOL_SIZE: 10
+          DATABASE_POOL_MAX_OVERFLOW: 10
+          DATABASE_POOL_TIMEOUT: 30
         run: |
           cd backend
-          uvicorn main:app --port "8081" --forwarded-allow-ips '*' &
+          uvicorn open_webui.main:app --port "8081" --forwarded-allow-ips '*' &
           UVICORN_PID=$!
           # Wait up to 20 seconds for the server to start
           for i in {1..20}; do
@@ -223,7 +233,7 @@ jobs:
 #          DATABASE_URL: mysql://root:mysql@localhost:3306/mysql
 #        run: |
 #          cd backend
-#          uvicorn main:app --port "8083" --forwarded-allow-ips '*' &
+#          uvicorn open_webui.main:app --port "8083" --forwarded-allow-ips '*' &
 #          UVICORN_PID=$!
 #          # Wait up to 20 seconds for the server to start
 #          for i in {1..20}; do

.github/workflows/lint-backend.disabled

@@ -16,7 +16,7 @@ jobs:
     steps:
       - uses: actions/checkout@v4
       - name: Use Python
-        uses: actions/setup-python@v4
+        uses: actions/setup-python@v5
       - name: Use Bun
         uses: oven-sh/setup-bun@v1
       - name: Install dependencies

.github/workflows/release-pypi.yml

@@ -4,6 +4,7 @@ on:
   push:
     branches:
       - main # or whatever branch you want to use
+      - pypi-release
 
 jobs:
   release:

CHANGELOG.md

@@ -5,6 +5,584 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [0.4.7] - 2024-12-01
+
+### Added
+
+- **✨ Prompt Input Auto-Completion**: Type a prompt and let AI intelligently suggest and complete your inputs. Simply press 'Tab' or swipe right on mobile to confirm. Available only with Rich Text Input (default setting). Disable via Admin Settings for full control.
+- **🌍 Improved Translations**: Enhanced localization for multiple languages, ensuring a more polished and accessible experience for international users.
+
+### Fixed
+
+- **🛠️ Tools Export Issue**: Resolved a critical issue where exporting tools wasn’t functioning, restoring seamless export capabilities.
+- **🔗 Model ID Registration**: Fixed an issue where model IDs weren’t registering correctly in the model editor, ensuring reliable model setup and tracking.
+- **🖋️ Textarea Auto-Expansion**: Corrected a bug where textareas didn’t expand automatically on certain browsers, improving usability for multi-line inputs.
+- **🔧 Ollama Embed Endpoint**: Addressed the /ollama/embed endpoint malfunction, ensuring consistent performance and functionality.
+
+### Changed
+
+- **🎨 Knowledge Base Styling**: Refined knowledge base visuals for a cleaner, more modern look, laying the groundwork for further enhancements in upcoming releases.
+
+## [0.4.6] - 2024-11-26
+
+### Added
+
+- **🌍 Enhanced Translations**: Various language translations improved to make the WebUI more accessible and user-friendly worldwide.
+
+### Fixed
+
+- **✏️ Textarea Shifting Bug**: Resolved the issue where the textarea shifted unexpectedly, ensuring a smoother typing experience.
+- **⚙️ Model Configuration Modal**: Fixed the issue where the models configuration modal introduced in 0.4.5 wasn’t working for some users.
+- **🔍 Legacy Query Support**: Restored functionality for custom query generation in RAG when using legacy prompts, ensuring both default and custom templates now work seamlessly.
+- **⚡ Improved General Reliability**: Various minor fixes improve platform stability and ensure a smoother overall experience across workflows.
+
+## [0.4.5] - 2024-11-26
+
+### Added
+
+- **🎨 Model Order/Defaults Reintroduced**: Brought back the ability to set model order and default models, now configurable via Admin Settings > Models > Configure (Gear Icon).
+
+### Fixed
+
+- **🔍 Query Generation Issue**: Resolved an error in web search query generation, enhancing search accuracy and ensuring smoother search workflows.
+- **📏 Textarea Auto Height Bug**: Fixed a layout issue where textarea input height was shifting unpredictably, particularly when editing system prompts.
+- **🔑 Ollama Authentication**: Corrected an issue with Ollama’s authorization headers, guaranteeing reliable authentication across all endpoints.
+- **⚙️ Missing Min_P Save**: Resolved an issue where the 'min_p' parameter was not being saved in configurations.
+- **🛠️ Tools Description**: Fixed a key issue that omitted tool descriptions in tools payload.
+
+## [0.4.4] - 2024-11-22
+
+### Added
+
+- **🌐 Translation Updates**: Refreshed Catalan, Brazilian Portuguese, German, and Ukrainian translations, further enhancing the platform's accessibility and improving the experience for international users.
+
+### Fixed
+
+- **📱 Mobile Controls Visibility**: Resolved an issue where the controls button was not displaying on the new chats page for mobile users, ensuring smoother navigation and functionality on smaller screens.
+- **📷 LDAP Profile Image Issue**: Fixed an LDAP integration bug related to profile images, ensuring seamless authentication and a reliable login experience for users.
+- **⏳ RAG Query Generation Issue**: Addressed a significant problem where RAG query generation occurred unnecessarily without attached files, drastically improving speed and reducing delays during chat completions.
+
+### Changed
+
+- **⚙️ Legacy Event Emitter Support**: Reintroduced compatibility with legacy "citation" types for event emitters in tools and functions, providing smoother workflows and broader tool support for users.
+
+## [0.4.3] - 2024-11-21
+
+### Added
+
+- **📚 Inline Citations for RAG Results**: Get seamless inline citations for Retrieval-Augmented Generation (RAG) responses using the default RAG prompt. Note: This feature only supports newly uploaded files, improving traceability and providing source clarity.
+- **🎨 Better Rich Text Input Support**: Enjoy smoother and more reliable rich text formatting for chats, enhancing communication quality.
+- **⚡ Faster Model Retrieval**: Implemented caching optimizations for faster model loading, providing a noticeable speed boost across workflows. Further improvements are on the way!
+
+### Fixed
+
+- **🔗 Pipelines Feature Restored**: Resolved a critical issue that previously prevented Pipelines from functioning, ensuring seamless workflows.
+- **✏️ Missing Suffix Field in Ollama Form**: Added the missing "suffix" field to the Ollama generate form, enhancing customization options.
+
+### Changed
+
+- **🗂️ Renamed "Citations" to "Sources"**: Improved clarity and consistency by renaming the "citations" field to "sources" in messages.
+
+## [0.4.2] - 2024-11-20
+
+### Fixed
+
+- **📁 Knowledge Files Visibility Issue**: Resolved the bug preventing individual files in knowledge collections from displaying when referenced with '#'.
+- **🔗 OpenAI Endpoint Prefix**: Fixed the issue where certain OpenAI connections that deviate from the official API spec weren’t working correctly with prefixes.
+- **⚔️ Arena Model Access Control**: Corrected an issue where arena model access control settings were not being saved.
+- **🔧 Usage Capability Selector**: Fixed the broken usage capabilities selector in the model editor.
+
+## [0.4.1] - 2024-11-19
+
+### Added
+
+- **📊 Enhanced Feedback System**: Introduced a detailed 1-10 rating scale for feedback alongside thumbs up/down, preparing for more precise model fine-tuning and improving feedback quality.
+- **ℹ️ Tool Descriptions on Hover**: Easily access tool descriptions by hovering over the message input, providing a smoother workflow with more context when utilizing tools.
+
+### Fixed
+
+- **🗑️ Graceful Handling of Deleted Users**: Resolved an issue where deleted users caused workspace items (models, knowledge, prompts, tools) to fail, ensuring reliable workspace loading.
+- **🔑 API Key Creation**: Fixed an issue preventing users from creating new API keys, restoring secure and seamless API management.
+- **🔗 HTTPS Proxy Fix**: Corrected HTTPS proxy issues affecting the '/api/v1/models/' endpoint, ensuring smoother, uninterrupted model management.
+
+## [0.4.0] - 2024-11-19
+
+### Added
+
+- **👥 User Groups**: You can now create and manage user groups, making user organization seamless.
+- **🔐 Group-Based Access Control**: Set granular access to models, knowledge, prompts, and tools based on user groups, allowing for more controlled and secure environments.
+- **🛠️ Group-Based User Permissions**: Easily manage workspace permissions. Grant users the ability to upload files, delete, edit, or create temporary chats, as well as define their ability to create models, knowledge, prompts, and tools.
+- **🔑 LDAP Support**: Newly introduced LDAP authentication adds robust security and scalability to user management.
+- **🌐 Enhanced OpenAI-Compatible Connections**: Added prefix ID support to avoid model ID clashes, with explicit model ID support for APIs lacking '/models' endpoint support, ensuring smooth operation with custom setups.
+- **🔐 Ollama API Key Support**: Now manage credentials for Ollama when set behind proxies, including the option to utilize prefix ID for proper distinction across multiple Ollama instances.
+- **🔄 Connection Enable/Disable Toggle**: Easily enable or disable individual OpenAI and Ollama connections as needed.
+- **🎨 Redesigned Model Workspace**: Freshly redesigned to improve usability for managing models across users and groups.
+- **🎨 Redesigned Prompt Workspace**: A fresh UI to conveniently organize and manage prompts.
+- **🧩 Sorted Functions Workspace**: Functions are now automatically categorized by type (Action, Filter, Pipe), streamlining management.
+- **💻 Redesigned Collaborative Workspace**: Enhanced support for multiple users contributing to models, knowledge, prompts, or tools, improving collaboration.
+- **🔧 Auto-Selected Tools in Model Editor**: Tools enabled through the model editor are now automatically selected, whereas previously it only gave users the option to enable the tool, reducing manual steps and enhancing efficiency.
+- **🔔 Web Search & Tools Indicator**: A clear indication now shows when web search or tools are active, reducing confusion.
+- **🔑 Toggle API Key Auth**: Tighten security by easily enabling or disabling API key authentication option for Open WebUI.
+- **🗂️ Agentic Retrieval**: Improve RAG accuracy via smart pre-processing of chat history to determine the best queries before retrieval.
+- **📁 Large Text as File Option**: Optionally convert large pasted text into a file upload, keeping the chat interface cleaner.
+- **🗂️ Toggle Citations for Models**: Ability to disable citations has been introduced in the model editor.
+- **🔍 User Settings Search**: Quickly search for settings fields, improving ease of use and navigation.
+- **🗣️ Experimental SpeechT5 TTS**: Local SpeechT5 support added for improved text-to-speech capabilities.
+- **🔄 Unified Reset for Models**: A one-click option has been introduced to reset and remove all models from the Admin Settings.
+- **🛠️ Initial Setup Wizard**: The setup process now explicitly informs users that they are creating an admin account during the first-time setup, ensuring clarity. Previously, users encountered the login page right away without this distinction.
+- **🌐 Enhanced Translations**: Several language translations, including Ukrainian, Norwegian, and Brazilian Portuguese, were refined for better localization.
+
+### Fixed
+
+- **🎥 YouTube Video Attachments**: Fixed issues preventing proper loading and attachment of YouTube videos as files.
+- **🔄 Shared Chat Update**: Corrected issues where shared chats were not updating, improving collaboration consistency.
+- **🔍 DuckDuckGo Rate Limit Fix**: Addressed issues with DuckDuckGo search integration, enhancing search stability and performance when operating within rate limits.
+- **🧾 Citations Relevance Fix**: Adjusted the relevance percentage calculation for citations, so that Open WebUI properly reflect the accuracy of a retrieved document in RAG, ensuring users get clearer insights into sources.
+- **🔑 Jina Search API Key Requirement**: Added the option to input an API key for Jina Search, ensuring smooth functionality as keys are now mandatory.
+
+### Changed
+
+- **🛠️ Functions Moved to Admin Panel**: As Functions operate as advanced plugins, they are now accessible from the Admin Panel instead of the workspace.
+- **🛠️ Manage Ollama Connections**: The "Models" section in Admin Settings has been relocated to Admin Settings > "Connections" > Ollama Connections. You can now manage Ollama instances via a dedicated "Manage Ollama" modal from "Connections", streamlining the setup and configuration of Ollama models.
+- **📊 Base Models in Admin Settings**: Admins can now find all base models, both connections or functions, in the "Models" Admin setting. Global model accessibility can be enabled or disabled here. Models are private by default, requiring explicit permission assignment for user access.
+- **📌 Sticky Model Selection for New Chats**: The model chosen from a previous chat now persists when creating a new chat. If you click "New Chat" again from the new chat page, it will revert to your default model.
+- **🎨 Design Refactoring**: Overall design refinements across the platform have been made, providing a more cohesive and polished user experience.
+
+### Removed
+
+- **📂 Model List Reordering**: Temporarily removed and will be reintroduced in upcoming user group settings improvements.
+- **⚙️ Default Model Setting**: Removed the ability to set a default model for users, will be reintroduced with user group settings in the future.
+
+## [0.3.35] - 2024-10-26
+
+### Added
+
+- **🌐 Translation Update**: Added translation labels in the SearchInput and CreateCollection components and updated Brazilian Portuguese translation (pt-BR)
+- **📁 Robust File Handling**: Enhanced file input handling for chat. If the content extraction fails or is empty, users will now receive a clear warning, preventing silent failures and ensuring you always know what's happening with your uploads.
+- **🌍 New Language Support**: Introduced Hungarian translations and updated French translations, expanding the platform's language accessibility for a more global user base.
+
+### Fixed
+
+- **📚 Knowledge Base Loading Issue**: Resolved a critical bug where the Knowledge Base was not loading, ensuring smooth access to your stored documents and improving information retrieval in RAG-enhanced workflows.
+- **🛠️ Tool Parameters Issue**: Fixed an error where tools were not functioning correctly when required parameters were missing, ensuring reliable tool performance and more efficient task completions.
+- **🔗 Merged Response Loss in Multi-Model Chats**: Addressed an issue where responses in multi-model chat workflows were being deleted after follow-up queries, improving consistency and ensuring smoother interactions across models.
+
+## [0.3.34] - 2024-10-26
+
+### Added
+
+- **🔧 Feedback Export Enhancements**: Feedback history data can now be exported to JSON, allowing for seamless integration in RLHF processing and further analysis.
+- **🗂️ Embedding Model Lazy Loading**: Search functionality for leaderboard reranking is now more efficient, as embedding models are lazy-loaded only when needed, optimizing performance.
+- **🎨 Rich Text Input Toggle**: Users can now switch back to legacy textarea input for chat if they prefer simpler text input, though rich text is still the default until deprecation.
+- **🛠️ Improved Tool Calling Mechanism**: Enhanced method for parsing and calling tools, improving the reliability and robustness of tool function calls.
+- **🌐 Globalization Enhancements**: Updates to internationalization (i18n) support, further refining multi-language compatibility and accuracy.
+
+### Fixed
+
+- **🖥️ Folder Rename Fix for Firefox**: Addressed a persistent issue where users could not rename folders by pressing enter in Firefox, now ensuring seamless folder management across browsers.
+- **🔠 Tiktoken Model Text Splitter Issue**: Resolved an issue where the tiktoken text splitter wasn’t working in Docker installations, restoring full functionality for tokenized text editing.
+- **💼 S3 File Upload Issue**: Fixed a problem affecting S3 file uploads, ensuring smooth operations for those who store files on cloud storage.
+- **🔒 Strict-Transport-Security Crash**: Resolved a crash when setting the Strict-Transport-Security (HSTS) header, improving stability and security enhancements.
+- **🚫 OIDC Boolean Access Fix**: Addressed an issue with boolean values not being accessed correctly during OIDC logins, ensuring login reliability.
+- **⚙️ Rich Text Paste Behavior**: Refined paste behavior in rich text input to make it smoother and more intuitive when pasting various content types.
+- **🔨 Model Exclusion for Arena Fix**: Corrected the filter function that was not properly excluding models from the arena, improving model management.
+- **🏷️ "Tags Generation Prompt" Fix**: Addressed an issue preventing custom "tags generation prompts" from registering properly, ensuring custom prompt work seamlessly.
+
+## [0.3.33] - 2024-10-24
+
+### Added
+
+- **🏆 Evaluation Leaderboard**: Easily track your performance through a new leaderboard system where your ratings contribute to a real-time ranking based on the Elo system. Sibling responses (regenerations, many model chats) are required for your ratings to count in the leaderboard. Additionally, you can opt-in to share your feedback history and be part of the community-wide leaderboard. Expect further improvements as we refine the algorithm—help us build the best community leaderboard!
+- **⚔️ Arena Model Evaluation**: Enable blind A/B testing of models directly from Admin Settings > Evaluation for a true side-by-side comparison. Ideal for pinpointing the best model for your needs.
+- **🎯 Topic-Based Leaderboard**: Discover more accurate rankings with experimental topic-based reranking, which adjusts leaderboard standings based on tag similarity in feedback. Get more relevant insights based on specific topics!
+- **📁 Folders Support for Chats**: Organize your chats better by grouping them into folders. Drag and drop chats between folders and export them seamlessly for easy sharing or analysis.
+- **📤 Easy Chat Import via Drag & Drop**: Save time by simply dragging and dropping chat exports (JSON) directly onto the sidebar to import them into your workspace—streamlined, efficient, and intuitive!
+- **📚 Enhanced Knowledge Collection**: Now, you can reference individual files from a knowledge collection—ideal for more precise Retrieval-Augmented Generations (RAG) queries and document analysis.
+- **🏷️ Enhanced Tagging System**: Tags now take up less space! Utilize the new 'tag:' query system to manage, search, and organize your conversations more effectively without cluttering the interface.
+- **🧠 Auto-Tagging for Chats**: Your conversations are now automatically tagged for improved organization, mirroring the efficiency of auto-generated titles.
+- **🔍 Backend Chat Query System**: Chat filtering has become more efficient, now handled through the backend\*\* instead of your browser, improving search performance and accuracy.
+- **🎮 Revamped Playground**: Experience a refreshed and optimized Playground for smoother testing, tweaks, and experimentation of your models and tools.
+- **🧩 Token-Based Text Splitter**: Introducing token-based text splitting (tiktoken), giving you more precise control over how text is processed. Previously, only character-based splitting was available.
+- **🔢 Ollama Batch Embeddings**: Leverage new batch embedding support for improved efficiency and performance with Ollama embedding models.
+- **🔍 Enhanced Add Text Content Modal**: Enjoy a cleaner, more intuitive workflow for adding and curating knowledge content with an upgraded input modal from our Knowledge workspace.
+- **🖋️ Rich Text Input for Chats**: Make your chat inputs more dynamic with support for rich text formatting. Your conversations just got a lot more polished and professional.
+- **⚡ Faster Whisper Model Configurability**: Customize your local faster whisper model directly from the WebUI.
+- **☁️ Experimental S3 Support**: Enable stateless WebUI instances with S3 support, greatly enhancing scalability and balancing heavy workloads.
+- **🔕 Disable Update Toast**: Now you can streamline your workspace even further—choose to disable update notifications for a more focused experience.
+- **🌟 RAG Citation Relevance Percentage**: Easily assess citation accuracy with the addition of relevance percentages in RAG results.
+- **⚙️ Mermaid Copy Button**: Mermaid diagrams now come with a handy copy button, simplifying the extraction and use of diagram contents directly in your workflow.
+- **🎨 UI Redesign**: Major interface redesign that will make navigation smoother, keep your focus where it matters, and ensure a modern look.
+
+### Fixed
+
+- **🎙️ Voice Note Mic Stopping Issue**: Fixed the issue where the microphone stayed active after ending a voice note recording, ensuring your audio workflow runs smoothly.
+
+### Removed
+
+- **👋 Goodbye Sidebar Tags**: Sidebar tag clutter is gone. We’ve shifted tag buttons to more effective query-based tag filtering for a sleeker, more agile interface.
+
+## [0.3.32] - 2024-10-06
+
+### Added
+
+- **🔢 Workspace Enhancements**: Added a display count for models, prompts, tools, and functions in the workspace, providing a clear overview and easier management.
+
+### Fixed
+
+- **🖥️ Web and YouTube Attachment Fix**: Resolved an issue where attaching web links and YouTube videos was malfunctioning, ensuring seamless integration and display within chats.
+- **📞 Call Mode Activation on Landing Page**: Fixed a bug where call mode was not operational from the landing page.
+
+### Changed
+
+- **🔄 URL Parameter Refinement**: Updated the 'tool_ids' URL parameter to 'tools' or 'tool-ids' for more intuitive and consistent user experience.
+- **🎨 Floating Buttons Styling Update**: Refactored the styling of floating buttons to intelligently adjust to the left side when there isn't enough room on the right, improving interface usability and aesthetic.
+- **🔧 Enhanced Accessibility for Floating Buttons**: Implemented the ability to close floating buttons with the 'Esc' key, making workflow smoother and more efficient for users navigating via keyboard.
+- **🖇️ Updated Information URL**: Information URLs now direct users to a general release page rather than a version-specific URL, ensuring access to the latest and relevant details all in one place.
+- **📦 Library Dependencies Update**: Upgraded dependencies to ensure compatibility and performance optimization for pip installs.
+
+## [0.3.31] - 2024-10-06
+
+### Added
+
+- **📚 Knowledge Feature**: Reimagined documents feature, now more performant with a better UI for enhanced organization; includes streamlined API integration for Retrieval-Augmented Generation (RAG). Detailed documentation forthcoming: https://docs.openwebui.com/
+- **🌐 New Landing Page**: Freshly designed landing page; toggle between the new UI and the classic chat UI from Settings > Interface for a personalized experience.
+- **📁 Full Document Retrieval Mode**: Toggle between full document retrieval or traditional snippets by clicking on the file item. This mode enhances document capabilities and supports comprehensive tasks like summarization by utilizing the entire content instead of RAG.
+- **📄 Extracted File Content Display**: View extracted content directly by clicking on the file item, simplifying file analysis.
+- **🎨 Artifacts Feature**: Render web content and SVGs directly in the interface, supporting quick iterations and live changes.
+- **🖊️ Editable Code Blocks**: Supercharged code blocks now allow live editing directly in the LLM response, with live reloads supported by artifacts.
+- **🔧 Code Block Enhancements**: Introduced a floating copy button in code blocks to facilitate easier code copying without scrolling.
+- **🔍 SVG Pan/Zoom**: Enhanced interaction with SVG images, including Mermaid diagrams, via new pan and zoom capabilities.
+- **🔍 Text Select Quick Actions**: New floating buttons appear when text is highlighted in LLM responses, offering deeper interactions like "Ask a Question" or "Explain".
+- **🗃️ Database Pool Configuration**: Enhanced database handling to support scalable user growth.
+- **🔊 Experimental Audio Compression**: Compress audio files to navigate around the 25MB limit for OpenAI's speech-to-text processing.
+- **🔍 Query Embedding**: Adjusted embedding behavior to enhance system performance by not repeating query embedding.
+- **💾 Lazy Load Optimizations**: Implemented lazy loading of large dependencies to minimize initial memory usage, boosting performance.
+- **🍏 Apple Touch Icon Support**: Optimizes the display of icons for web bookmarks on Apple mobile devices.
+- **🔽 Expandable Content Markdown Support**: Introducing 'details', 'summary' tag support for creating expandable content sections in markdown, facilitating cleaner, organized documentation and interactive content display.
+
+### Fixed
+
+- **🔘 Action Button Issue**: Resolved a bug where action buttons were not functioning, enhancing UI reliability.
+- **🔄 Multi-Model Chat Loop**: Fixed an infinite loop issue in multi-model chat environments, ensuring smoother chat operations.
+- **📄 Chat PDF/TXT Export Issue**: Resolved problems with exporting chat logs to PDF and TXT formats.
+- **🔊 Call to Text-to-Speech Issues**: Rectified problems with text-to-speech functions to improve audio interactions.
+
+### Changed
+
+- **⚙️ Endpoint Renaming**: Renamed 'rag' endpoints to 'retrieval' for clearer function description.
+- **🎨 Styling and Interface Updates**: Multiple refinements across the platform to enhance visual appeal and user interaction.
+
+### Removed
+
+- **🗑️ Deprecated 'DOCS_DIR'**: Removed the outdated 'docs_dir' variable in favor of more direct file management solutions, with direct file directory syncing and API uploads for a more integrated experience.
+
+## [0.3.30] - 2024-09-26
+
+### Fixed
+
+- **🍞 Update Available Toast Dismissal**: Enhanced user experience by ensuring that once the update available notification is dismissed, it won't reappear for 24 hours.
+- **📋 Ollama /embed Form Data**: Adjusted the integration inaccuracies in the /embed form data to ensure it perfectly matches with Ollama's specifications.
+- **🔧 O1 Max Completion Tokens Issue**: Resolved compatibility issues with OpenAI's o1 models max_completion_tokens param to ensure smooth operation.
+- **🔄 Pip Install Database Issue**: Fixed a critical issue where database changes during pip installations were reverting and not saving chat logs, now ensuring data persistence and reliability in chat operations.
+- **🏷️ Chat Rename Tab Update**: Fixed the functionality to change the web browser's tab title simultaneously when a chat is renamed, keeping tab titles consistent.
+
+## [0.3.29] - 2023-09-25
+
+### Fixed
+
+- **🔧 KaTeX Rendering Improvement**: Resolved specific corner cases in KaTeX rendering to enhance the display of complex mathematical notation.
+- **📞 'Call' URL Parameter Fix**: Corrected functionality for 'call' URL search parameter ensuring reliable activation of voice calls through URL triggers.
+- **🔄 Configuration Reset Fix**: Fixed the RESET_CONFIG_ON_START to ensure settings revert to default correctly upon each startup, improving reliability in configuration management.
+- **🌍 Filter Outlet Hook Fix**: Addressed issues in the filter outlet hook, ensuring all filter functions operate as intended.
+
+## [0.3.28] - 2024-09-24
+
+### Fixed
+
+- **🔍 Web Search Functionality**: Corrected an issue where the web search option was not functioning properly.
+
+## [0.3.27] - 2024-09-24
+
+### Fixed
+
+- **🔄 Periodic Cleanup Error Resolved**: Fixed a critical RuntimeError related to the 'periodic_usage_pool_cleanup' coroutine, ensuring smooth and efficient performance post-pip install, correcting a persisting issue from version 0.3.26.
+- **📊 Enhanced LaTeX Rendering**: Improved rendering for LaTeX content, enhancing clarity and visual presentation in documents and mathematical models.
+
+## [0.3.26] - 2024-09-24
+
+### Fixed
+
+- **🔄 Event Loop Error Resolution**: Addressed a critical error where a missing running event loop caused 'periodic_usage_pool_cleanup' to fail with pip installs. This fix ensures smoother and more reliable updates and installations, enhancing overall system stability.
+
+## [0.3.25] - 2024-09-24
+
+### Fixed
+
+- **🖼️ Image Generation Functionality**: Resolved an issue where image generation was not functioning, restoring full capability for visual content creation.
+- **⚖️ Rate Response Corrections**: Addressed a problem where rate responses were not working, ensuring reliable feedback mechanisms are operational.
+
+## [0.3.24] - 2024-09-24
+
+### Added
+
+- **🚀 Rendering Optimization**: Significantly improved message rendering performance, enhancing user experience and webui responsiveness.
+- **💖 Favorite Response Feature in Chat Overview**: Users can now mark responses as favorite directly from the chat overview, enhancing ease of retrieval and organization of preferred responses.
+- **💬 Create Message Pairs with Shortcut**: Implemented creation of new message pairs using Cmd/Ctrl+Shift+Enter, making conversation editing faster and more intuitive.
+- **🌍 Expanded User Prompt Variables**: Added weekday, timezone, and language information variables to user prompts to match system prompt variables.
+- **🎵 Enhanced Audio Support**: Now includes support for 'audio/x-m4a' files, broadening compatibility with audio content within the platform.
+- **🔏 Model URL Search Parameter**: Added an ability to select a model directly via URL parameters, streamlining navigation and model access.
+- **📄 Enhanced PDF Citations**: PDF citations now open at the associated page, streamlining reference checks and document handling.
+- **🔧Use of Redis in Sockets**: Enhanced socket implementation to fully support Redis, enabling effective stateless instances suitable for scalable load balancing.
+- **🌍 Stream Individual Model Responses**: Allows specific models to have individualized streaming settings, enhancing performance and customization.
+- **🕒 Display Model Hash and Last Modified Timestamp for Ollama Models**: Provides critical model details directly in the Models workspace for enhanced tracking.
+- **❗ Update Info Notification for Admins**: Ensures administrators receive immediate updates upon login, keeping them informed of the latest changes and system statuses.
+
+### Fixed
+
+- **🗑️ Temporary File Handling On Windows**: Fixed an issue causing errors when accessing a temporary file being used by another process, Tools & Functions should now work as intended.
+- **🔓 Authentication Toggle Issue**: Resolved the malfunction where setting 'WEBUI_AUTH=False' did not appropriately disable authentication, ensuring that user experience and system security settings function as configured.
+- **🔧 Save As Copy Issue for Many Model Chats**: Resolved an error preventing users from save messages as copies in many model chats.
+- **🔒 Sidebar Closure on Mobile**: Resolved an issue where the mobile sidebar remained open after menu engagement, improving user interface responsivity and comfort.
+- **🛡️ Tooltip XSS Vulnerability**: Resolved a cross-site scripting (XSS) issue within tooltips, ensuring enhanced security and data integrity during user interactions.
+
+### Changed
+
+- **↩️ Deprecated Interface Stream Response Settings**: Moved to advanced parameters to streamline interface settings and enhance user clarity.
+- **⚙️ Renamed 'speedRate' to 'playbackRate'**: Standardizes terminology, improving usability and understanding in media settings.
+
+## [0.3.23] - 2024-09-21
+
+### Added
+
+- **🚀 WebSocket Redis Support**: Enhanced load balancing capabilities for multiple instance setups, promoting better performance and reliability in WebUI.
+- **🔧 Adjustable Chat Controls**: Introduced width-adjustable chat controls, enabling a personalized and more comfortable user interface.
+- **🌎 i18n Updates**: Improved and updated the Chinese translations.
+
+### Fixed
+
+- **🌐 Task Model Unloading Issue**: Modified task handling to use the Ollama /api/chat endpoint instead of OpenAI compatible endpoint, ensuring models stay loaded and ready with custom parameters, thus minimizing delays in task execution.
+- **📝 Title Generation Fix for OpenAI Compatible APIs**: Resolved an issue preventing the generation of titles, enhancing consistency and reliability when using multiple API providers.
+- **🗃️ RAG Duplicate Collection Issue**: Fixed a bug causing repeated processing of the same uploaded file. Now utilizes indexed files to prevent unnecessary duplications, optimizing resource usage.
+- **🖼️ Image Generation Enhancement**: Refactored OpenAI image generation endpoint to be asynchronous, preventing the WebUI from becoming unresponsive during processing, thus enhancing user experience.
+- **🔓 Downgrade Authlib**: Reverted Authlib to version 1.3.1 to address and resolve issues concerning OAuth functionality.
+
+### Changed
+
+- **🔍 Improved Message Interaction**: Enhanced the message node interface to allow for easier focus redirection with a simple click, streamlining user interaction.
+- **✨ Styling Refactor**: Updated WebUI styling for a cleaner, more modern look, enhancing user experience across the platform.
+
+## [0.3.22] - 2024-09-19
+
+### Added
+
+- **⭐ Chat Overview**: Introducing a node-based interactive messages diagram for improved visualization of conversation flows.
+- **🔗 Multiple Vector DB Support**: Now supports multiple vector databases, including the newly added Milvus support. Community contributions for additional database support are highly encouraged!
+- **📡 Experimental Non-Stream Chat Completion**: Experimental feature allowing the use of OpenAI o1 models, which do not support streaming, ensuring more versatile model deployment.
+- **🔍 Experimental Colbert-AI Reranker Integration**: Added support for "jinaai/jina-colbert-v2" as a reranker, enhancing search relevance and accuracy. Note: it may not function at all on low-spec computers.
+- **🕸️ ENABLE_WEBSOCKET_SUPPORT**: Added environment variable for instances to ignore websocket upgrades, stabilizing connections on platforms with websocket issues.
+- **🔊 Azure Speech Service Integration**: Added support for Azure Speech services for Text-to-Speech (TTS).
+- **🎚️ Customizable Playback Speed**: Playback speed control is now available in Call mode settings, allowing users to adjust audio playback speed to their preferences.
+- **🧠 Enhanced Error Messaging**: System now displays helpful error messages directly to users during chat completion issues.
+- **📂 Save Model as Transparent PNG**: Model profile images are now saved as PNGs, supporting transparency and improving visual integration.
+- **📱 iPhone Compatibility Adjustments**: Added padding to accommodate the iPhone navigation bar, improving UI display on these devices.
+- **🔗 Secure Response Headers**: Implemented security response headers, bolstering web application security.
+- **🔧 Enhanced AUTOMATIC1111 Settings**: Users can now configure 'CFG Scale', 'Sampler', and 'Scheduler' parameters directly in the admin settings, enhancing workflow flexibility without source code modifications.
+- **🌍 i18n Updates**: Enhanced translations for Chinese, Ukrainian, Russian, and French, fostering a better localized experience.
+
+### Fixed
+
+- **🛠️ Chat Message Deletion**: Resolved issues with chat message deletion, ensuring a smoother user interaction and system stability.
+- **🔢 Ordered List Numbering**: Fixed the incorrect ordering in lists.
+
+### Changed
+
+- **🎨 Transparent Icon Handling**: Allowed model icons to be displayed on transparent backgrounds, improving UI aesthetics.
+- **📝 Improved RAG Template**: Enhanced Retrieval-Augmented Generation template, optimizing context handling and error checking for more precise operation.
+
+## [0.3.21] - 2024-09-08
+
+### Added
+
+- **📊 Document Count Display**: Now displays the total number of documents directly within the dashboard.
+- **🚀 Ollama Embed API Endpoint**: Enabled /api/embed endpoint proxy support.
+
+### Fixed
+
+- **🐳 Docker Launch Issue**: Resolved the problem preventing Open-WebUI from launching correctly when using Docker.
+
+### Changed
+
+- **🔍 Enhanced Search Prompts**: Improved the search query generation prompts for better accuracy and user interaction, enhancing the overall search experience.
+
+## [0.3.20] - 2024-09-07
+
+### Added
+
+- **🌐 Translation Update**: Updated Catalan translations to improve user experience for Catalan speakers.
+
+### Fixed
+
+- **📄 PDF Download**: Resolved a configuration issue with fonts directory, ensuring PDFs are now downloaded with the correct formatting.
+- **🛠️ Installation of Tools & Functions Requirements**: Fixed a bug where necessary requirements for tools and functions were not properly installing.
+- **🔗 Inline Image Link Rendering**: Enabled rendering of images directly from links in chat.
+- **📞 Post-Call User Interface Cleanup**: Adjusted UI behavior to automatically close chat controls after a voice call ends, reducing screen clutter.
+- **🎙️ Microphone Deactivation Post-Call**: Addressed an issue where the microphone remained active after calls.
+- **✍️ Markdown Spacing Correction**: Corrected spacing in Markdown rendering, ensuring text appears neatly and as expected.
+- **🔄 Message Re-rendering**: Fixed an issue causing all response messages to re-render with each new message, now improving chat performance.
+
+### Changed
+
+- **🌐 Refined Web Search Integration**: Deprecated the Search Query Generation Prompt threshold; introduced a toggle button for "Enable Web Search Query Generation" allowing users to opt-in to using web search more judiciously.
+- **📝 Default Prompt Templates Update**: Emptied environment variable templates for search and title generation now default to the Open WebUI default prompt templates, simplifying configuration efforts.
+
+## [0.3.19] - 2024-09-05
+
+### Added
+
+- **🌐 Translation Update**: Improved Chinese translations.
+
+### Fixed
+
+- **📂 DATA_DIR Overriding**: Fixed an issue to avoid overriding DATA_DIR, preventing errors when directories are set identically, ensuring smoother operation and data management.
+- **🛠️ Frontmatter Extraction**: Fixed the extraction process for frontmatter in tools and functions.
+
+### Changed
+
+- **🎨 UI Styling**: Refined the user interface styling for enhanced visual coherence and user experience.
+
+## [0.3.18] - 2024-09-04
+
+### Added
+
+- **🛠️ Direct Database Execution for Tools & Functions**: Enhanced the execution of Python files for tools and functions, now directly loading from the database for a more streamlined backend process.
+
+### Fixed
+
+- **🔄 Automatic Rewrite of Import Statements in Tools & Functions**: Tool and function scripts that import 'utils', 'apps', 'main', 'config' will now automatically rename these with 'open_webui.', ensuring compatibility and consistency across different modules.
+- **🎨 Styling Adjustments**: Minor fixes in the visual styling to improve user experience and interface consistency.
+
+## [0.3.17] - 2024-09-04
+
+### Added
+
+- **🔄 Import/Export Configuration**: Users can now import and export webui configurations from admin settings > Database, simplifying setup replication across systems.
+- **🌍 Web Search via URL Parameter**: Added support for activating web search directly through URL by setting 'web-search=true'.
+- **🌐 SearchApi Integration**: Added support for SearchApi as an alternative web search provider, enhancing search capabilities within the platform.
+- **🔍 Literal Type Support in Tools**: Tools now support the Literal type.
+- **🌍 Updated Translations**: Improved translations for Chinese, Ukrainian, and Catalan.
+
+### Fixed
+
+- **🔧 Pip Install Issue**: Resolved the issue where pip install failed due to missing 'alembic.ini', ensuring smoother installation processes.
+- **🌃 Automatic Theme Update**: Fixed an issue where the color theme did not update dynamically with system changes.
+- **🛠️ User Agent in ComfyUI**: Added default headers in ComfyUI to fix access issues, improving reliability in network communications.
+- **🔄 Missing Chat Completion Response Headers**: Ensured proper return of proxied response headers during chat completion, improving API reliability.
+- **🔗 Websocket Connection Prioritization**: Modified socket.io configuration to prefer websockets and more reliably fallback to polling, enhancing connection stability.
+- **🎭 Accessibility Enhancements**: Added missing ARIA labels for buttons, improving accessibility for visually impaired users.
+- **⚖️ Advanced Parameter**: Fixed an issue ensuring that advanced parameters are correctly applied in all scenarios, ensuring consistent behavior of user-defined settings.
+
+### Changed
+
+- **🔁 Namespace Reorganization**: Reorganized all Python files under the 'open_webui' namespace to streamline the project structure and improve maintainability. Tools and functions importing from 'utils' should now use 'open_webui.utils'.
+- **🚧 Dependency Updates**: Updated several backend dependencies like 'aiohttp', 'authlib', 'duckduckgo-search', 'flask-cors', and 'langchain' to their latest versions, enhancing performance and security.
+
+## [0.3.16] - 2024-08-27
+
+### Added
+
+- **🚀 Config DB Migration**: Migrated configuration handling from config.json to the database, enabling high-availability setups and load balancing across multiple Open WebUI instances.
+- **🔗 Call Mode Activation via URL**: Added a 'call=true' URL search parameter enabling direct shortcuts to activate call mode, enhancing user interaction on mobile devices.
+- **✨ TTS Content Control**: Added functionality to control how message content is segmented for Text-to-Speech (TTS) generation requests, allowing for more flexible speech output options.
+- **😄 Show Knowledge Search Status**: Enhanced model usage transparency by displaying status when working with knowledge-augmented models, helping users understand the system's state during queries.
+- **👆 Click-to-Copy for Codespan**: Enhanced interactive experience in the WebUI by allowing users to click to copy content from code spans directly.
+- **🚫 API User Blocking via Model Filter**: Introduced the ability to block API users based on customized model filters, enhancing security and control over API access.
+- **🎬 Call Overlay Styling**: Adjusted call overlay styling on large screens to not cover the entire interface, but only the chat control area, for a more unobtrusive interaction experience.
+
+### Fixed
+
+- **🔧 LaTeX Rendering Issue**: Addressed an issue that affected the correct rendering of LaTeX.
+- **📁 File Leak Prevention**: Resolved the issue of uploaded files mistakenly being accessible across user chats.
+- **🔧 Pipe Functions with '**files**' Param**: Fixed issues with '**files**' parameter not functioning correctly in pipe functions.
+- **📝 Markdown Processing for RAG**: Fixed issues with processing Markdown in files.
+- **🚫 Duplicate System Prompts**: Fixed bugs causing system prompts to duplicate.
+
+### Changed
+
+- **🔋 Wakelock Permission**: Optimized the activation of wakelock to only engage during call mode, conserving device resources and improving battery performance during idle periods.
+- **🔍 Content-Type for Ollama Chats**: Added 'application/x-ndjson' content-type to '/api/chat' endpoint responses to match raw Ollama responses.
+- **✋ Disable Signups Conditionally**: Implemented conditional logic to disable sign-ups when 'ENABLE_LOGIN_FORM' is set to false.
+
+## [0.3.15] - 2024-08-21
+
+### Added
+
+- **🔗 Temporary Chat Activation**: Integrated a new URL parameter 'temporary-chat=true' to enable temporary chat sessions directly through the URL.
+- **🌄 ComfyUI Seed Node Support**: Introduced seed node support in ComfyUI for image generation, allowing users to specify node IDs for randomized seed assignment.
+
+### Fixed
+
+- **🛠️ Tools and Functions**: Resolved a critical issue where Tools and Functions were not properly functioning, restoring full capability and reliability to these essential features.
+- **🔘 Chat Action Button in Many Model Chat**: Fixed the malfunctioning of chat action buttons in many model chat environments, ensuring a smoother and more responsive user interaction.
+- **⏪ Many Model Chat Compatibility**: Restored backward compatibility for many model chats.
+
+## [0.3.14] - 2024-08-21
+
+### Added
+
+- **🛠️ Custom ComfyUI Workflow**: Deprecating several older environment variables, this enhancement introduces a new, customizable workflow for a more tailored user experience.
+- **🔀 Merge Responses in Many Model Chat**: Enhances the dialogue by merging responses from multiple models into a single, coherent reply, improving the interaction quality in many model chats.
+- **✅ Multiple Instances of Same Model in Chats**: Enhanced many model chat to support adding multiple instances of the same model.
+- **🔧 Quick Actions in Model Workspace**: Enhanced Shift key quick actions for hiding/unhiding and deleting models, facilitating a smoother workflow.
+- **🗨️ Markdown Rendering in User Messages**: User messages are now rendered in Markdown, enhancing readability and interaction.
+- **💬 Temporary Chat Feature**: Introduced a temporary chat feature, deprecating the old chat history setting to enhance user interaction flexibility.
+- **🖋️ User Message Editing**: Enhanced the user chat editing feature to allow saving changes without sending, providing more flexibility in message management.
+- **🛡️ Security Enhancements**: Various security improvements implemented across the platform to ensure safer user experiences.
+- **🌍 Updated Translations**: Enhanced translations for Chinese, Ukrainian, and Bahasa Malaysia, improving localization and user comprehension.
+
+### Fixed
+
+- **📑 Mermaid Rendering Issue**: Addressed issues with Mermaid chart rendering to ensure clean and clear visual data representation.
+- **🎭 PWA Icon Maskability**: Fixed the Progressive Web App icon to be maskable, ensuring proper display on various device home screens.
+- **🔀 Cloned Model Chat Freezing Issue**: Fixed a bug where cloning many model chats would cause freezing, enhancing stability and responsiveness.
+- **🔍 Generic Error Handling and Refinements**: Various minor fixes and refinements to address previously untracked issues, ensuring smoother operations.
+
+### Changed
+
+- **🖼️ Image Generation Refactor**: Overhauled image generation processes for improved efficiency and quality.
+- **🔨 Refactor Tool and Function Calling**: Refactored tool and function calling mechanisms for improved clarity and maintainability.
+- **🌐 Backend Library Updates**: Updated critical backend libraries including SQLAlchemy, uvicorn[standard], faster-whisper, bcrypt, and boto3 for enhanced performance and security.
+
+### Removed
+
+- **🚫 Deprecated ComfyUI Environment Variables**: Removed several outdated environment variables related to ComfyUI settings, simplifying configuration management.
+
+## [0.3.13] - 2024-08-14
+
+### Added
+
+- **🎨 Enhanced Markdown Rendering**: Significant improvements in rendering markdown, ensuring smooth and reliable display of LaTeX and Mermaid charts, enhancing user experience with more robust visual content.
+- **🔄 Auto-Install Tools & Functions Python Dependencies**: For 'Tools' and 'Functions', Open WebUI now automatically install extra python requirements specified in the frontmatter, streamlining setup processes and customization.
+- **🌀 OAuth Email Claim Customization**: Introduced an 'OAUTH_EMAIL_CLAIM' variable to allow customization of the default "email" claim within OAuth configurations, providing greater flexibility in authentication processes.
+- **📶 Websocket Reconnection**: Enhanced reliability with the capability to automatically reconnect when a websocket is closed, ensuring consistent and stable communication.
+- **🤳 Haptic Feedback on Support Devices**: Android devices now support haptic feedback for an immersive tactile experience during certain interactions.
+
+### Fixed
+
+- **🛠️ ComfyUI Performance Improvement**: Addressed an issue causing FastAPI to stall when ComfyUI image generation was active; now runs in a separate thread to prevent UI unresponsiveness.
+- **🔀 Session Handling**: Fixed an issue mandating session_id on client-side to ensure smoother session management and transitions.
+- **🖋️ Minor Bug Fixes and Format Corrections**: Various minor fixes including typo corrections, backend formatting improvements, and test amendments enhancing overall system stability and performance.
+
+### Changed
+
+- **🚀 Migration to SvelteKit 2**: Upgraded the underlying framework to SvelteKit version 2, offering enhanced speed, better code structure, and improved deployment capabilities.
+- **🧹 General Cleanup and Refactoring**: Performed broad cleanup and refactoring across the platform, improving code efficiency and maintaining high standards of code health.
+- **🚧 Integration Testing Improvements**: Modified how Cypress integration tests detect chat messages and updated sharing tests for better reliability and accuracy.
+- **📁 Standardized '.safetensors' File Extension**: Renamed the '.sft' file extension to '.safetensors' for ComfyUI workflows, standardizing file formats across the platform.
+
+### Removed
+
+- **🗑️ Deprecated Frontend Functions**: Removed frontend functions that were migrated to backend to declutter the codebase and reduce redundancy.
+
 ## [0.3.12] - 2024-08-07
 
 ### Added

CODE_OF_CONDUCT.md

@@ -2,76 +2,98 @@
 
 ## Our Pledge
 
-We as members, contributors, and leaders pledge to make participation in our
-community a harassment-free experience for everyone, regardless of age, body
-size, visible or invisible disability, ethnicity, sex characteristics, gender
-identity and expression, level of experience, education, socio-economic status,
-nationality, personal appearance, race, religion, or sexual identity
-and orientation.
+As members, contributors, and leaders of this community, we pledge to make participation in our open-source project a harassment-free experience for everyone, regardless of age, body size, visible or invisible disability, ethnicity, sex characteristics, gender identity and expression, level of experience, education, socio-economic status, nationality, personal appearance, race, religion, or sexual identity and orientation.
 
-We pledge to act and interact in ways that contribute to an open, welcoming, diverse, inclusive, and healthy community.
+We are committed to creating and maintaining an open, respectful, and professional environment where positive contributions and meaningful discussions can flourish. By participating in this project, you agree to uphold these values and align your behavior to the standards outlined in this Code of Conduct.
+
+## Why These Standards Are Important
+
+Open-source projects rely on a community of volunteers dedicating their time, expertise, and effort toward a shared goal. These projects are inherently collaborative but also fragile, as the success of the project depends on the goodwill, energy, and productivity of those involved.
+
+Maintaining a positive and respectful environment is essential to safeguarding the integrity of this project and protecting contributors' efforts. Behavior that disrupts this atmosphere—whether through hostility, entitlement, or unprofessional conduct—can severely harm the morale and productivity of the community. **Strict enforcement of these standards ensures a safe and supportive space for meaningful collaboration.**
+
+This is a community where **respect and professionalism are mandatory.** Violations of these standards will result in **zero tolerance** and immediate enforcement to prevent disruption and ensure the well-being of all participants.
 
 ## Our Standards
 
-Examples of behavior that contribute to a positive environment for our community include:
+Examples of behavior that contribute to a positive and professional community include:
 
-- Demonstrating empathy and kindness toward other people
-- Being respectful of differing opinions, viewpoints, and experiences
-- Giving and gracefully accepting constructive feedback
-- Accepting responsibility and apologizing to those affected by our mistakes, and learning from the experience
-- Focusing on what is best not just for us as individuals, but for the overall community
+- **Respecting others.** Be considerate, listen actively, and engage with empathy toward others' viewpoints and experiences.
+- **Constructive feedback.** Provide actionable, thoughtful, and respectful feedback that helps improve the project and encourages collaboration. Avoid unproductive negativity or hypercriticism.
+- **Recognizing volunteer contributions.** Appreciate that contributors dedicate their free time and resources selflessly. Approach them with gratitude and patience.
+- **Focusing on shared goals.** Collaborate in ways that prioritize the health, success, and sustainability of the community over individual agendas.
 
 Examples of unacceptable behavior include:
 
-- The use of sexualized language or imagery, and sexual attention or advances of any kind
-- Trolling, insulting or derogatory comments, and personal or political attacks
-- Public or private harassment
-- Publishing others' private information, such as a physical or email address, without their explicit permission
-- **Spamming of any kind**
-- Aggressive sales tactics targeting our community members are strictly prohibited. You can mention your product if it's relevant to the discussion, but under no circumstances should you push it forcefully
-- Other conduct which could reasonably be considered inappropriate in a professional setting
+- The use of discriminatory, demeaning, or sexualized language or behavior.
+- Personal attacks, derogatory comments, trolling, or inflammatory political or ideological arguments.
+- Harassment, intimidation, or any behavior intended to create a hostile, uncomfortable, or unsafe environment.
+- Publishing others' private information (e.g., physical or email addresses) without explicit permission.
+- **Entitlement, demand, or aggression toward contributors.** Volunteers are under no obligation to provide immediate or personalized support. Rude or dismissive behavior will not be tolerated.
+- **Unproductive or destructive behavior.** This includes venting frustration as hostility ("tantrums"), hypercriticism, attention-seeking negativity, or anything that distracts from the project's goals.
+- **Spamming and promotional exploitation.** Sharing irrelevant product promotions or self-promotion in the community is not allowed unless it directly contributes value to the discussion.
+
+### Feedback and Community Engagement
+
+- **Constructive feedback is encouraged, but hostile or entitled behavior will result in immediate action.** If you disagree with elements of the project, we encourage you to offer meaningful improvements or fork the project if necessary. Healthy discussions and technical disagreements are welcome only when handled with professionalism.
+- **Respect contributors' time and efforts.** No one is entitled to personalized or on-demand assistance. This is a community built on collaboration and shared effort; demanding or demeaning behavior undermines that trust and will not be allowed.
+
+### Zero Tolerance: No Warnings, Immediate Action
+
+This community operates under a **zero-tolerance policy.** Any behavior deemed unacceptable under this Code of Conduct will result in **immediate enforcement, without prior warning.**
+
+We employ this approach to ensure that unproductive or disruptive behavior does not escalate further or cause unnecessary harm to other contributors. The standards are clear, and violations of any kind—whether mild or severe—will be addressed decisively to protect the community.
 
 ## Enforcement Responsibilities
 
-Community leaders are responsible for clarifying and enforcing our standards of acceptable behavior and will take appropriate and fair corrective action in response to any behavior that they deem inappropriate, threatening, offensive, or harmful.
+Community leaders are responsible for upholding and enforcing these standards. They are empowered to take **immediate and appropriate action** to address any behaviors they deem unacceptable under this Code of Conduct. These actions are taken with the goal of protecting the community and preserving its safe, positive, and productive environment.
 
 ## Scope
 
-This Code of Conduct applies within all community spaces and also applies when an individual is officially representing the community in public spaces. Examples of representing our community include using an official e-mail address, posting via an official social media account, or acting as an appointed representative at an online or offline event.
+This Code of Conduct applies to all community spaces, including forums, repositories, social media accounts, and in-person events. It also applies when an individual represents the community in public settings, such as conferences or official communications.
 
-## Enforcement
+Additionally, any behavior outside of these defined spaces that negatively impacts the community or its members may fall within the scope of this Code of Conduct.
 
-Instances of abusive, harassing, spamming, or otherwise unacceptable behavior may be reported to the community leaders responsible for enforcement at hello@openwebui.com. All complaints will be reviewed and investigated promptly and fairly.
+## Reporting Violations
 
-All community leaders are obligated to respect the privacy and security of the reporter of any incident.
+Instances of unacceptable behavior can be reported to the leadership team at **hello@openwebui.com**. Reports will be handled promptly, confidentially, and with consideration for the safety and well-being of the reporter.
+
+All community leaders are required to uphold confidentiality and impartiality when addressing reports of violations.
 
 ## Enforcement Guidelines
 
-Community leaders will follow these Community Impact Guidelines in determining the consequences for any action they deem in violation of this Code of Conduct:
+### Ban
 
-### 1. Temporary Ban
+**Community Impact**: Community leaders will issue a ban to any participant whose behavior is deemed unacceptable according to this Code of Conduct. Bans are enforced immediately and without prior notice.
 
-**Community Impact**: Any violation of community standards, including but not limited to inappropriate language, unprofessional behavior, harassment, or spamming.
+A ban may be temporary or permanent, depending on the severity of the violation. This includes—but is not limited to—behavior such as:
 
-**Consequence**: A temporary ban from any sort of interaction or public communication with the community for a specified period of time. No public or private interaction with the people involved, including unsolicited interaction with those enforcing the Code of Conduct, is allowed during this period. Violating these terms may lead to a permanent ban.
+- Harassment or abusive behavior toward contributors.
+- Persistent negativity or hostility that disrupts the collaborative environment.
+- Disrespectful, demanding, or aggressive interactions with others.
+- Attempts to cause harm or sabotage the community.
 
-### 2. Permanent Ban
+**Consequence**: A banned individual is immediately removed from access to all community spaces, communication channels, and events. Community leaders reserve the right to enforce either a time-limited suspension or a permanent ban based on the specific circumstances of the violation.
 
-**Community Impact**: Repeated or severe violations of community standards, including sustained inappropriate behavior, harassment of an individual, or aggression toward or disparagement of classes of individuals.
+This approach ensures that disruptive behaviors are addressed swiftly and decisively in order to maintain the integrity and productivity of the community.
 
-**Consequence**: A permanent ban from any sort of public interaction within the community.
+## Why Zero Tolerance Is Necessary
+
+Open-source projects thrive on collaboration, goodwill, and mutual respect. Toxic behaviors—such as entitlement, hostility, or persistent negativity—threaten not just individual contributors but the health of the project as a whole. Allowing such behaviors to persist robs contributors of their time, energy, and enthusiasm for the work they do.
+
+By enforcing a zero-tolerance policy, we ensure that the community remains a safe, welcoming space for all participants. These measures are not about harshness—they are about protecting contributors and fostering a productive environment where innovation can thrive.
+
+Our expectations are clear, and our enforcement reflects our commitment to this project's long-term success.
 
 ## Attribution
 
-This Code of Conduct is adapted from the [Contributor Covenant][homepage],
-version 2.0, available at
+This Code of Conduct is adapted from the [Contributor Covenant][homepage], version 2.0, available at  
 https://www.contributor-covenant.org/version/2/0/code_of_conduct.html.
 
-Community Impact Guidelines were inspired by [Mozilla's code of conduct
-enforcement ladder](https://github.com/mozilla/diversity).
+Community Impact Guidelines were inspired by [Mozilla's code of conduct enforcement ladder](https://github.com/mozilla/diversity).
 
 [homepage]: https://www.contributor-covenant.org
 
-For answers to common questions about this code of conduct, see the FAQ at
-https://www.contributor-covenant.org/faq. Translations are available at
+For answers to common questions about this code of conduct, see the FAQ at  
+https://www.contributor-covenant.org/faq. Translations are available at  
 https://www.contributor-covenant.org/translations.

Dockerfile

@@ -1,6 +1,6 @@
 # syntax=docker/dockerfile:1
 # Initialize device type args
-# use build args in the docker build commmand with --build-arg="BUILDARG=true"
+# use build args in the docker build command with --build-arg="BUILDARG=true"
 ARG USE_CUDA=false
 ARG USE_OLLAMA=false
 # Tested with cu117 for CUDA 11 and cu121 for CUDA 12 (default)
@@ -11,13 +11,17 @@ ARG USE_CUDA_VER=cu121
 # IMPORTANT: If you change the embedding model (sentence-transformers/all-MiniLM-L6-v2) and vice versa, you aren't able to use RAG Chat with your previous documents loaded in the WebUI! You need to re-embed them.
 ARG USE_EMBEDDING_MODEL=sentence-transformers/all-MiniLM-L6-v2
 ARG USE_RERANKING_MODEL=""
+
+# Tiktoken encoding name; models to use can be found at https://huggingface.co/models?library=tiktoken
+ARG USE_TIKTOKEN_ENCODING_NAME="cl100k_base"
+
 ARG BUILD_HASH=dev-build
 # Override at your own risk - non-root configurations are untested
 ARG UID=0
 ARG GID=0
 
 ######## WebUI frontend ########
-FROM --platform=$BUILDPLATFORM node:21-alpine3.19 as build
+FROM --platform=$BUILDPLATFORM node:22-alpine3.20 AS build
 ARG BUILD_HASH
 
 WORKDIR /app
@@ -30,7 +34,7 @@ ENV APP_BUILD_HASH=${BUILD_HASH}
 RUN npm run build
 
 ######## WebUI backend ########
-FROM python:3.11-slim-bookworm as base
+FROM python:3.11-slim-bookworm AS base
 
 # Use args
 ARG USE_CUDA
@@ -72,13 +76,21 @@ ENV RAG_EMBEDDING_MODEL="$USE_EMBEDDING_MODEL_DOCKER" \
     RAG_RERANKING_MODEL="$USE_RERANKING_MODEL_DOCKER" \
     SENTENCE_TRANSFORMERS_HOME="/app/backend/data/cache/embedding/models"
 
+## Tiktoken model settings ##
+ENV TIKTOKEN_ENCODING_NAME="cl100k_base" \
+    TIKTOKEN_CACHE_DIR="/app/backend/data/cache/tiktoken"
+
 ## Hugging Face download cache ##
 ENV HF_HOME="/app/backend/data/cache/embedding/models"
+
+## Torch Extensions ##
+# ENV TORCH_EXTENSIONS_DIR="/.cache/torch_extensions"
+
 #### Other models ##########################################################
 
 WORKDIR /app/backend
 
-ENV HOME /root
+ENV HOME=/root
 # Create user and group if not root
 RUN if [ $UID -ne 0 ]; then \
     if [ $GID -ne 0 ]; then \
@@ -96,7 +108,7 @@ RUN chown -R $UID:$GID /app $HOME
 RUN if [ "$USE_OLLAMA" = "true" ]; then \
     apt-get update && \
     # Install pandoc and netcat
-    apt-get install -y --no-install-recommends pandoc netcat-openbsd curl && \
+    apt-get install -y --no-install-recommends git build-essential pandoc netcat-openbsd curl && \
     apt-get install -y --no-install-recommends gcc python3-dev && \
     # for RAG OCR
     apt-get install -y --no-install-recommends ffmpeg libsm6 libxext6 && \
@@ -109,7 +121,7 @@ RUN if [ "$USE_OLLAMA" = "true" ]; then \
     else \
     apt-get update && \
     # Install pandoc, netcat and gcc
-    apt-get install -y --no-install-recommends pandoc gcc netcat-openbsd curl jq && \
+    apt-get install -y --no-install-recommends git build-essential pandoc gcc netcat-openbsd curl jq && \
     apt-get install -y --no-install-recommends gcc python3-dev && \
     # for RAG OCR
     apt-get install -y --no-install-recommends ffmpeg libsm6 libxext6 && \
@@ -127,11 +139,13 @@ RUN pip3 install uv && \
     uv pip install --system -r requirements.txt --no-cache-dir && \
     python -c "import os; from sentence_transformers import SentenceTransformer; SentenceTransformer(os.environ['RAG_EMBEDDING_MODEL'], device='cpu')" && \
     python -c "import os; from faster_whisper import WhisperModel; WhisperModel(os.environ['WHISPER_MODEL'], device='cpu', compute_type='int8', download_root=os.environ['WHISPER_MODEL_DIR'])"; \
+    python -c "import os; import tiktoken; tiktoken.get_encoding(os.environ['TIKTOKEN_ENCODING_NAME'])"; \
     else \
     pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cpu --no-cache-dir && \
     uv pip install --system -r requirements.txt --no-cache-dir && \
     python -c "import os; from sentence_transformers import SentenceTransformer; SentenceTransformer(os.environ['RAG_EMBEDDING_MODEL'], device='cpu')" && \
     python -c "import os; from faster_whisper import WhisperModel; WhisperModel(os.environ['WHISPER_MODEL'], device='cpu', compute_type='int8', download_root=os.environ['WHISPER_MODEL_DIR'])"; \
+    python -c "import os; import tiktoken; tiktoken.get_encoding(os.environ['TIKTOKEN_ENCODING_NAME'])"; \
     fi; \
     chown -R $UID:$GID /app/backend/data/
 
@@ -157,5 +171,6 @@ USER $UID:$GID
 
 ARG BUILD_HASH
 ENV WEBUI_BUILD_VERSION=${BUILD_HASH}
+ENV DOCKER=true
 
 CMD [ "bash", "start.sh"]

README.md

@@ -1,4 +1,4 @@
-# Open WebUI (Formerly Ollama WebUI) 👋
+# Open WebUI 👋
 
 ![GitHub stars](https://img.shields.io/github/stars/open-webui/open-webui?style=social)
 ![GitHub forks](https://img.shields.io/github/forks/open-webui/open-webui?style=social)
@@ -21,7 +21,7 @@ Open WebUI is an [extensible](https://github.com/open-webui/pipelines), feature-
 
 - 🤝 **Ollama/OpenAI API Integration**: Effortlessly integrate OpenAI-compatible APIs for versatile conversations alongside Ollama models. Customize the OpenAI API URL to link with **LMStudio, GroqCloud, Mistral, OpenRouter, and more**.
 
-- 🧩 **Pipelines, Open WebUI Plugin Support**: Seamlessly integrate custom logic and Python libraries into Open WebUI using [Pipelines Plugin Framework](https://github.com/open-webui/pipelines). Launch your Pipelines instance, set the OpenAI URL to the Pipelines URL, and explore endless possibilities. [Examples](https://github.com/open-webui/pipelines/tree/main/examples) include **Function Calling**, User **Rate Limiting** to control access, **Usage Monitoring** with tools like Langfuse, **Live Translation with LibreTranslate** for multilingual support, **Toxic Message Filtering** and much more.
+- 🛡️ **Granular Permissions and User Groups**: By allowing administrators to create detailed user roles and permissions, we ensure a secure user environment. This granularity not only enhances security but also allows for customized user experiences, fostering a sense of ownership and responsibility amongst users.
 
 - 📱 **Responsive Design**: Enjoy a seamless experience across Desktop PC, Laptop, and Mobile devices.
 
@@ -37,7 +37,7 @@ Open WebUI is an [extensible](https://github.com/open-webui/pipelines), feature-
 
 - 📚 **Local RAG Integration**: Dive into the future of chat interactions with groundbreaking Retrieval Augmented Generation (RAG) support. This feature seamlessly integrates document interactions into your chat experience. You can load documents directly into the chat or add files to your document library, effortlessly accessing them using the `#` command before a query.
 
-- 🔍 **Web Search for RAG**: Perform web searches using providers like `SearXNG`, `Google PSE`, `Brave Search`, `serpstack`, `serper`, `Serply`, `DuckDuckGo` and `TavilySearch` and inject the results directly into your chat experience.
+- 🔍 **Web Search for RAG**: Perform web searches using providers like `SearXNG`, `Google PSE`, `Brave Search`, `serpstack`, `serper`, `Serply`, `DuckDuckGo`, `TavilySearch`, `SearchApi` and `Bing` and inject the results directly into your chat experience.
 
 - 🌐 **Web Browsing Capability**: Seamlessly integrate websites into your chat experience using the `#` command followed by a URL. This feature allows you to incorporate web content directly into your conversations, enhancing the richness and depth of your interactions.
 
@@ -49,6 +49,8 @@ Open WebUI is an [extensible](https://github.com/open-webui/pipelines), feature-
 
 - 🌐🌍 **Multilingual Support**: Experience Open WebUI in your preferred language with our internationalization (i18n) support. Join us in expanding our supported languages! We're actively seeking contributors!
 
+- 🧩 **Pipelines, Open WebUI Plugin Support**: Seamlessly integrate custom logic and Python libraries into Open WebUI using [Pipelines Plugin Framework](https://github.com/open-webui/pipelines). Launch your Pipelines instance, set the OpenAI URL to the Pipelines URL, and explore endless possibilities. [Examples](https://github.com/open-webui/pipelines/tree/main/examples) include **Function Calling**, User **Rate Limiting** to control access, **Usage Monitoring** with tools like Langfuse, **Live Translation with LibreTranslate** for multilingual support, **Toxic Message Filtering** and much more.
+
 - 🌟 **Continuous Updates**: We are committed to improving Open WebUI with regular updates, fixes, and new features.
 
 Want to learn more about Open WebUI's features? Check out our [Open WebUI documentation](https://docs.openwebui.com/features) for a comprehensive overview!
@@ -59,11 +61,31 @@ Don't forget to explore our sibling project, [Open WebUI Community](https://open
 
 ## How to Install 🚀
 
-> [!NOTE]  
-> Please note that for certain Docker environments, additional configurations might be needed. If you encounter any connection issues, our detailed guide on [Open WebUI Documentation](https://docs.openwebui.com/) is ready to assist you.
+### Installation via Python pip 🐍
+
+Open WebUI can be installed using pip, the Python package installer. Before proceeding, ensure you're using **Python 3.11** to avoid compatibility issues.
+
+1. **Install Open WebUI**:
+   Open your terminal and run the following command to install Open WebUI:
+
+   ```bash
+   pip install open-webui
+   ```
+
+2. **Running Open WebUI**:
+   After installation, you can start Open WebUI by executing:
+
+   ```bash
+   open-webui serve
+   ```
+
+This will start the Open WebUI server, which you can access at [http://localhost:8080](http://localhost:8080)
 
 ### Quick Start with Docker 🐳
 
+> [!NOTE]  
+> Please note that for certain Docker environments, additional configurations might be needed. If you encounter any connection issues, our detailed guide on [Open WebUI Documentation](https://docs.openwebui.com/) is ready to assist you.
+
 > [!WARNING]
 > When using Docker to install Open WebUI, make sure to include the `-v open-webui:/app/backend/data` in your Docker command. This step is crucial as it ensures your database is properly mounted and prevents any loss of data.
 
@@ -86,7 +108,7 @@ Don't forget to explore our sibling project, [Open WebUI Community](https://open
   docker run -d -p 3000:8080 -e OLLAMA_BASE_URL=https://example.com -v open-webui:/app/backend/data --name open-webui --restart always ghcr.io/open-webui/open-webui:main
   ```
 
-  - **To run Open WebUI with Nvidia GPU support**, use this command:
+- **To run Open WebUI with Nvidia GPU support**, use this command:
 
   ```bash
   docker run -d -p 3000:8080 --gpus all --add-host=host.docker.internal:host-gateway -v open-webui:/app/backend/data --name open-webui --restart always ghcr.io/open-webui/open-webui:cuda
@@ -150,7 +172,7 @@ docker run --rm --volume /var/run/docker.sock:/var/run/docker.sock containrrr/wa
 
 In the last part of the command, replace `open-webui` with your container name if it is different.
 
-Check our Migration Guide available in our [Open WebUI Documentation](https://docs.openwebui.com/migration/).
+Check our Migration Guide available in our [Open WebUI Documentation](https://docs.openwebui.com/tutorials/migration/).
 
 ### Using the Dev Branch 🌙
 
@@ -167,18 +189,6 @@ docker run -d -p 3000:8080 -v open-webui:/app/backend/data --name open-webui --a
 
 Discover upcoming features on our roadmap in the [Open WebUI Documentation](https://docs.openwebui.com/roadmap/).
 
-## Supporters ✨
-
-A big shoutout to our amazing supporters who's helping to make this project possible! 🙏
-
-### Platinum Sponsors 🤍
-
-- We're looking for Sponsors!
-
-### Acknowledgments
-
-Special thanks to [Prof. Lawrence Kim](https://www.lhkim.com/) and [Prof. Nick Vincent](https://www.nickmvincent.com/) for their invaluable support and guidance in shaping this project into a research endeavor. Grateful for your mentorship throughout the journey! 🙌
-
 ## License 📜
 
 This project is licensed under the [MIT License](LICENSE) - see the [LICENSE](LICENSE) file for details. 📄
@@ -200,4 +210,4 @@ If you have any questions, suggestions, or need assistance, please open an issue
 
 ---
 
-Created by [Timothy J. Baek](https://github.com/tjbck) - Let's make Open WebUI even more amazing together! 💪
+Created by [Timothy Jaeryang Baek](https://github.com/tjbck) - Let's make Open WebUI even more amazing together! 💪

TROUBLESHOOTING.md

@@ -18,7 +18,7 @@ If you're experiencing connection issues, it’s often due to the WebUI docker c
 docker run -d --network=host -v open-webui:/app/backend/data -e OLLAMA_BASE_URL=http://127.0.0.1:11434 --name open-webui --restart always ghcr.io/open-webui/open-webui:main
 ```
 
-### Error on Slow Reponses for Ollama
+### Error on Slow Responses for Ollama
 
 Open WebUI has a default timeout of 5 minutes for Ollama to finish generating the response. If needed, this can be adjusted via the environment variable AIOHTTP_CLIENT_TIMEOUT, which sets the timeout in seconds.
 

backend/.gitignore

@@ -8,9 +8,5 @@ _test
 Pipfile
 !/data
 /data/*
-!/data/litellm
-/data/litellm/*
-!data/litellm/config.yaml
-
-!data/config.json
+/open_webui/data/*
 .webui_secret_key

backend/apps/audio/main.py

@@ -1,504 +0,0 @@
-import os
-import logging
-from fastapi import (
-    FastAPI,
-    Request,
-    Depends,
-    HTTPException,
-    status,
-    UploadFile,
-    File,
-    Form,
-)
-from fastapi.responses import StreamingResponse, JSONResponse, FileResponse
-
-from fastapi.middleware.cors import CORSMiddleware
-from pydantic import BaseModel
-
-from typing import List
-import uuid
-import requests
-import hashlib
-from pathlib import Path
-import json
-
-from constants import ERROR_MESSAGES
-from utils.utils import (
-    decode_token,
-    get_current_user,
-    get_verified_user,
-    get_admin_user,
-)
-from utils.misc import calculate_sha256
-
-
-from config import (
-    SRC_LOG_LEVELS,
-    CACHE_DIR,
-    UPLOAD_DIR,
-    WHISPER_MODEL,
-    WHISPER_MODEL_DIR,
-    WHISPER_MODEL_AUTO_UPDATE,
-    DEVICE_TYPE,
-    AUDIO_STT_OPENAI_API_BASE_URL,
-    AUDIO_STT_OPENAI_API_KEY,
-    AUDIO_TTS_OPENAI_API_BASE_URL,
-    AUDIO_TTS_OPENAI_API_KEY,
-    AUDIO_TTS_API_KEY,
-    AUDIO_STT_ENGINE,
-    AUDIO_STT_MODEL,
-    AUDIO_TTS_ENGINE,
-    AUDIO_TTS_MODEL,
-    AUDIO_TTS_VOICE,
-    AppConfig,
-)
-
-log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["AUDIO"])
-
-app = FastAPI()
-app.add_middleware(
-    CORSMiddleware,
-    allow_origins=["*"],
-    allow_credentials=True,
-    allow_methods=["*"],
-    allow_headers=["*"],
-)
-
-app.state.config = AppConfig()
-
-app.state.config.STT_OPENAI_API_BASE_URL = AUDIO_STT_OPENAI_API_BASE_URL
-app.state.config.STT_OPENAI_API_KEY = AUDIO_STT_OPENAI_API_KEY
-app.state.config.STT_ENGINE = AUDIO_STT_ENGINE
-app.state.config.STT_MODEL = AUDIO_STT_MODEL
-
-app.state.config.TTS_OPENAI_API_BASE_URL = AUDIO_TTS_OPENAI_API_BASE_URL
-app.state.config.TTS_OPENAI_API_KEY = AUDIO_TTS_OPENAI_API_KEY
-app.state.config.TTS_ENGINE = AUDIO_TTS_ENGINE
-app.state.config.TTS_MODEL = AUDIO_TTS_MODEL
-app.state.config.TTS_VOICE = AUDIO_TTS_VOICE
-app.state.config.TTS_API_KEY = AUDIO_TTS_API_KEY
-
-# setting device type for whisper model
-whisper_device_type = DEVICE_TYPE if DEVICE_TYPE and DEVICE_TYPE == "cuda" else "cpu"
-log.info(f"whisper_device_type: {whisper_device_type}")
-
-SPEECH_CACHE_DIR = Path(CACHE_DIR).joinpath("./audio/speech/")
-SPEECH_CACHE_DIR.mkdir(parents=True, exist_ok=True)
-
-
-class TTSConfigForm(BaseModel):
-    OPENAI_API_BASE_URL: str
-    OPENAI_API_KEY: str
-    API_KEY: str
-    ENGINE: str
-    MODEL: str
-    VOICE: str
-
-
-class STTConfigForm(BaseModel):
-    OPENAI_API_BASE_URL: str
-    OPENAI_API_KEY: str
-    ENGINE: str
-    MODEL: str
-
-
-class AudioConfigUpdateForm(BaseModel):
-    tts: TTSConfigForm
-    stt: STTConfigForm
-
-
-from pydub import AudioSegment
-from pydub.utils import mediainfo
-
-
-def is_mp4_audio(file_path):
-    """Check if the given file is an MP4 audio file."""
-    if not os.path.isfile(file_path):
-        print(f"File not found: {file_path}")
-        return False
-
-    info = mediainfo(file_path)
-    if (
-        info.get("codec_name") == "aac"
-        and info.get("codec_type") == "audio"
-        and info.get("codec_tag_string") == "mp4a"
-    ):
-        return True
-    return False
-
-
-def convert_mp4_to_wav(file_path, output_path):
-    """Convert MP4 audio file to WAV format."""
-    audio = AudioSegment.from_file(file_path, format="mp4")
-    audio.export(output_path, format="wav")
-    print(f"Converted {file_path} to {output_path}")
-
-
-@app.get("/config")
-async def get_audio_config(user=Depends(get_admin_user)):
-    return {
-        "tts": {
-            "OPENAI_API_BASE_URL": app.state.config.TTS_OPENAI_API_BASE_URL,
-            "OPENAI_API_KEY": app.state.config.TTS_OPENAI_API_KEY,
-            "API_KEY": app.state.config.TTS_API_KEY,
-            "ENGINE": app.state.config.TTS_ENGINE,
-            "MODEL": app.state.config.TTS_MODEL,
-            "VOICE": app.state.config.TTS_VOICE,
-        },
-        "stt": {
-            "OPENAI_API_BASE_URL": app.state.config.STT_OPENAI_API_BASE_URL,
-            "OPENAI_API_KEY": app.state.config.STT_OPENAI_API_KEY,
-            "ENGINE": app.state.config.STT_ENGINE,
-            "MODEL": app.state.config.STT_MODEL,
-        },
-    }
-
-
-@app.post("/config/update")
-async def update_audio_config(
-    form_data: AudioConfigUpdateForm, user=Depends(get_admin_user)
-):
-    app.state.config.TTS_OPENAI_API_BASE_URL = form_data.tts.OPENAI_API_BASE_URL
-    app.state.config.TTS_OPENAI_API_KEY = form_data.tts.OPENAI_API_KEY
-    app.state.config.TTS_API_KEY = form_data.tts.API_KEY
-    app.state.config.TTS_ENGINE = form_data.tts.ENGINE
-    app.state.config.TTS_MODEL = form_data.tts.MODEL
-    app.state.config.TTS_VOICE = form_data.tts.VOICE
-
-    app.state.config.STT_OPENAI_API_BASE_URL = form_data.stt.OPENAI_API_BASE_URL
-    app.state.config.STT_OPENAI_API_KEY = form_data.stt.OPENAI_API_KEY
-    app.state.config.STT_ENGINE = form_data.stt.ENGINE
-    app.state.config.STT_MODEL = form_data.stt.MODEL
-
-    return {
-        "tts": {
-            "OPENAI_API_BASE_URL": app.state.config.TTS_OPENAI_API_BASE_URL,
-            "OPENAI_API_KEY": app.state.config.TTS_OPENAI_API_KEY,
-            "API_KEY": app.state.config.TTS_API_KEY,
-            "ENGINE": app.state.config.TTS_ENGINE,
-            "MODEL": app.state.config.TTS_MODEL,
-            "VOICE": app.state.config.TTS_VOICE,
-        },
-        "stt": {
-            "OPENAI_API_BASE_URL": app.state.config.STT_OPENAI_API_BASE_URL,
-            "OPENAI_API_KEY": app.state.config.STT_OPENAI_API_KEY,
-            "ENGINE": app.state.config.STT_ENGINE,
-            "MODEL": app.state.config.STT_MODEL,
-        },
-    }
-
-
-@app.post("/speech")
-async def speech(request: Request, user=Depends(get_verified_user)):
-    body = await request.body()
-    name = hashlib.sha256(body).hexdigest()
-
-    file_path = SPEECH_CACHE_DIR.joinpath(f"{name}.mp3")
-    file_body_path = SPEECH_CACHE_DIR.joinpath(f"{name}.json")
-
-    # Check if the file already exists in the cache
-    if file_path.is_file():
-        return FileResponse(file_path)
-
-    if app.state.config.TTS_ENGINE == "openai":
-        headers = {}
-        headers["Authorization"] = f"Bearer {app.state.config.TTS_OPENAI_API_KEY}"
-        headers["Content-Type"] = "application/json"
-
-        try:
-            body = body.decode("utf-8")
-            body = json.loads(body)
-            body["model"] = app.state.config.TTS_MODEL
-            body = json.dumps(body).encode("utf-8")
-        except Exception as e:
-            pass
-
-        r = None
-        try:
-            r = requests.post(
-                url=f"{app.state.config.TTS_OPENAI_API_BASE_URL}/audio/speech",
-                data=body,
-                headers=headers,
-                stream=True,
-            )
-
-            r.raise_for_status()
-
-            # Save the streaming content to a file
-            with open(file_path, "wb") as f:
-                for chunk in r.iter_content(chunk_size=8192):
-                    f.write(chunk)
-
-            with open(file_body_path, "w") as f:
-                json.dump(json.loads(body.decode("utf-8")), f)
-
-            # Return the saved file
-            return FileResponse(file_path)
-
-        except Exception as e:
-            log.exception(e)
-            error_detail = "Open WebUI: Server Connection Error"
-            if r is not None:
-                try:
-                    res = r.json()
-                    if "error" in res:
-                        error_detail = f"External: {res['error']['message']}"
-                except:
-                    error_detail = f"External: {e}"
-
-            raise HTTPException(
-                status_code=r.status_code if r != None else 500,
-                detail=error_detail,
-            )
-
-    elif app.state.config.TTS_ENGINE == "elevenlabs":
-        payload = None
-        try:
-            payload = json.loads(body.decode("utf-8"))
-        except Exception as e:
-            log.exception(e)
-            raise HTTPException(status_code=400, detail="Invalid JSON payload")
-
-        voice_id = payload.get("voice", "")
-        url = f"https://api.elevenlabs.io/v1/text-to-speech/{voice_id}"
-
-        headers = {
-            "Accept": "audio/mpeg",
-            "Content-Type": "application/json",
-            "xi-api-key": app.state.config.TTS_API_KEY,
-        }
-
-        data = {
-            "text": payload["input"],
-            "model_id": app.state.config.TTS_MODEL,
-            "voice_settings": {"stability": 0.5, "similarity_boost": 0.5},
-        }
-
-        try:
-            r = requests.post(url, json=data, headers=headers)
-
-            r.raise_for_status()
-
-            # Save the streaming content to a file
-            with open(file_path, "wb") as f:
-                for chunk in r.iter_content(chunk_size=8192):
-                    f.write(chunk)
-
-            with open(file_body_path, "w") as f:
-                json.dump(json.loads(body.decode("utf-8")), f)
-
-            # Return the saved file
-            return FileResponse(file_path)
-
-        except Exception as e:
-            log.exception(e)
-            error_detail = "Open WebUI: Server Connection Error"
-            if r is not None:
-                try:
-                    res = r.json()
-                    if "error" in res:
-                        error_detail = f"External: {res['error']['message']}"
-                except:
-                    error_detail = f"External: {e}"
-
-            raise HTTPException(
-                status_code=r.status_code if r != None else 500,
-                detail=error_detail,
-            )
-
-
-@app.post("/transcriptions")
-def transcribe(
-    file: UploadFile = File(...),
-    user=Depends(get_current_user),
-):
-    log.info(f"file.content_type: {file.content_type}")
-
-    if file.content_type not in ["audio/mpeg", "audio/wav"]:
-        raise HTTPException(
-            status_code=status.HTTP_400_BAD_REQUEST,
-            detail=ERROR_MESSAGES.FILE_NOT_SUPPORTED,
-        )
-
-    try:
-        ext = file.filename.split(".")[-1]
-
-        id = uuid.uuid4()
-        filename = f"{id}.{ext}"
-
-        file_dir = f"{CACHE_DIR}/audio/transcriptions"
-        os.makedirs(file_dir, exist_ok=True)
-        file_path = f"{file_dir}/{filename}"
-
-        print(filename)
-
-        contents = file.file.read()
-        with open(file_path, "wb") as f:
-            f.write(contents)
-            f.close()
-
-        if app.state.config.STT_ENGINE == "":
-            from faster_whisper import WhisperModel
-
-            whisper_kwargs = {
-                "model_size_or_path": WHISPER_MODEL,
-                "device": whisper_device_type,
-                "compute_type": "int8",
-                "download_root": WHISPER_MODEL_DIR,
-                "local_files_only": not WHISPER_MODEL_AUTO_UPDATE,
-            }
-
-            log.debug(f"whisper_kwargs: {whisper_kwargs}")
-
-            try:
-                model = WhisperModel(**whisper_kwargs)
-            except:
-                log.warning(
-                    "WhisperModel initialization failed, attempting download with local_files_only=False"
-                )
-                whisper_kwargs["local_files_only"] = False
-                model = WhisperModel(**whisper_kwargs)
-
-            segments, info = model.transcribe(file_path, beam_size=5)
-            log.info(
-                "Detected language '%s' with probability %f"
-                % (info.language, info.language_probability)
-            )
-
-            transcript = "".join([segment.text for segment in list(segments)])
-
-            data = {"text": transcript.strip()}
-
-            # save the transcript to a json file
-            transcript_file = f"{file_dir}/{id}.json"
-            with open(transcript_file, "w") as f:
-                json.dump(data, f)
-
-            print(data)
-
-            return data
-
-        elif app.state.config.STT_ENGINE == "openai":
-            if is_mp4_audio(file_path):
-                print("is_mp4_audio")
-                os.rename(file_path, file_path.replace(".wav", ".mp4"))
-                # Convert MP4 audio file to WAV format
-                convert_mp4_to_wav(file_path.replace(".wav", ".mp4"), file_path)
-
-            headers = {"Authorization": f"Bearer {app.state.config.STT_OPENAI_API_KEY}"}
-
-            files = {"file": (filename, open(file_path, "rb"))}
-            data = {"model": app.state.config.STT_MODEL}
-
-            print(files, data)
-
-            r = None
-            try:
-                r = requests.post(
-                    url=f"{app.state.config.STT_OPENAI_API_BASE_URL}/audio/transcriptions",
-                    headers=headers,
-                    files=files,
-                    data=data,
-                )
-
-                r.raise_for_status()
-
-                data = r.json()
-
-                # save the transcript to a json file
-                transcript_file = f"{file_dir}/{id}.json"
-                with open(transcript_file, "w") as f:
-                    json.dump(data, f)
-
-                print(data)
-                return data
-            except Exception as e:
-                log.exception(e)
-                error_detail = "Open WebUI: Server Connection Error"
-                if r is not None:
-                    try:
-                        res = r.json()
-                        if "error" in res:
-                            error_detail = f"External: {res['error']['message']}"
-                    except:
-                        error_detail = f"External: {e}"
-
-                raise HTTPException(
-                    status_code=r.status_code if r != None else 500,
-                    detail=error_detail,
-                )
-
-    except Exception as e:
-        log.exception(e)
-
-        raise HTTPException(
-            status_code=status.HTTP_400_BAD_REQUEST,
-            detail=ERROR_MESSAGES.DEFAULT(e),
-        )
-
-
-def get_available_models() -> List[dict]:
-    if app.state.config.TTS_ENGINE == "openai":
-        return [{"id": "tts-1"}, {"id": "tts-1-hd"}]
-    elif app.state.config.TTS_ENGINE == "elevenlabs":
-        headers = {
-            "xi-api-key": app.state.config.TTS_API_KEY,
-            "Content-Type": "application/json",
-        }
-
-        try:
-            response = requests.get(
-                "https://api.elevenlabs.io/v1/models", headers=headers
-            )
-            response.raise_for_status()
-            models = response.json()
-            return [
-                {"name": model["name"], "id": model["model_id"]} for model in models
-            ]
-        except requests.RequestException as e:
-            log.error(f"Error fetching voices: {str(e)}")
-    return []
-
-
-@app.get("/models")
-async def get_models(user=Depends(get_verified_user)):
-    return {"models": get_available_models()}
-
-
-def get_available_voices() -> List[dict]:
-    if app.state.config.TTS_ENGINE == "openai":
-        return [
-            {"name": "alloy", "id": "alloy"},
-            {"name": "echo", "id": "echo"},
-            {"name": "fable", "id": "fable"},
-            {"name": "onyx", "id": "onyx"},
-            {"name": "nova", "id": "nova"},
-            {"name": "shimmer", "id": "shimmer"},
-        ]
-    elif app.state.config.TTS_ENGINE == "elevenlabs":
-        headers = {
-            "xi-api-key": app.state.config.TTS_API_KEY,
-            "Content-Type": "application/json",
-        }
-
-        try:
-            response = requests.get(
-                "https://api.elevenlabs.io/v1/voices", headers=headers
-            )
-            response.raise_for_status()
-            voices_data = response.json()
-
-            voices = []
-            for voice in voices_data.get("voices", []):
-                voices.append({"name": voice["name"], "id": voice["voice_id"]})
-            return voices
-        except requests.RequestException as e:
-            log.error(f"Error fetching voices: {str(e)}")
-
-    return []
-
-
-@app.get("/voices")
-async def get_voices(user=Depends(get_verified_user)):
-    return {"voices": get_available_voices()}

backend/apps/images/utils/comfyui.py

@@ -1,407 +0,0 @@
-import websocket  # NOTE: websocket-client (https://github.com/websocket-client/websocket-client)
-import uuid
-import json
-import urllib.request
-import urllib.parse
-import random
-import logging
-
-from config import SRC_LOG_LEVELS
-
-log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["COMFYUI"])
-
-from pydantic import BaseModel
-
-from typing import Optional
-
-COMFYUI_DEFAULT_PROMPT = """
-{
-  "3": {
-    "inputs": {
-      "seed": 0,
-      "steps": 20,
-      "cfg": 8,
-      "sampler_name": "euler",
-      "scheduler": "normal",
-      "denoise": 1,
-      "model": [
-        "4",
-        0
-      ],
-      "positive": [
-        "6",
-        0
-      ],
-      "negative": [
-        "7",
-        0
-      ],
-      "latent_image": [
-        "5",
-        0
-      ]
-    },
-    "class_type": "KSampler",
-    "_meta": {
-      "title": "KSampler"
-    }
-  },
-  "4": {
-    "inputs": {
-      "ckpt_name": "model.safetensors"
-    },
-    "class_type": "CheckpointLoaderSimple",
-    "_meta": {
-      "title": "Load Checkpoint"
-    }
-  },
-  "5": {
-    "inputs": {
-      "width": 512,
-      "height": 512,
-      "batch_size": 1
-    },
-    "class_type": "EmptyLatentImage",
-    "_meta": {
-      "title": "Empty Latent Image"
-    }
-  },
-  "6": {
-    "inputs": {
-      "text": "Prompt",
-      "clip": [
-        "4",
-        1
-      ]
-    },
-    "class_type": "CLIPTextEncode",
-    "_meta": {
-      "title": "CLIP Text Encode (Prompt)"
-    }
-  },
-  "7": {
-    "inputs": {
-      "text": "Negative Prompt",
-      "clip": [
-        "4",
-        1
-      ]
-    },
-    "class_type": "CLIPTextEncode",
-    "_meta": {
-      "title": "CLIP Text Encode (Prompt)"
-    }
-  },
-  "8": {
-    "inputs": {
-      "samples": [
-        "3",
-        0
-      ],
-      "vae": [
-        "4",
-        2
-      ]
-    },
-    "class_type": "VAEDecode",
-    "_meta": {
-      "title": "VAE Decode"
-    }
-  },
-  "9": {
-    "inputs": {
-      "filename_prefix": "ComfyUI",
-      "images": [
-        "8",
-        0
-      ]
-    },
-    "class_type": "SaveImage",
-    "_meta": {
-      "title": "Save Image"
-    }
-  }
-}
-"""
-
-FLUX_DEFAULT_PROMPT = """
-{
-    "5": {
-        "inputs": {
-            "width": 1024,
-            "height": 1024,
-            "batch_size": 1
-        },
-        "class_type": "EmptyLatentImage"
-    },
-    "6": {
-        "inputs": {
-            "text": "Input Text Here",
-            "clip": [
-                "11",
-                0
-            ]
-        },
-        "class_type": "CLIPTextEncode"
-    },
-    "8": {
-        "inputs": {
-            "samples": [
-                "13",
-                0
-            ],
-            "vae": [
-                "10",
-                0
-            ]
-        },
-        "class_type": "VAEDecode"
-    },
-    "9": {
-        "inputs": {
-            "filename_prefix": "ComfyUI",
-            "images": [
-                "8",
-                0
-            ]
-        },
-        "class_type": "SaveImage"
-    },
-    "10": {
-        "inputs": {
-            "vae_name": "ae.sft"
-        },
-        "class_type": "VAELoader"
-    },
-    "11": {
-        "inputs": {
-            "clip_name1": "clip_l.safetensors",
-            "clip_name2": "t5xxl_fp16.safetensors",
-            "type": "flux"
-        },
-        "class_type": "DualCLIPLoader"
-    },
-    "12": {
-        "inputs": {
-            "unet_name": "flux1-dev.sft",
-            "weight_dtype": "default"
-        },
-        "class_type": "UNETLoader"
-    },
-    "13": {
-        "inputs": {
-            "noise": [
-                "25",
-                0
-            ],
-            "guider": [
-                "22",
-                0
-            ],
-            "sampler": [
-                "16",
-                0
-            ],
-            "sigmas": [
-                "17",
-                0
-            ],
-            "latent_image": [
-                "5",
-                0
-            ]
-        },
-        "class_type": "SamplerCustomAdvanced"
-    },
-    "16": {
-        "inputs": {
-            "sampler_name": "euler"
-        },
-        "class_type": "KSamplerSelect"
-    },
-    "17": {
-        "inputs": {
-            "scheduler": "simple",
-            "steps": 20,
-            "denoise": 1,
-            "model": [
-                "12",
-                0
-            ]
-        },
-        "class_type": "BasicScheduler"
-    },
-    "22": {
-        "inputs": {
-            "model": [
-                "12",
-                0
-            ],
-            "conditioning": [
-                "6",
-                0
-            ]
-        },
-        "class_type": "BasicGuider"
-    },
-    "25": {
-        "inputs": {
-            "noise_seed": 778937779713005
-        },
-        "class_type": "RandomNoise"
-    }
-}
-"""
-
-
-def queue_prompt(prompt, client_id, base_url):
-    log.info("queue_prompt")
-    p = {"prompt": prompt, "client_id": client_id}
-    data = json.dumps(p).encode("utf-8")
-    req = urllib.request.Request(f"{base_url}/prompt", data=data)
-    return json.loads(urllib.request.urlopen(req).read())
-
-
-def get_image(filename, subfolder, folder_type, base_url):
-    log.info("get_image")
-    data = {"filename": filename, "subfolder": subfolder, "type": folder_type}
-    url_values = urllib.parse.urlencode(data)
-    with urllib.request.urlopen(f"{base_url}/view?{url_values}") as response:
-        return response.read()
-
-
-def get_image_url(filename, subfolder, folder_type, base_url):
-    log.info("get_image")
-    data = {"filename": filename, "subfolder": subfolder, "type": folder_type}
-    url_values = urllib.parse.urlencode(data)
-    return f"{base_url}/view?{url_values}"
-
-
-def get_history(prompt_id, base_url):
-    log.info("get_history")
-    with urllib.request.urlopen(f"{base_url}/history/{prompt_id}") as response:
-        return json.loads(response.read())
-
-
-def get_images(ws, prompt, client_id, base_url):
-    prompt_id = queue_prompt(prompt, client_id, base_url)["prompt_id"]
-    output_images = []
-    while True:
-        out = ws.recv()
-        if isinstance(out, str):
-            message = json.loads(out)
-            if message["type"] == "executing":
-                data = message["data"]
-                if data["node"] is None and data["prompt_id"] == prompt_id:
-                    break  # Execution is done
-        else:
-            continue  # previews are binary data
-
-    history = get_history(prompt_id, base_url)[prompt_id]
-    for o in history["outputs"]:
-        for node_id in history["outputs"]:
-            node_output = history["outputs"][node_id]
-            if "images" in node_output:
-                for image in node_output["images"]:
-                    url = get_image_url(
-                        image["filename"], image["subfolder"], image["type"], base_url
-                    )
-                    output_images.append({"url": url})
-    return {"data": output_images}
-
-
-class ImageGenerationPayload(BaseModel):
-    prompt: str
-    negative_prompt: Optional[str] = ""
-    steps: Optional[int] = None
-    seed: Optional[int] = None
-    width: int
-    height: int
-    n: int = 1
-    cfg_scale: Optional[float] = None
-    sampler: Optional[str] = None
-    scheduler: Optional[str] = None
-    sd3: Optional[bool] = None
-    flux: Optional[bool] = None
-    flux_weight_dtype: Optional[str] = None
-    flux_fp8_clip: Optional[bool] = None
-
-
-def comfyui_generate_image(
-    model: str, payload: ImageGenerationPayload, client_id, base_url
-):
-    ws_url = base_url.replace("http://", "ws://").replace("https://", "wss://")
-
-    comfyui_prompt = json.loads(COMFYUI_DEFAULT_PROMPT)
-
-    if payload.cfg_scale:
-        comfyui_prompt["3"]["inputs"]["cfg"] = payload.cfg_scale
-
-    if payload.sampler:
-        comfyui_prompt["3"]["inputs"]["sampler"] = payload.sampler
-
-    if payload.scheduler:
-        comfyui_prompt["3"]["inputs"]["scheduler"] = payload.scheduler
-
-    if payload.sd3:
-        comfyui_prompt["5"]["class_type"] = "EmptySD3LatentImage"
-
-    if payload.steps:
-        comfyui_prompt["3"]["inputs"]["steps"] = payload.steps
-
-    comfyui_prompt["4"]["inputs"]["ckpt_name"] = model
-    comfyui_prompt["7"]["inputs"]["text"] = payload.negative_prompt
-    comfyui_prompt["3"]["inputs"]["seed"] = (
-        payload.seed if payload.seed else random.randint(0, 18446744073709551614)
-    )
-
-    # as Flux uses a completely different workflow, we must treat it specially
-    if payload.flux:
-        comfyui_prompt = json.loads(FLUX_DEFAULT_PROMPT)
-        comfyui_prompt["12"]["inputs"]["unet_name"] = model
-        comfyui_prompt["25"]["inputs"]["noise_seed"] = (
-            payload.seed if payload.seed else random.randint(0, 18446744073709551614)
-        )
-
-        if payload.sampler:
-            comfyui_prompt["16"]["inputs"]["sampler_name"] = payload.sampler
-
-        if payload.steps:
-            comfyui_prompt["17"]["inputs"]["steps"] = payload.steps
-
-        if payload.scheduler:
-            comfyui_prompt["17"]["inputs"]["scheduler"] = payload.scheduler
-
-        if payload.flux_weight_dtype:
-            comfyui_prompt["12"]["inputs"]["weight_dtype"] = payload.flux_weight_dtype
-
-        if payload.flux_fp8_clip:
-            comfyui_prompt["11"]["inputs"][
-                "clip_name2"
-            ] = "t5xxl_fp8_e4m3fn.safetensors"
-
-    comfyui_prompt["5"]["inputs"]["batch_size"] = payload.n
-    comfyui_prompt["5"]["inputs"]["width"] = payload.width
-    comfyui_prompt["5"]["inputs"]["height"] = payload.height
-
-    # set the text prompt for our positive CLIPTextEncode
-    comfyui_prompt["6"]["inputs"]["text"] = payload.prompt
-
-    try:
-        ws = websocket.WebSocket()
-        ws.connect(f"{ws_url}/ws?clientId={client_id}")
-        log.info("WebSocket connection established.")
-    except Exception as e:
-        log.exception(f"Failed to connect to WebSocket server: {e}")
-        return None
-
-    try:
-        images = get_images(ws, comfyui_prompt, client_id, base_url)
-    except Exception as e:
-        log.exception(f"Error while receiving images: {e}")
-        images = None
-
-    ws.close()
-
-    return images

backend/apps/openai/main.py

@@ -1,508 +0,0 @@
-from fastapi import FastAPI, Request, HTTPException, Depends
-from fastapi.middleware.cors import CORSMiddleware
-from fastapi.responses import StreamingResponse, FileResponse
-
-import requests
-import aiohttp
-import asyncio
-import json
-import logging
-
-from pydantic import BaseModel
-from starlette.background import BackgroundTask
-
-from apps.webui.models.models import Models
-from constants import ERROR_MESSAGES
-from utils.utils import (
-    get_verified_user,
-    get_admin_user,
-)
-from utils.misc import apply_model_params_to_body, apply_model_system_prompt_to_body
-
-from config import (
-    SRC_LOG_LEVELS,
-    ENABLE_OPENAI_API,
-    AIOHTTP_CLIENT_TIMEOUT,
-    OPENAI_API_BASE_URLS,
-    OPENAI_API_KEYS,
-    CACHE_DIR,
-    ENABLE_MODEL_FILTER,
-    MODEL_FILTER_LIST,
-    AppConfig,
-)
-from typing import List, Optional, Literal, overload
-
-
-import hashlib
-from pathlib import Path
-
-log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["OPENAI"])
-
-app = FastAPI()
-app.add_middleware(
-    CORSMiddleware,
-    allow_origins=["*"],
-    allow_credentials=True,
-    allow_methods=["*"],
-    allow_headers=["*"],
-)
-
-
-app.state.config = AppConfig()
-
-app.state.config.ENABLE_MODEL_FILTER = ENABLE_MODEL_FILTER
-app.state.config.MODEL_FILTER_LIST = MODEL_FILTER_LIST
-
-app.state.config.ENABLE_OPENAI_API = ENABLE_OPENAI_API
-app.state.config.OPENAI_API_BASE_URLS = OPENAI_API_BASE_URLS
-app.state.config.OPENAI_API_KEYS = OPENAI_API_KEYS
-
-app.state.MODELS = {}
-
-
-@app.middleware("http")
-async def check_url(request: Request, call_next):
-    if len(app.state.MODELS) == 0:
-        await get_all_models()
-
-    response = await call_next(request)
-    return response
-
-
-@app.get("/config")
-async def get_config(user=Depends(get_admin_user)):
-    return {"ENABLE_OPENAI_API": app.state.config.ENABLE_OPENAI_API}
-
-
-class OpenAIConfigForm(BaseModel):
-    enable_openai_api: Optional[bool] = None
-
-
-@app.post("/config/update")
-async def update_config(form_data: OpenAIConfigForm, user=Depends(get_admin_user)):
-    app.state.config.ENABLE_OPENAI_API = form_data.enable_openai_api
-    return {"ENABLE_OPENAI_API": app.state.config.ENABLE_OPENAI_API}
-
-
-class UrlsUpdateForm(BaseModel):
-    urls: List[str]
-
-
-class KeysUpdateForm(BaseModel):
-    keys: List[str]
-
-
-@app.get("/urls")
-async def get_openai_urls(user=Depends(get_admin_user)):
-    return {"OPENAI_API_BASE_URLS": app.state.config.OPENAI_API_BASE_URLS}
-
-
-@app.post("/urls/update")
-async def update_openai_urls(form_data: UrlsUpdateForm, user=Depends(get_admin_user)):
-    await get_all_models()
-    app.state.config.OPENAI_API_BASE_URLS = form_data.urls
-    return {"OPENAI_API_BASE_URLS": app.state.config.OPENAI_API_BASE_URLS}
-
-
-@app.get("/keys")
-async def get_openai_keys(user=Depends(get_admin_user)):
-    return {"OPENAI_API_KEYS": app.state.config.OPENAI_API_KEYS}
-
-
-@app.post("/keys/update")
-async def update_openai_key(form_data: KeysUpdateForm, user=Depends(get_admin_user)):
-    app.state.config.OPENAI_API_KEYS = form_data.keys
-    return {"OPENAI_API_KEYS": app.state.config.OPENAI_API_KEYS}
-
-
-@app.post("/audio/speech")
-async def speech(request: Request, user=Depends(get_verified_user)):
-    idx = None
-    try:
-        idx = app.state.config.OPENAI_API_BASE_URLS.index("https://api.openai.com/v1")
-        body = await request.body()
-        name = hashlib.sha256(body).hexdigest()
-
-        SPEECH_CACHE_DIR = Path(CACHE_DIR).joinpath("./audio/speech/")
-        SPEECH_CACHE_DIR.mkdir(parents=True, exist_ok=True)
-        file_path = SPEECH_CACHE_DIR.joinpath(f"{name}.mp3")
-        file_body_path = SPEECH_CACHE_DIR.joinpath(f"{name}.json")
-
-        # Check if the file already exists in the cache
-        if file_path.is_file():
-            return FileResponse(file_path)
-
-        headers = {}
-        headers["Authorization"] = f"Bearer {app.state.config.OPENAI_API_KEYS[idx]}"
-        headers["Content-Type"] = "application/json"
-        if "openrouter.ai" in app.state.config.OPENAI_API_BASE_URLS[idx]:
-            headers["HTTP-Referer"] = "https://openwebui.com/"
-            headers["X-Title"] = "Open WebUI"
-        r = None
-        try:
-            r = requests.post(
-                url=f"{app.state.config.OPENAI_API_BASE_URLS[idx]}/audio/speech",
-                data=body,
-                headers=headers,
-                stream=True,
-            )
-
-            r.raise_for_status()
-
-            # Save the streaming content to a file
-            with open(file_path, "wb") as f:
-                for chunk in r.iter_content(chunk_size=8192):
-                    f.write(chunk)
-
-            with open(file_body_path, "w") as f:
-                json.dump(json.loads(body.decode("utf-8")), f)
-
-            # Return the saved file
-            return FileResponse(file_path)
-
-        except Exception as e:
-            log.exception(e)
-            error_detail = "Open WebUI: Server Connection Error"
-            if r is not None:
-                try:
-                    res = r.json()
-                    if "error" in res:
-                        error_detail = f"External: {res['error']}"
-                except Exception:
-                    error_detail = f"External: {e}"
-
-            raise HTTPException(
-                status_code=r.status_code if r else 500, detail=error_detail
-            )
-
-    except ValueError:
-        raise HTTPException(status_code=401, detail=ERROR_MESSAGES.OPENAI_NOT_FOUND)
-
-
-async def fetch_url(url, key):
-    timeout = aiohttp.ClientTimeout(total=5)
-    try:
-        headers = {"Authorization": f"Bearer {key}"}
-        async with aiohttp.ClientSession(timeout=timeout, trust_env=True) as session:
-            async with session.get(url, headers=headers) as response:
-                return await response.json()
-    except Exception as e:
-        # Handle connection error here
-        log.error(f"Connection error: {e}")
-        return None
-
-
-async def cleanup_response(
-    response: Optional[aiohttp.ClientResponse],
-    session: Optional[aiohttp.ClientSession],
-):
-    if response:
-        response.close()
-    if session:
-        await session.close()
-
-
-def merge_models_lists(model_lists):
-    log.debug(f"merge_models_lists {model_lists}")
-    merged_list = []
-
-    for idx, models in enumerate(model_lists):
-        if models is not None and "error" not in models:
-            merged_list.extend(
-                [
-                    {
-                        **model,
-                        "name": model.get("name", model["id"]),
-                        "owned_by": "openai",
-                        "openai": model,
-                        "urlIdx": idx,
-                    }
-                    for model in models
-                    if "api.openai.com"
-                    not in app.state.config.OPENAI_API_BASE_URLS[idx]
-                    or "gpt" in model["id"]
-                ]
-            )
-
-    return merged_list
-
-
-def is_openai_api_disabled():
-    api_keys = app.state.config.OPENAI_API_KEYS
-    no_keys = len(api_keys) == 1 and api_keys[0] == ""
-    return no_keys or not app.state.config.ENABLE_OPENAI_API
-
-
-async def get_all_models_raw() -> list:
-    if is_openai_api_disabled():
-        return []
-
-    # Check if API KEYS length is same than API URLS length
-    num_urls = len(app.state.config.OPENAI_API_BASE_URLS)
-    num_keys = len(app.state.config.OPENAI_API_KEYS)
-
-    if num_keys != num_urls:
-        # if there are more keys than urls, remove the extra keys
-        if num_keys > num_urls:
-            new_keys = app.state.config.OPENAI_API_KEYS[:num_urls]
-            app.state.config.OPENAI_API_KEYS = new_keys
-        # if there are more urls than keys, add empty keys
-        else:
-            app.state.config.OPENAI_API_KEYS += [""] * (num_urls - num_keys)
-
-    tasks = [
-        fetch_url(f"{url}/models", app.state.config.OPENAI_API_KEYS[idx])
-        for idx, url in enumerate(app.state.config.OPENAI_API_BASE_URLS)
-    ]
-
-    responses = await asyncio.gather(*tasks)
-    log.debug(f"get_all_models:responses() {responses}")
-
-    return responses
-
-
-@overload
-async def get_all_models(raw: Literal[True]) -> list: ...
-
-
-@overload
-async def get_all_models(raw: Literal[False] = False) -> dict[str, list]: ...
-
-
-async def get_all_models(raw=False) -> dict[str, list] | list:
-    log.info("get_all_models()")
-    if is_openai_api_disabled():
-        return [] if raw else {"data": []}
-
-    responses = await get_all_models_raw()
-    if raw:
-        return responses
-
-    def extract_data(response):
-        if response and "data" in response:
-            return response["data"]
-        if isinstance(response, list):
-            return response
-        return None
-
-    models = {"data": merge_models_lists(map(extract_data, responses))}
-
-    log.debug(f"models: {models}")
-    app.state.MODELS = {model["id"]: model for model in models["data"]}
-
-    return models
-
-
-@app.get("/models")
-@app.get("/models/{url_idx}")
-async def get_models(url_idx: Optional[int] = None, user=Depends(get_verified_user)):
-    if url_idx is None:
-        models = await get_all_models()
-        if app.state.config.ENABLE_MODEL_FILTER:
-            if user.role == "user":
-                models["data"] = list(
-                    filter(
-                        lambda model: model["id"] in app.state.config.MODEL_FILTER_LIST,
-                        models["data"],
-                    )
-                )
-                return models
-        return models
-    else:
-        url = app.state.config.OPENAI_API_BASE_URLS[url_idx]
-        key = app.state.config.OPENAI_API_KEYS[url_idx]
-
-        headers = {}
-        headers["Authorization"] = f"Bearer {key}"
-        headers["Content-Type"] = "application/json"
-
-        r = None
-
-        try:
-            r = requests.request(method="GET", url=f"{url}/models", headers=headers)
-            r.raise_for_status()
-
-            response_data = r.json()
-            if "api.openai.com" in url:
-                response_data["data"] = list(
-                    filter(lambda model: "gpt" in model["id"], response_data["data"])
-                )
-
-            return response_data
-        except Exception as e:
-            log.exception(e)
-            error_detail = "Open WebUI: Server Connection Error"
-            if r is not None:
-                try:
-                    res = r.json()
-                    if "error" in res:
-                        error_detail = f"External: {res['error']}"
-                except Exception:
-                    error_detail = f"External: {e}"
-
-            raise HTTPException(
-                status_code=r.status_code if r else 500,
-                detail=error_detail,
-            )
-
-
-@app.post("/chat/completions")
-@app.post("/chat/completions/{url_idx}")
-async def generate_chat_completion(
-    form_data: dict,
-    url_idx: Optional[int] = None,
-    user=Depends(get_verified_user),
-):
-    idx = 0
-    payload = {**form_data}
-    payload.pop("metadata")
-
-    model_id = form_data.get("model")
-    model_info = Models.get_model_by_id(model_id)
-
-    if model_info:
-        if model_info.base_model_id:
-            payload["model"] = model_info.base_model_id
-
-        params = model_info.params.model_dump()
-        payload = apply_model_params_to_body(params, payload)
-        payload = apply_model_system_prompt_to_body(params, payload, user)
-
-    model = app.state.MODELS[payload.get("model")]
-    idx = model["urlIdx"]
-
-    if "pipeline" in model and model.get("pipeline"):
-        payload["user"] = {
-            "name": user.name,
-            "id": user.id,
-            "email": user.email,
-            "role": user.role,
-        }
-
-    # Convert the modified body back to JSON
-    payload = json.dumps(payload)
-
-    log.debug(payload)
-
-    url = app.state.config.OPENAI_API_BASE_URLS[idx]
-    key = app.state.config.OPENAI_API_KEYS[idx]
-
-    headers = {}
-    headers["Authorization"] = f"Bearer {key}"
-    headers["Content-Type"] = "application/json"
-    if "openrouter.ai" in app.state.config.OPENAI_API_BASE_URLS[idx]:
-        headers["HTTP-Referer"] = "https://openwebui.com/"
-        headers["X-Title"] = "Open WebUI"
-
-    r = None
-    session = None
-    streaming = False
-
-    try:
-        session = aiohttp.ClientSession(
-            trust_env=True, timeout=aiohttp.ClientTimeout(total=AIOHTTP_CLIENT_TIMEOUT)
-        )
-        r = await session.request(
-            method="POST",
-            url=f"{url}/chat/completions",
-            data=payload,
-            headers=headers,
-        )
-
-        r.raise_for_status()
-
-        # Check if response is SSE
-        if "text/event-stream" in r.headers.get("Content-Type", ""):
-            streaming = True
-            return StreamingResponse(
-                r.content,
-                status_code=r.status,
-                headers=dict(r.headers),
-                background=BackgroundTask(
-                    cleanup_response, response=r, session=session
-                ),
-            )
-        else:
-            response_data = await r.json()
-            return response_data
-    except Exception as e:
-        log.exception(e)
-        error_detail = "Open WebUI: Server Connection Error"
-        if r is not None:
-            try:
-                res = await r.json()
-                print(res)
-                if "error" in res:
-                    error_detail = f"External: {res['error']['message'] if 'message' in res['error'] else res['error']}"
-            except Exception:
-                error_detail = f"External: {e}"
-        raise HTTPException(status_code=r.status if r else 500, detail=error_detail)
-    finally:
-        if not streaming and session:
-            if r:
-                r.close()
-            await session.close()
-
-
-@app.api_route("/{path:path}", methods=["GET", "POST", "PUT", "DELETE"])
-async def proxy(path: str, request: Request, user=Depends(get_verified_user)):
-    idx = 0
-
-    body = await request.body()
-
-    url = app.state.config.OPENAI_API_BASE_URLS[idx]
-    key = app.state.config.OPENAI_API_KEYS[idx]
-
-    target_url = f"{url}/{path}"
-
-    headers = {}
-    headers["Authorization"] = f"Bearer {key}"
-    headers["Content-Type"] = "application/json"
-
-    r = None
-    session = None
-    streaming = False
-
-    try:
-        session = aiohttp.ClientSession(trust_env=True)
-        r = await session.request(
-            method=request.method,
-            url=target_url,
-            data=body,
-            headers=headers,
-        )
-
-        r.raise_for_status()
-
-        # Check if response is SSE
-        if "text/event-stream" in r.headers.get("Content-Type", ""):
-            streaming = True
-            return StreamingResponse(
-                r.content,
-                status_code=r.status,
-                headers=dict(r.headers),
-                background=BackgroundTask(
-                    cleanup_response, response=r, session=session
-                ),
-            )
-        else:
-            response_data = await r.json()
-            return response_data
-    except Exception as e:
-        log.exception(e)
-        error_detail = "Open WebUI: Server Connection Error"
-        if r is not None:
-            try:
-                res = await r.json()
-                print(res)
-                if "error" in res:
-                    error_detail = f"External: {res['error']['message'] if 'message' in res['error'] else res['error']}"
-            except Exception:
-                error_detail = f"External: {e}"
-        raise HTTPException(status_code=r.status if r else 500, detail=error_detail)
-    finally:
-        if not streaming and session:
-            if r:
-                r.close()
-            await session.close()

backend/apps/socket/main.py

@@ -1,171 +0,0 @@
-import socketio
-import asyncio
-
-
-from apps.webui.models.users import Users
-from utils.utils import decode_token
-
-sio = socketio.AsyncServer(cors_allowed_origins=[], async_mode="asgi")
-app = socketio.ASGIApp(sio, socketio_path="/ws/socket.io")
-
-# Dictionary to maintain the user pool
-
-SESSION_POOL = {}
-USER_POOL = {}
-USAGE_POOL = {}
-# Timeout duration in seconds
-TIMEOUT_DURATION = 3
-
-
-@sio.event
-async def connect(sid, environ, auth):
-    user = None
-    if auth and "token" in auth:
-        data = decode_token(auth["token"])
-
-        if data is not None and "id" in data:
-            user = Users.get_user_by_id(data["id"])
-
-        if user:
-            SESSION_POOL[sid] = user.id
-            if user.id in USER_POOL:
-                USER_POOL[user.id].append(sid)
-            else:
-                USER_POOL[user.id] = [sid]
-
-            print(f"user {user.name}({user.id}) connected with session ID {sid}")
-
-            await sio.emit("user-count", {"count": len(set(USER_POOL))})
-            await sio.emit("usage", {"models": get_models_in_use()})
-
-
-@sio.on("user-join")
-async def user_join(sid, data):
-    print("user-join", sid, data)
-
-    auth = data["auth"] if "auth" in data else None
-    if not auth or "token" not in auth:
-        return
-
-    data = decode_token(auth["token"])
-    if data is None or "id" not in data:
-        return
-
-    user = Users.get_user_by_id(data["id"])
-    if not user:
-        return
-
-    SESSION_POOL[sid] = user.id
-    if user.id in USER_POOL:
-        USER_POOL[user.id].append(sid)
-    else:
-        USER_POOL[user.id] = [sid]
-
-    print(f"user {user.name}({user.id}) connected with session ID {sid}")
-
-    await sio.emit("user-count", {"count": len(set(USER_POOL))})
-
-
-@sio.on("user-count")
-async def user_count(sid):
-    await sio.emit("user-count", {"count": len(set(USER_POOL))})
-
-
-def get_models_in_use():
-    # Aggregate all models in use
-    models_in_use = []
-    for model_id, data in USAGE_POOL.items():
-        models_in_use.append(model_id)
-
-    return models_in_use
-
-
-@sio.on("usage")
-async def usage(sid, data):
-    model_id = data["model"]
-
-    # Cancel previous callback if there is one
-    if model_id in USAGE_POOL:
-        USAGE_POOL[model_id]["callback"].cancel()
-
-    # Store the new usage data and task
-
-    if model_id in USAGE_POOL:
-        USAGE_POOL[model_id]["sids"].append(sid)
-        USAGE_POOL[model_id]["sids"] = list(set(USAGE_POOL[model_id]["sids"]))
-
-    else:
-        USAGE_POOL[model_id] = {"sids": [sid]}
-
-    # Schedule a task to remove the usage data after TIMEOUT_DURATION
-    USAGE_POOL[model_id]["callback"] = asyncio.create_task(
-        remove_after_timeout(sid, model_id)
-    )
-
-    # Broadcast the usage data to all clients
-    await sio.emit("usage", {"models": get_models_in_use()})
-
-
-async def remove_after_timeout(sid, model_id):
-    try:
-        await asyncio.sleep(TIMEOUT_DURATION)
-        if model_id in USAGE_POOL:
-            print(USAGE_POOL[model_id]["sids"])
-            USAGE_POOL[model_id]["sids"].remove(sid)
-            USAGE_POOL[model_id]["sids"] = list(set(USAGE_POOL[model_id]["sids"]))
-
-            if len(USAGE_POOL[model_id]["sids"]) == 0:
-                del USAGE_POOL[model_id]
-
-            # Broadcast the usage data to all clients
-            await sio.emit("usage", {"models": get_models_in_use()})
-    except asyncio.CancelledError:
-        # Task was cancelled due to new 'usage' event
-        pass
-
-
-@sio.event
-async def disconnect(sid):
-    if sid in SESSION_POOL:
-        user_id = SESSION_POOL[sid]
-        del SESSION_POOL[sid]
-
-        USER_POOL[user_id].remove(sid)
-
-        if len(USER_POOL[user_id]) == 0:
-            del USER_POOL[user_id]
-
-        await sio.emit("user-count", {"count": len(USER_POOL)})
-    else:
-        print(f"Unknown session ID {sid} disconnected")
-
-
-def get_event_emitter(request_info):
-    async def __event_emitter__(event_data):
-        await sio.emit(
-            "chat-events",
-            {
-                "chat_id": request_info["chat_id"],
-                "message_id": request_info["message_id"],
-                "data": event_data,
-            },
-            to=request_info["session_id"],
-        )
-
-    return __event_emitter__
-
-
-def get_event_call(request_info):
-    async def __event_call__(event_data):
-        response = await sio.call(
-            "chat-events",
-            {
-                "chat_id": request_info["chat_id"],
-                "message_id": request_info["message_id"],
-                "data": event_data,
-            },
-            to=request_info["session_id"],
-        )
-        return response
-
-    return __event_call__

backend/apps/webui/models/chats.py

@@ -1,407 +0,0 @@
-from pydantic import BaseModel, ConfigDict
-from typing import List, Union, Optional
-
-import json
-import uuid
-import time
-
-from sqlalchemy import Column, String, BigInteger, Boolean, Text
-
-from apps.webui.internal.db import Base, get_db
-
-
-####################
-# Chat DB Schema
-####################
-
-
-class Chat(Base):
-    __tablename__ = "chat"
-
-    id = Column(String, primary_key=True)
-    user_id = Column(String)
-    title = Column(Text)
-    chat = Column(Text)  # Save Chat JSON as Text
-
-    created_at = Column(BigInteger)
-    updated_at = Column(BigInteger)
-
-    share_id = Column(Text, unique=True, nullable=True)
-    archived = Column(Boolean, default=False)
-
-
-class ChatModel(BaseModel):
-    model_config = ConfigDict(from_attributes=True)
-
-    id: str
-    user_id: str
-    title: str
-    chat: str
-
-    created_at: int  # timestamp in epoch
-    updated_at: int  # timestamp in epoch
-
-    share_id: Optional[str] = None
-    archived: bool = False
-
-
-####################
-# Forms
-####################
-
-
-class ChatForm(BaseModel):
-    chat: dict
-
-
-class ChatTitleForm(BaseModel):
-    title: str
-
-
-class ChatResponse(BaseModel):
-    id: str
-    user_id: str
-    title: str
-    chat: dict
-    updated_at: int  # timestamp in epoch
-    created_at: int  # timestamp in epoch
-    share_id: Optional[str] = None  # id of the chat to be shared
-    archived: bool
-
-
-class ChatTitleIdResponse(BaseModel):
-    id: str
-    title: str
-    updated_at: int
-    created_at: int
-
-
-class ChatTable:
-
-    def insert_new_chat(self, user_id: str, form_data: ChatForm) -> Optional[ChatModel]:
-        with get_db() as db:
-
-            id = str(uuid.uuid4())
-            chat = ChatModel(
-                **{
-                    "id": id,
-                    "user_id": user_id,
-                    "title": (
-                        form_data.chat["title"]
-                        if "title" in form_data.chat
-                        else "New Chat"
-                    ),
-                    "chat": json.dumps(form_data.chat),
-                    "created_at": int(time.time()),
-                    "updated_at": int(time.time()),
-                }
-            )
-
-            result = Chat(**chat.model_dump())
-            db.add(result)
-            db.commit()
-            db.refresh(result)
-            return ChatModel.model_validate(result) if result else None
-
-    def update_chat_by_id(self, id: str, chat: dict) -> Optional[ChatModel]:
-        try:
-            with get_db() as db:
-
-                chat_obj = db.get(Chat, id)
-                chat_obj.chat = json.dumps(chat)
-                chat_obj.title = chat["title"] if "title" in chat else "New Chat"
-                chat_obj.updated_at = int(time.time())
-                db.commit()
-                db.refresh(chat_obj)
-
-                return ChatModel.model_validate(chat_obj)
-        except Exception as e:
-            return None
-
-    def insert_shared_chat_by_chat_id(self, chat_id: str) -> Optional[ChatModel]:
-        with get_db() as db:
-
-            # Get the existing chat to share
-            chat = db.get(Chat, chat_id)
-            # Check if the chat is already shared
-            if chat.share_id:
-                return self.get_chat_by_id_and_user_id(chat.share_id, "shared")
-            # Create a new chat with the same data, but with a new ID
-            shared_chat = ChatModel(
-                **{
-                    "id": str(uuid.uuid4()),
-                    "user_id": f"shared-{chat_id}",
-                    "title": chat.title,
-                    "chat": chat.chat,
-                    "created_at": chat.created_at,
-                    "updated_at": int(time.time()),
-                }
-            )
-            shared_result = Chat(**shared_chat.model_dump())
-            db.add(shared_result)
-            db.commit()
-            db.refresh(shared_result)
-
-            # Update the original chat with the share_id
-            result = (
-                db.query(Chat)
-                .filter_by(id=chat_id)
-                .update({"share_id": shared_chat.id})
-            )
-            db.commit()
-            return shared_chat if (shared_result and result) else None
-
-    def update_shared_chat_by_chat_id(self, chat_id: str) -> Optional[ChatModel]:
-        try:
-            with get_db() as db:
-
-                print("update_shared_chat_by_id")
-                chat = db.get(Chat, chat_id)
-                print(chat)
-                chat.title = chat.title
-                chat.chat = chat.chat
-                db.commit()
-                db.refresh(chat)
-
-                return self.get_chat_by_id(chat.share_id)
-        except:
-            return None
-
-    def delete_shared_chat_by_chat_id(self, chat_id: str) -> bool:
-        try:
-            with get_db() as db:
-
-                db.query(Chat).filter_by(user_id=f"shared-{chat_id}").delete()
-                db.commit()
-
-                return True
-        except:
-            return False
-
-    def update_chat_share_id_by_id(
-        self, id: str, share_id: Optional[str]
-    ) -> Optional[ChatModel]:
-        try:
-            with get_db() as db:
-
-                chat = db.get(Chat, id)
-                chat.share_id = share_id
-                db.commit()
-                db.refresh(chat)
-                return ChatModel.model_validate(chat)
-        except:
-            return None
-
-    def toggle_chat_archive_by_id(self, id: str) -> Optional[ChatModel]:
-        try:
-            with get_db() as db:
-
-                chat = db.get(Chat, id)
-                chat.archived = not chat.archived
-                db.commit()
-                db.refresh(chat)
-                return ChatModel.model_validate(chat)
-        except:
-            return None
-
-    def archive_all_chats_by_user_id(self, user_id: str) -> bool:
-        try:
-            with get_db() as db:
-                db.query(Chat).filter_by(user_id=user_id).update({"archived": True})
-                db.commit()
-                return True
-        except:
-            return False
-
-    def get_archived_chat_list_by_user_id(
-        self, user_id: str, skip: int = 0, limit: int = 50
-    ) -> List[ChatModel]:
-        with get_db() as db:
-
-            all_chats = (
-                db.query(Chat)
-                .filter_by(user_id=user_id, archived=True)
-                .order_by(Chat.updated_at.desc())
-                # .limit(limit).offset(skip)
-                .all()
-            )
-            return [ChatModel.model_validate(chat) for chat in all_chats]
-
-    def get_chat_list_by_user_id(
-        self,
-        user_id: str,
-        include_archived: bool = False,
-        skip: int = 0,
-        limit: int = 50,
-    ) -> List[ChatModel]:
-        with get_db() as db:
-            query = db.query(Chat).filter_by(user_id=user_id)
-            if not include_archived:
-                query = query.filter_by(archived=False)
-            all_chats = (
-                query.order_by(Chat.updated_at.desc())
-                # .limit(limit).offset(skip)
-                .all()
-            )
-            return [ChatModel.model_validate(chat) for chat in all_chats]
-
-    def get_chat_title_id_list_by_user_id(
-        self,
-        user_id: str,
-        include_archived: bool = False,
-        skip: int = 0,
-        limit: int = -1,
-    ) -> List[ChatTitleIdResponse]:
-        with get_db() as db:
-            query = db.query(Chat).filter_by(user_id=user_id)
-            if not include_archived:
-                query = query.filter_by(archived=False)
-
-            all_chats = (
-                query.order_by(Chat.updated_at.desc())
-                # limit cols
-                .with_entities(Chat.id, Chat.title, Chat.updated_at, Chat.created_at)
-                .limit(limit)
-                .offset(skip)
-                .all()
-            )
-            # result has to be destrctured from sqlalchemy `row` and mapped to a dict since the `ChatModel`is not the returned dataclass.
-            return [
-                ChatTitleIdResponse.model_validate(
-                    {
-                        "id": chat[0],
-                        "title": chat[1],
-                        "updated_at": chat[2],
-                        "created_at": chat[3],
-                    }
-                )
-                for chat in all_chats
-            ]
-
-    def get_chat_list_by_chat_ids(
-        self, chat_ids: List[str], skip: int = 0, limit: int = 50
-    ) -> List[ChatModel]:
-        with get_db() as db:
-            all_chats = (
-                db.query(Chat)
-                .filter(Chat.id.in_(chat_ids))
-                .filter_by(archived=False)
-                .order_by(Chat.updated_at.desc())
-                .all()
-            )
-            return [ChatModel.model_validate(chat) for chat in all_chats]
-
-    def get_chat_by_id(self, id: str) -> Optional[ChatModel]:
-        try:
-            with get_db() as db:
-
-                chat = db.get(Chat, id)
-                return ChatModel.model_validate(chat)
-        except:
-            return None
-
-    def get_chat_by_share_id(self, id: str) -> Optional[ChatModel]:
-        try:
-            with get_db() as db:
-
-                chat = db.query(Chat).filter_by(share_id=id).first()
-
-                if chat:
-                    return self.get_chat_by_id(id)
-                else:
-                    return None
-        except Exception as e:
-            return None
-
-    def get_chat_by_id_and_user_id(self, id: str, user_id: str) -> Optional[ChatModel]:
-        try:
-            with get_db() as db:
-
-                chat = db.query(Chat).filter_by(id=id, user_id=user_id).first()
-                return ChatModel.model_validate(chat)
-        except:
-            return None
-
-    def get_chats(self, skip: int = 0, limit: int = 50) -> List[ChatModel]:
-        with get_db() as db:
-
-            all_chats = (
-                db.query(Chat)
-                # .limit(limit).offset(skip)
-                .order_by(Chat.updated_at.desc())
-            )
-            return [ChatModel.model_validate(chat) for chat in all_chats]
-
-    def get_chats_by_user_id(self, user_id: str) -> List[ChatModel]:
-        with get_db() as db:
-
-            all_chats = (
-                db.query(Chat)
-                .filter_by(user_id=user_id)
-                .order_by(Chat.updated_at.desc())
-            )
-            return [ChatModel.model_validate(chat) for chat in all_chats]
-
-    def get_archived_chats_by_user_id(self, user_id: str) -> List[ChatModel]:
-        with get_db() as db:
-
-            all_chats = (
-                db.query(Chat)
-                .filter_by(user_id=user_id, archived=True)
-                .order_by(Chat.updated_at.desc())
-            )
-            return [ChatModel.model_validate(chat) for chat in all_chats]
-
-    def delete_chat_by_id(self, id: str) -> bool:
-        try:
-            with get_db() as db:
-
-                db.query(Chat).filter_by(id=id).delete()
-                db.commit()
-
-                return True and self.delete_shared_chat_by_chat_id(id)
-        except:
-            return False
-
-    def delete_chat_by_id_and_user_id(self, id: str, user_id: str) -> bool:
-        try:
-            with get_db() as db:
-
-                db.query(Chat).filter_by(id=id, user_id=user_id).delete()
-                db.commit()
-
-                return True and self.delete_shared_chat_by_chat_id(id)
-        except:
-            return False
-
-    def delete_chats_by_user_id(self, user_id: str) -> bool:
-        try:
-
-            with get_db() as db:
-
-                self.delete_shared_chats_by_user_id(user_id)
-
-                db.query(Chat).filter_by(user_id=user_id).delete()
-                db.commit()
-
-                return True
-        except:
-            return False
-
-    def delete_shared_chats_by_user_id(self, user_id: str) -> bool:
-        try:
-
-            with get_db() as db:
-
-                chats_by_user = db.query(Chat).filter_by(user_id=user_id).all()
-                shared_chat_ids = [f"shared-{chat.id}" for chat in chats_by_user]
-
-                db.query(Chat).filter(Chat.user_id.in_(shared_chat_ids)).delete()
-                db.commit()
-
-                return True
-        except:
-            return False
-
-
-Chats = ChatTable()

backend/apps/webui/models/documents.py

@@ -1,167 +0,0 @@
-from pydantic import BaseModel, ConfigDict
-from typing import List, Optional
-import time
-import logging
-
-from sqlalchemy import String, Column, BigInteger, Text
-
-from apps.webui.internal.db import Base, get_db
-
-import json
-
-from config import SRC_LOG_LEVELS
-
-log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["MODELS"])
-
-####################
-# Documents DB Schema
-####################
-
-
-class Document(Base):
-    __tablename__ = "document"
-
-    collection_name = Column(String, primary_key=True)
-    name = Column(String, unique=True)
-    title = Column(Text)
-    filename = Column(Text)
-    content = Column(Text, nullable=True)
-    user_id = Column(String)
-    timestamp = Column(BigInteger)
-
-
-class DocumentModel(BaseModel):
-    model_config = ConfigDict(from_attributes=True)
-
-    collection_name: str
-    name: str
-    title: str
-    filename: str
-    content: Optional[str] = None
-    user_id: str
-    timestamp: int  # timestamp in epoch
-
-
-####################
-# Forms
-####################
-
-
-class DocumentResponse(BaseModel):
-    collection_name: str
-    name: str
-    title: str
-    filename: str
-    content: Optional[dict] = None
-    user_id: str
-    timestamp: int  # timestamp in epoch
-
-
-class DocumentUpdateForm(BaseModel):
-    name: str
-    title: str
-
-
-class DocumentForm(DocumentUpdateForm):
-    collection_name: str
-    filename: str
-    content: Optional[str] = None
-
-
-class DocumentsTable:
-
-    def insert_new_doc(
-        self, user_id: str, form_data: DocumentForm
-    ) -> Optional[DocumentModel]:
-        with get_db() as db:
-
-            document = DocumentModel(
-                **{
-                    **form_data.model_dump(),
-                    "user_id": user_id,
-                    "timestamp": int(time.time()),
-                }
-            )
-
-            try:
-                result = Document(**document.model_dump())
-                db.add(result)
-                db.commit()
-                db.refresh(result)
-                if result:
-                    return DocumentModel.model_validate(result)
-                else:
-                    return None
-            except:
-                return None
-
-    def get_doc_by_name(self, name: str) -> Optional[DocumentModel]:
-        try:
-            with get_db() as db:
-
-                document = db.query(Document).filter_by(name=name).first()
-                return DocumentModel.model_validate(document) if document else None
-        except:
-            return None
-
-    def get_docs(self) -> List[DocumentModel]:
-        with get_db() as db:
-
-            return [
-                DocumentModel.model_validate(doc) for doc in db.query(Document).all()
-            ]
-
-    def update_doc_by_name(
-        self, name: str, form_data: DocumentUpdateForm
-    ) -> Optional[DocumentModel]:
-        try:
-            with get_db() as db:
-
-                db.query(Document).filter_by(name=name).update(
-                    {
-                        "title": form_data.title,
-                        "name": form_data.name,
-                        "timestamp": int(time.time()),
-                    }
-                )
-                db.commit()
-                return self.get_doc_by_name(form_data.name)
-        except Exception as e:
-            log.exception(e)
-            return None
-
-    def update_doc_content_by_name(
-        self, name: str, updated: dict
-    ) -> Optional[DocumentModel]:
-        try:
-            doc = self.get_doc_by_name(name)
-            doc_content = json.loads(doc.content if doc.content else "{}")
-            doc_content = {**doc_content, **updated}
-
-            with get_db() as db:
-
-                db.query(Document).filter_by(name=name).update(
-                    {
-                        "content": json.dumps(doc_content),
-                        "timestamp": int(time.time()),
-                    }
-                )
-                db.commit()
-                return self.get_doc_by_name(name)
-        except Exception as e:
-            log.exception(e)
-            return None
-
-    def delete_doc_by_name(self, name: str) -> bool:
-        try:
-            with get_db() as db:
-
-                db.query(Document).filter_by(name=name).delete()
-                db.commit()
-                return True
-        except:
-            return False
-
-
-Documents = DocumentsTable()

backend/apps/webui/models/files.py

@@ -1,126 +0,0 @@
-from pydantic import BaseModel, ConfigDict
-from typing import List, Union, Optional
-import time
-import logging
-
-from sqlalchemy import Column, String, BigInteger, Text
-
-from apps.webui.internal.db import JSONField, Base, get_db
-
-import json
-
-from config import SRC_LOG_LEVELS
-
-log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["MODELS"])
-
-####################
-# Files DB Schema
-####################
-
-
-class File(Base):
-    __tablename__ = "file"
-
-    id = Column(String, primary_key=True)
-    user_id = Column(String)
-    filename = Column(Text)
-    meta = Column(JSONField)
-    created_at = Column(BigInteger)
-
-
-class FileModel(BaseModel):
-    id: str
-    user_id: str
-    filename: str
-    meta: dict
-    created_at: int  # timestamp in epoch
-
-    model_config = ConfigDict(from_attributes=True)
-
-
-####################
-# Forms
-####################
-
-
-class FileModelResponse(BaseModel):
-    id: str
-    user_id: str
-    filename: str
-    meta: dict
-    created_at: int  # timestamp in epoch
-
-
-class FileForm(BaseModel):
-    id: str
-    filename: str
-    meta: dict = {}
-
-
-class FilesTable:
-
-    def insert_new_file(self, user_id: str, form_data: FileForm) -> Optional[FileModel]:
-        with get_db() as db:
-
-            file = FileModel(
-                **{
-                    **form_data.model_dump(),
-                    "user_id": user_id,
-                    "created_at": int(time.time()),
-                }
-            )
-
-            try:
-                result = File(**file.model_dump())
-                db.add(result)
-                db.commit()
-                db.refresh(result)
-                if result:
-                    return FileModel.model_validate(result)
-                else:
-                    return None
-            except Exception as e:
-                print(f"Error creating tool: {e}")
-                return None
-
-    def get_file_by_id(self, id: str) -> Optional[FileModel]:
-        with get_db() as db:
-
-            try:
-                file = db.get(File, id)
-                return FileModel.model_validate(file)
-            except:
-                return None
-
-    def get_files(self) -> List[FileModel]:
-        with get_db() as db:
-
-            return [FileModel.model_validate(file) for file in db.query(File).all()]
-
-    def delete_file_by_id(self, id: str) -> bool:
-
-        with get_db() as db:
-
-            try:
-                db.query(File).filter_by(id=id).delete()
-                db.commit()
-
-                return True
-            except:
-                return False
-
-    def delete_all_files(self) -> bool:
-
-        with get_db() as db:
-
-            try:
-                db.query(File).delete()
-                db.commit()
-
-                return True
-            except:
-                return False
-
-
-Files = FilesTable()

backend/apps/webui/models/tags.py

@@ -1,272 +0,0 @@
-from pydantic import BaseModel, ConfigDict
-from typing import List, Optional
-
-import json
-import uuid
-import time
-import logging
-
-from sqlalchemy import String, Column, BigInteger, Text
-
-from apps.webui.internal.db import Base, get_db
-
-from config import SRC_LOG_LEVELS
-
-log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["MODELS"])
-
-####################
-# Tag DB Schema
-####################
-
-
-class Tag(Base):
-    __tablename__ = "tag"
-
-    id = Column(String, primary_key=True)
-    name = Column(String)
-    user_id = Column(String)
-    data = Column(Text, nullable=True)
-
-
-class ChatIdTag(Base):
-    __tablename__ = "chatidtag"
-
-    id = Column(String, primary_key=True)
-    tag_name = Column(String)
-    chat_id = Column(String)
-    user_id = Column(String)
-    timestamp = Column(BigInteger)
-
-
-class TagModel(BaseModel):
-    id: str
-    name: str
-    user_id: str
-    data: Optional[str] = None
-
-    model_config = ConfigDict(from_attributes=True)
-
-
-class ChatIdTagModel(BaseModel):
-    id: str
-    tag_name: str
-    chat_id: str
-    user_id: str
-    timestamp: int
-
-    model_config = ConfigDict(from_attributes=True)
-
-
-####################
-# Forms
-####################
-
-
-class ChatIdTagForm(BaseModel):
-    tag_name: str
-    chat_id: str
-
-
-class TagChatIdsResponse(BaseModel):
-    chat_ids: List[str]
-
-
-class ChatTagsResponse(BaseModel):
-    tags: List[str]
-
-
-class TagTable:
-
-    def insert_new_tag(self, name: str, user_id: str) -> Optional[TagModel]:
-        with get_db() as db:
-
-            id = str(uuid.uuid4())
-            tag = TagModel(**{"id": id, "user_id": user_id, "name": name})
-            try:
-                result = Tag(**tag.model_dump())
-                db.add(result)
-                db.commit()
-                db.refresh(result)
-                if result:
-                    return TagModel.model_validate(result)
-                else:
-                    return None
-            except Exception as e:
-                return None
-
-    def get_tag_by_name_and_user_id(
-        self, name: str, user_id: str
-    ) -> Optional[TagModel]:
-        try:
-            with get_db() as db:
-                tag = db.query(Tag).filter_by(name=name, user_id=user_id).first()
-                return TagModel.model_validate(tag)
-        except Exception as e:
-            return None
-
-    def add_tag_to_chat(
-        self, user_id: str, form_data: ChatIdTagForm
-    ) -> Optional[ChatIdTagModel]:
-        tag = self.get_tag_by_name_and_user_id(form_data.tag_name, user_id)
-        if tag == None:
-            tag = self.insert_new_tag(form_data.tag_name, user_id)
-
-        id = str(uuid.uuid4())
-        chatIdTag = ChatIdTagModel(
-            **{
-                "id": id,
-                "user_id": user_id,
-                "chat_id": form_data.chat_id,
-                "tag_name": tag.name,
-                "timestamp": int(time.time()),
-            }
-        )
-        try:
-            with get_db() as db:
-                result = ChatIdTag(**chatIdTag.model_dump())
-                db.add(result)
-                db.commit()
-                db.refresh(result)
-                if result:
-                    return ChatIdTagModel.model_validate(result)
-                else:
-                    return None
-        except:
-            return None
-
-    def get_tags_by_user_id(self, user_id: str) -> List[TagModel]:
-        with get_db() as db:
-            tag_names = [
-                chat_id_tag.tag_name
-                for chat_id_tag in (
-                    db.query(ChatIdTag)
-                    .filter_by(user_id=user_id)
-                    .order_by(ChatIdTag.timestamp.desc())
-                    .all()
-                )
-            ]
-
-            return [
-                TagModel.model_validate(tag)
-                for tag in (
-                    db.query(Tag)
-                    .filter_by(user_id=user_id)
-                    .filter(Tag.name.in_(tag_names))
-                    .all()
-                )
-            ]
-
-    def get_tags_by_chat_id_and_user_id(
-        self, chat_id: str, user_id: str
-    ) -> List[TagModel]:
-        with get_db() as db:
-
-            tag_names = [
-                chat_id_tag.tag_name
-                for chat_id_tag in (
-                    db.query(ChatIdTag)
-                    .filter_by(user_id=user_id, chat_id=chat_id)
-                    .order_by(ChatIdTag.timestamp.desc())
-                    .all()
-                )
-            ]
-
-            return [
-                TagModel.model_validate(tag)
-                for tag in (
-                    db.query(Tag)
-                    .filter_by(user_id=user_id)
-                    .filter(Tag.name.in_(tag_names))
-                    .all()
-                )
-            ]
-
-    def get_chat_ids_by_tag_name_and_user_id(
-        self, tag_name: str, user_id: str
-    ) -> List[ChatIdTagModel]:
-        with get_db() as db:
-
-            return [
-                ChatIdTagModel.model_validate(chat_id_tag)
-                for chat_id_tag in (
-                    db.query(ChatIdTag)
-                    .filter_by(user_id=user_id, tag_name=tag_name)
-                    .order_by(ChatIdTag.timestamp.desc())
-                    .all()
-                )
-            ]
-
-    def count_chat_ids_by_tag_name_and_user_id(
-        self, tag_name: str, user_id: str
-    ) -> int:
-        with get_db() as db:
-
-            return (
-                db.query(ChatIdTag)
-                .filter_by(tag_name=tag_name, user_id=user_id)
-                .count()
-            )
-
-    def delete_tag_by_tag_name_and_user_id(self, tag_name: str, user_id: str) -> bool:
-        try:
-            with get_db() as db:
-                res = (
-                    db.query(ChatIdTag)
-                    .filter_by(tag_name=tag_name, user_id=user_id)
-                    .delete()
-                )
-                log.debug(f"res: {res}")
-                db.commit()
-
-                tag_count = self.count_chat_ids_by_tag_name_and_user_id(
-                    tag_name, user_id
-                )
-                if tag_count == 0:
-                    # Remove tag item from Tag col as well
-                    db.query(Tag).filter_by(name=tag_name, user_id=user_id).delete()
-                    db.commit()
-                return True
-        except Exception as e:
-            log.error(f"delete_tag: {e}")
-            return False
-
-    def delete_tag_by_tag_name_and_chat_id_and_user_id(
-        self, tag_name: str, chat_id: str, user_id: str
-    ) -> bool:
-        try:
-            with get_db() as db:
-
-                res = (
-                    db.query(ChatIdTag)
-                    .filter_by(tag_name=tag_name, chat_id=chat_id, user_id=user_id)
-                    .delete()
-                )
-                log.debug(f"res: {res}")
-                db.commit()
-
-                tag_count = self.count_chat_ids_by_tag_name_and_user_id(
-                    tag_name, user_id
-                )
-                if tag_count == 0:
-                    # Remove tag item from Tag col as well
-                    db.query(Tag).filter_by(name=tag_name, user_id=user_id).delete()
-                    db.commit()
-
-                return True
-        except Exception as e:
-            log.error(f"delete_tag: {e}")
-            return False
-
-    def delete_tags_by_chat_id_and_user_id(self, chat_id: str, user_id: str) -> bool:
-        tags = self.get_tags_by_chat_id_and_user_id(chat_id, user_id)
-
-        for tag in tags:
-            self.delete_tag_by_tag_name_and_chat_id_and_user_id(
-                tag.tag_name, chat_id, user_id
-            )
-
-        return True
-
-
-Tags = TagTable()

backend/apps/webui/routers/auths.py

@@ -1,429 +0,0 @@
-import logging
-
-from fastapi import Request, UploadFile, File
-from fastapi import Depends, HTTPException, status
-from fastapi.responses import Response
-
-from fastapi import APIRouter
-from pydantic import BaseModel
-import re
-import uuid
-import csv
-
-from apps.webui.models.auths import (
-    SigninForm,
-    SignupForm,
-    AddUserForm,
-    UpdateProfileForm,
-    UpdatePasswordForm,
-    UserResponse,
-    SigninResponse,
-    Auths,
-    ApiKey,
-)
-from apps.webui.models.users import Users
-
-from utils.utils import (
-    get_password_hash,
-    get_current_user,
-    get_admin_user,
-    create_token,
-    create_api_key,
-)
-from utils.misc import parse_duration, validate_email_format
-from utils.webhook import post_webhook
-from constants import ERROR_MESSAGES, WEBHOOK_MESSAGES
-from config import (
-    WEBUI_AUTH,
-    WEBUI_AUTH_TRUSTED_EMAIL_HEADER,
-    WEBUI_AUTH_TRUSTED_NAME_HEADER,
-)
-
-router = APIRouter()
-
-############################
-# GetSessionUser
-############################
-
-
-@router.get("/", response_model=UserResponse)
-async def get_session_user(
-    request: Request, response: Response, user=Depends(get_current_user)
-):
-    token = create_token(
-        data={"id": user.id},
-        expires_delta=parse_duration(request.app.state.config.JWT_EXPIRES_IN),
-    )
-
-    # Set the cookie token
-    response.set_cookie(
-        key="token",
-        value=token,
-        httponly=True,  # Ensures the cookie is not accessible via JavaScript
-    )
-
-    return {
-        "id": user.id,
-        "email": user.email,
-        "name": user.name,
-        "role": user.role,
-        "profile_image_url": user.profile_image_url,
-    }
-
-
-############################
-# Update Profile
-############################
-
-
-@router.post("/update/profile", response_model=UserResponse)
-async def update_profile(
-    form_data: UpdateProfileForm, session_user=Depends(get_current_user)
-):
-    if session_user:
-        user = Users.update_user_by_id(
-            session_user.id,
-            {"profile_image_url": form_data.profile_image_url, "name": form_data.name},
-        )
-        if user:
-            return user
-        else:
-            raise HTTPException(400, detail=ERROR_MESSAGES.DEFAULT())
-    else:
-        raise HTTPException(400, detail=ERROR_MESSAGES.INVALID_CRED)
-
-
-############################
-# Update Password
-############################
-
-
-@router.post("/update/password", response_model=bool)
-async def update_password(
-    form_data: UpdatePasswordForm, session_user=Depends(get_current_user)
-):
-    if WEBUI_AUTH_TRUSTED_EMAIL_HEADER:
-        raise HTTPException(400, detail=ERROR_MESSAGES.ACTION_PROHIBITED)
-    if session_user:
-        user = Auths.authenticate_user(session_user.email, form_data.password)
-
-        if user:
-            hashed = get_password_hash(form_data.new_password)
-            return Auths.update_user_password_by_id(user.id, hashed)
-        else:
-            raise HTTPException(400, detail=ERROR_MESSAGES.INVALID_PASSWORD)
-    else:
-        raise HTTPException(400, detail=ERROR_MESSAGES.INVALID_CRED)
-
-
-############################
-# SignIn
-############################
-
-
-@router.post("/signin", response_model=SigninResponse)
-async def signin(request: Request, response: Response, form_data: SigninForm):
-    if WEBUI_AUTH_TRUSTED_EMAIL_HEADER:
-        if WEBUI_AUTH_TRUSTED_EMAIL_HEADER not in request.headers:
-            raise HTTPException(400, detail=ERROR_MESSAGES.INVALID_TRUSTED_HEADER)
-
-        trusted_email = request.headers[WEBUI_AUTH_TRUSTED_EMAIL_HEADER].lower()
-        trusted_name = trusted_email
-        if WEBUI_AUTH_TRUSTED_NAME_HEADER:
-            trusted_name = request.headers.get(
-                WEBUI_AUTH_TRUSTED_NAME_HEADER, trusted_email
-            )
-        if not Users.get_user_by_email(trusted_email.lower()):
-            await signup(
-                request,
-                response,
-                SignupForm(
-                    email=trusted_email, password=str(uuid.uuid4()), name=trusted_name
-                ),
-            )
-        user = Auths.authenticate_user_by_trusted_header(trusted_email)
-    elif WEBUI_AUTH == False:
-        admin_email = "admin@localhost"
-        admin_password = "admin"
-
-        if Users.get_user_by_email(admin_email.lower()):
-            user = Auths.authenticate_user(admin_email.lower(), admin_password)
-        else:
-            if Users.get_num_users() != 0:
-                raise HTTPException(400, detail=ERROR_MESSAGES.EXISTING_USERS)
-
-            await signup(
-                request,
-                response,
-                SignupForm(email=admin_email, password=admin_password, name="User"),
-            )
-
-            user = Auths.authenticate_user(admin_email.lower(), admin_password)
-    else:
-        user = Auths.authenticate_user(form_data.email.lower(), form_data.password)
-
-    if user:
-        token = create_token(
-            data={"id": user.id},
-            expires_delta=parse_duration(request.app.state.config.JWT_EXPIRES_IN),
-        )
-
-        # Set the cookie token
-        response.set_cookie(
-            key="token",
-            value=token,
-            httponly=True,  # Ensures the cookie is not accessible via JavaScript
-        )
-
-        return {
-            "token": token,
-            "token_type": "Bearer",
-            "id": user.id,
-            "email": user.email,
-            "name": user.name,
-            "role": user.role,
-            "profile_image_url": user.profile_image_url,
-        }
-    else:
-        raise HTTPException(400, detail=ERROR_MESSAGES.INVALID_CRED)
-
-
-############################
-# SignUp
-############################
-
-
-@router.post("/signup", response_model=SigninResponse)
-async def signup(request: Request, response: Response, form_data: SignupForm):
-    if not request.app.state.config.ENABLE_SIGNUP and WEBUI_AUTH:
-        raise HTTPException(
-            status.HTTP_403_FORBIDDEN, detail=ERROR_MESSAGES.ACCESS_PROHIBITED
-        )
-
-    if not validate_email_format(form_data.email.lower()):
-        raise HTTPException(
-            status.HTTP_400_BAD_REQUEST, detail=ERROR_MESSAGES.INVALID_EMAIL_FORMAT
-        )
-
-    if Users.get_user_by_email(form_data.email.lower()):
-        raise HTTPException(400, detail=ERROR_MESSAGES.EMAIL_TAKEN)
-
-    try:
-        role = (
-            "admin"
-            if Users.get_num_users() == 0
-            else request.app.state.config.DEFAULT_USER_ROLE
-        )
-        hashed = get_password_hash(form_data.password)
-        user = Auths.insert_new_auth(
-            form_data.email.lower(),
-            hashed,
-            form_data.name,
-            form_data.profile_image_url,
-            role,
-        )
-
-        if user:
-            token = create_token(
-                data={"id": user.id},
-                expires_delta=parse_duration(request.app.state.config.JWT_EXPIRES_IN),
-            )
-            # response.set_cookie(key='token', value=token, httponly=True)
-
-            # Set the cookie token
-            response.set_cookie(
-                key="token",
-                value=token,
-                httponly=True,  # Ensures the cookie is not accessible via JavaScript
-            )
-
-            if request.app.state.config.WEBHOOK_URL:
-                post_webhook(
-                    request.app.state.config.WEBHOOK_URL,
-                    WEBHOOK_MESSAGES.USER_SIGNUP(user.name),
-                    {
-                        "action": "signup",
-                        "message": WEBHOOK_MESSAGES.USER_SIGNUP(user.name),
-                        "user": user.model_dump_json(exclude_none=True),
-                    },
-                )
-
-            return {
-                "token": token,
-                "token_type": "Bearer",
-                "id": user.id,
-                "email": user.email,
-                "name": user.name,
-                "role": user.role,
-                "profile_image_url": user.profile_image_url,
-            }
-        else:
-            raise HTTPException(500, detail=ERROR_MESSAGES.CREATE_USER_ERROR)
-    except Exception as err:
-        raise HTTPException(500, detail=ERROR_MESSAGES.DEFAULT(err))
-
-
-############################
-# AddUser
-############################
-
-
-@router.post("/add", response_model=SigninResponse)
-async def add_user(form_data: AddUserForm, user=Depends(get_admin_user)):
-
-    if not validate_email_format(form_data.email.lower()):
-        raise HTTPException(
-            status.HTTP_400_BAD_REQUEST, detail=ERROR_MESSAGES.INVALID_EMAIL_FORMAT
-        )
-
-    if Users.get_user_by_email(form_data.email.lower()):
-        raise HTTPException(400, detail=ERROR_MESSAGES.EMAIL_TAKEN)
-
-    try:
-
-        print(form_data)
-        hashed = get_password_hash(form_data.password)
-        user = Auths.insert_new_auth(
-            form_data.email.lower(),
-            hashed,
-            form_data.name,
-            form_data.profile_image_url,
-            form_data.role,
-        )
-
-        if user:
-            token = create_token(data={"id": user.id})
-            return {
-                "token": token,
-                "token_type": "Bearer",
-                "id": user.id,
-                "email": user.email,
-                "name": user.name,
-                "role": user.role,
-                "profile_image_url": user.profile_image_url,
-            }
-        else:
-            raise HTTPException(500, detail=ERROR_MESSAGES.CREATE_USER_ERROR)
-    except Exception as err:
-        raise HTTPException(500, detail=ERROR_MESSAGES.DEFAULT(err))
-
-
-############################
-# GetAdminDetails
-############################
-
-
-@router.get("/admin/details")
-async def get_admin_details(request: Request, user=Depends(get_current_user)):
-    if request.app.state.config.SHOW_ADMIN_DETAILS:
-        admin_email = request.app.state.config.ADMIN_EMAIL
-        admin_name = None
-
-        print(admin_email, admin_name)
-
-        if admin_email:
-            admin = Users.get_user_by_email(admin_email)
-            if admin:
-                admin_name = admin.name
-        else:
-            admin = Users.get_first_user()
-            if admin:
-                admin_email = admin.email
-                admin_name = admin.name
-
-        return {
-            "name": admin_name,
-            "email": admin_email,
-        }
-    else:
-        raise HTTPException(400, detail=ERROR_MESSAGES.ACTION_PROHIBITED)
-
-
-############################
-# ToggleSignUp
-############################
-
-
-@router.get("/admin/config")
-async def get_admin_config(request: Request, user=Depends(get_admin_user)):
-    return {
-        "SHOW_ADMIN_DETAILS": request.app.state.config.SHOW_ADMIN_DETAILS,
-        "ENABLE_SIGNUP": request.app.state.config.ENABLE_SIGNUP,
-        "DEFAULT_USER_ROLE": request.app.state.config.DEFAULT_USER_ROLE,
-        "JWT_EXPIRES_IN": request.app.state.config.JWT_EXPIRES_IN,
-        "ENABLE_COMMUNITY_SHARING": request.app.state.config.ENABLE_COMMUNITY_SHARING,
-    }
-
-
-class AdminConfig(BaseModel):
-    SHOW_ADMIN_DETAILS: bool
-    ENABLE_SIGNUP: bool
-    DEFAULT_USER_ROLE: str
-    JWT_EXPIRES_IN: str
-    ENABLE_COMMUNITY_SHARING: bool
-
-
-@router.post("/admin/config")
-async def update_admin_config(
-    request: Request, form_data: AdminConfig, user=Depends(get_admin_user)
-):
-    request.app.state.config.SHOW_ADMIN_DETAILS = form_data.SHOW_ADMIN_DETAILS
-    request.app.state.config.ENABLE_SIGNUP = form_data.ENABLE_SIGNUP
-
-    if form_data.DEFAULT_USER_ROLE in ["pending", "user", "admin"]:
-        request.app.state.config.DEFAULT_USER_ROLE = form_data.DEFAULT_USER_ROLE
-
-    pattern = r"^(-1|0|(-?\d+(\.\d+)?)(ms|s|m|h|d|w))$"
-
-    # Check if the input string matches the pattern
-    if re.match(pattern, form_data.JWT_EXPIRES_IN):
-        request.app.state.config.JWT_EXPIRES_IN = form_data.JWT_EXPIRES_IN
-
-    request.app.state.config.ENABLE_COMMUNITY_SHARING = (
-        form_data.ENABLE_COMMUNITY_SHARING
-    )
-
-    return {
-        "SHOW_ADMIN_DETAILS": request.app.state.config.SHOW_ADMIN_DETAILS,
-        "ENABLE_SIGNUP": request.app.state.config.ENABLE_SIGNUP,
-        "DEFAULT_USER_ROLE": request.app.state.config.DEFAULT_USER_ROLE,
-        "JWT_EXPIRES_IN": request.app.state.config.JWT_EXPIRES_IN,
-        "ENABLE_COMMUNITY_SHARING": request.app.state.config.ENABLE_COMMUNITY_SHARING,
-    }
-
-
-############################
-# API Key
-############################
-
-
-# create api key
-@router.post("/api_key", response_model=ApiKey)
-async def create_api_key_(user=Depends(get_current_user)):
-    api_key = create_api_key()
-    success = Users.update_user_api_key_by_id(user.id, api_key)
-    if success:
-        return {
-            "api_key": api_key,
-        }
-    else:
-        raise HTTPException(500, detail=ERROR_MESSAGES.CREATE_API_KEY_ERROR)
-
-
-# delete api key
-@router.delete("/api_key", response_model=bool)
-async def delete_api_key(user=Depends(get_current_user)):
-    success = Users.update_user_api_key_by_id(user.id, None)
-    return success
-
-
-# get api key
-@router.get("/api_key", response_model=ApiKey)
-async def get_api_key(user=Depends(get_current_user)):
-    api_key = Users.get_user_api_key_by_id(user.id)
-    if api_key:
-        return {
-            "api_key": api_key,
-        }
-    else:
-        raise HTTPException(404, detail=ERROR_MESSAGES.API_KEY_NOT_FOUND)

backend/apps/webui/routers/chats.py

@@ -1,494 +0,0 @@
-from fastapi import Depends, Request, HTTPException, status
-from datetime import datetime, timedelta
-from typing import List, Union, Optional
-from utils.utils import get_verified_user, get_admin_user
-from fastapi import APIRouter
-from pydantic import BaseModel
-import json
-import logging
-
-from apps.webui.models.users import Users
-from apps.webui.models.chats import (
-    ChatModel,
-    ChatResponse,
-    ChatTitleForm,
-    ChatForm,
-    ChatTitleIdResponse,
-    Chats,
-)
-
-
-from apps.webui.models.tags import (
-    TagModel,
-    ChatIdTagModel,
-    ChatIdTagForm,
-    ChatTagsResponse,
-    Tags,
-)
-
-from constants import ERROR_MESSAGES
-
-from config import SRC_LOG_LEVELS, ENABLE_ADMIN_EXPORT, ENABLE_ADMIN_CHAT_ACCESS
-
-log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["MODELS"])
-
-router = APIRouter()
-
-############################
-# GetChatList
-############################
-
-
-@router.get("/", response_model=List[ChatTitleIdResponse])
-@router.get("/list", response_model=List[ChatTitleIdResponse])
-async def get_session_user_chat_list(
-    user=Depends(get_verified_user), page: Optional[int] = None
-):
-    if page is not None:
-        limit = 60
-        skip = (page - 1) * limit
-
-        return Chats.get_chat_title_id_list_by_user_id(user.id, skip=skip, limit=limit)
-    else:
-        return Chats.get_chat_title_id_list_by_user_id(user.id)
-
-
-############################
-# DeleteAllChats
-############################
-
-
-@router.delete("/", response_model=bool)
-async def delete_all_user_chats(request: Request, user=Depends(get_verified_user)):
-
-    if (
-        user.role == "user"
-        and not request.app.state.config.USER_PERMISSIONS["chat"]["deletion"]
-    ):
-        raise HTTPException(
-            status_code=status.HTTP_401_UNAUTHORIZED,
-            detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
-        )
-
-    result = Chats.delete_chats_by_user_id(user.id)
-    return result
-
-
-############################
-# GetUserChatList
-############################
-
-
-@router.get("/list/user/{user_id}", response_model=List[ChatTitleIdResponse])
-async def get_user_chat_list_by_user_id(
-    user_id: str,
-    user=Depends(get_admin_user),
-    skip: int = 0,
-    limit: int = 50,
-):
-    if not ENABLE_ADMIN_CHAT_ACCESS:
-        raise HTTPException(
-            status_code=status.HTTP_401_UNAUTHORIZED,
-            detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
-        )
-    return Chats.get_chat_list_by_user_id(
-        user_id, include_archived=True, skip=skip, limit=limit
-    )
-
-
-############################
-# CreateNewChat
-############################
-
-
-@router.post("/new", response_model=Optional[ChatResponse])
-async def create_new_chat(form_data: ChatForm, user=Depends(get_verified_user)):
-    try:
-        chat = Chats.insert_new_chat(user.id, form_data)
-        return ChatResponse(**{**chat.model_dump(), "chat": json.loads(chat.chat)})
-    except Exception as e:
-        log.exception(e)
-        raise HTTPException(
-            status_code=status.HTTP_400_BAD_REQUEST, detail=ERROR_MESSAGES.DEFAULT()
-        )
-
-
-############################
-# GetChats
-############################
-
-
-@router.get("/all", response_model=List[ChatResponse])
-async def get_user_chats(user=Depends(get_verified_user)):
-    return [
-        ChatResponse(**{**chat.model_dump(), "chat": json.loads(chat.chat)})
-        for chat in Chats.get_chats_by_user_id(user.id)
-    ]
-
-
-############################
-# GetArchivedChats
-############################
-
-
-@router.get("/all/archived", response_model=List[ChatResponse])
-async def get_user_archived_chats(user=Depends(get_verified_user)):
-    return [
-        ChatResponse(**{**chat.model_dump(), "chat": json.loads(chat.chat)})
-        for chat in Chats.get_archived_chats_by_user_id(user.id)
-    ]
-
-
-############################
-# GetAllChatsInDB
-############################
-
-
-@router.get("/all/db", response_model=List[ChatResponse])
-async def get_all_user_chats_in_db(user=Depends(get_admin_user)):
-    if not ENABLE_ADMIN_EXPORT:
-        raise HTTPException(
-            status_code=status.HTTP_401_UNAUTHORIZED,
-            detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
-        )
-    return [
-        ChatResponse(**{**chat.model_dump(), "chat": json.loads(chat.chat)})
-        for chat in Chats.get_chats()
-    ]
-
-
-############################
-# GetArchivedChats
-############################
-
-
-@router.get("/archived", response_model=List[ChatTitleIdResponse])
-async def get_archived_session_user_chat_list(
-    user=Depends(get_verified_user), skip: int = 0, limit: int = 50
-):
-    return Chats.get_archived_chat_list_by_user_id(user.id, skip, limit)
-
-
-############################
-# ArchiveAllChats
-############################
-
-
-@router.post("/archive/all", response_model=bool)
-async def archive_all_chats(user=Depends(get_verified_user)):
-    return Chats.archive_all_chats_by_user_id(user.id)
-
-
-############################
-# GetSharedChatById
-############################
-
-
-@router.get("/share/{share_id}", response_model=Optional[ChatResponse])
-async def get_shared_chat_by_id(share_id: str, user=Depends(get_verified_user)):
-    if user.role == "pending":
-        raise HTTPException(
-            status_code=status.HTTP_401_UNAUTHORIZED, detail=ERROR_MESSAGES.NOT_FOUND
-        )
-
-    if user.role == "user" or (user.role == "admin" and not ENABLE_ADMIN_CHAT_ACCESS):
-        chat = Chats.get_chat_by_share_id(share_id)
-    elif user.role == "admin" and ENABLE_ADMIN_CHAT_ACCESS:
-        chat = Chats.get_chat_by_id(share_id)
-
-    if chat:
-        return ChatResponse(**{**chat.model_dump(), "chat": json.loads(chat.chat)})
-    else:
-        raise HTTPException(
-            status_code=status.HTTP_401_UNAUTHORIZED, detail=ERROR_MESSAGES.NOT_FOUND
-        )
-
-
-############################
-# GetChatsByTags
-############################
-
-
-class TagNameForm(BaseModel):
-    name: str
-    skip: Optional[int] = 0
-    limit: Optional[int] = 50
-
-
-@router.post("/tags", response_model=List[ChatTitleIdResponse])
-async def get_user_chat_list_by_tag_name(
-    form_data: TagNameForm, user=Depends(get_verified_user)
-):
-
-    chat_ids = [
-        chat_id_tag.chat_id
-        for chat_id_tag in Tags.get_chat_ids_by_tag_name_and_user_id(
-            form_data.name, user.id
-        )
-    ]
-
-    chats = Chats.get_chat_list_by_chat_ids(chat_ids, form_data.skip, form_data.limit)
-
-    if len(chats) == 0:
-        Tags.delete_tag_by_tag_name_and_user_id(form_data.name, user.id)
-
-    return chats
-
-
-############################
-# GetAllTags
-############################
-
-
-@router.get("/tags/all", response_model=List[TagModel])
-async def get_all_tags(user=Depends(get_verified_user)):
-    try:
-        tags = Tags.get_tags_by_user_id(user.id)
-        return tags
-    except Exception as e:
-        log.exception(e)
-        raise HTTPException(
-            status_code=status.HTTP_400_BAD_REQUEST, detail=ERROR_MESSAGES.DEFAULT()
-        )
-
-
-############################
-# GetChatById
-############################
-
-
-@router.get("/{id}", response_model=Optional[ChatResponse])
-async def get_chat_by_id(id: str, user=Depends(get_verified_user)):
-    chat = Chats.get_chat_by_id_and_user_id(id, user.id)
-
-    if chat:
-        return ChatResponse(**{**chat.model_dump(), "chat": json.loads(chat.chat)})
-    else:
-        raise HTTPException(
-            status_code=status.HTTP_401_UNAUTHORIZED, detail=ERROR_MESSAGES.NOT_FOUND
-        )
-
-
-############################
-# UpdateChatById
-############################
-
-
-@router.post("/{id}", response_model=Optional[ChatResponse])
-async def update_chat_by_id(
-    id: str, form_data: ChatForm, user=Depends(get_verified_user)
-):
-    chat = Chats.get_chat_by_id_and_user_id(id, user.id)
-    if chat:
-        updated_chat = {**json.loads(chat.chat), **form_data.chat}
-
-        chat = Chats.update_chat_by_id(id, updated_chat)
-        return ChatResponse(**{**chat.model_dump(), "chat": json.loads(chat.chat)})
-    else:
-        raise HTTPException(
-            status_code=status.HTTP_401_UNAUTHORIZED,
-            detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
-        )
-
-
-############################
-# DeleteChatById
-############################
-
-
-@router.delete("/{id}", response_model=bool)
-async def delete_chat_by_id(request: Request, id: str, user=Depends(get_verified_user)):
-
-    if user.role == "admin":
-        result = Chats.delete_chat_by_id(id)
-        return result
-    else:
-        if not request.app.state.config.USER_PERMISSIONS["chat"]["deletion"]:
-            raise HTTPException(
-                status_code=status.HTTP_401_UNAUTHORIZED,
-                detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
-            )
-
-        result = Chats.delete_chat_by_id_and_user_id(id, user.id)
-        return result
-
-
-############################
-# CloneChat
-############################
-
-
-@router.get("/{id}/clone", response_model=Optional[ChatResponse])
-async def clone_chat_by_id(id: str, user=Depends(get_verified_user)):
-    chat = Chats.get_chat_by_id_and_user_id(id, user.id)
-    if chat:
-
-        chat_body = json.loads(chat.chat)
-        updated_chat = {
-            **chat_body,
-            "originalChatId": chat.id,
-            "branchPointMessageId": chat_body["history"]["currentId"],
-            "title": f"Clone of {chat.title}",
-        }
-
-        chat = Chats.insert_new_chat(user.id, ChatForm(**{"chat": updated_chat}))
-        return ChatResponse(**{**chat.model_dump(), "chat": json.loads(chat.chat)})
-    else:
-        raise HTTPException(
-            status_code=status.HTTP_401_UNAUTHORIZED, detail=ERROR_MESSAGES.DEFAULT()
-        )
-
-
-############################
-# ArchiveChat
-############################
-
-
-@router.get("/{id}/archive", response_model=Optional[ChatResponse])
-async def archive_chat_by_id(id: str, user=Depends(get_verified_user)):
-    chat = Chats.get_chat_by_id_and_user_id(id, user.id)
-    if chat:
-        chat = Chats.toggle_chat_archive_by_id(id)
-        return ChatResponse(**{**chat.model_dump(), "chat": json.loads(chat.chat)})
-    else:
-        raise HTTPException(
-            status_code=status.HTTP_401_UNAUTHORIZED, detail=ERROR_MESSAGES.DEFAULT()
-        )
-
-
-############################
-# ShareChatById
-############################
-
-
-@router.post("/{id}/share", response_model=Optional[ChatResponse])
-async def share_chat_by_id(id: str, user=Depends(get_verified_user)):
-    chat = Chats.get_chat_by_id_and_user_id(id, user.id)
-    if chat:
-        if chat.share_id:
-            shared_chat = Chats.update_shared_chat_by_chat_id(chat.id)
-            return ChatResponse(
-                **{**shared_chat.model_dump(), "chat": json.loads(shared_chat.chat)}
-            )
-
-        shared_chat = Chats.insert_shared_chat_by_chat_id(chat.id)
-        if not shared_chat:
-            raise HTTPException(
-                status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
-                detail=ERROR_MESSAGES.DEFAULT(),
-            )
-
-        return ChatResponse(
-            **{**shared_chat.model_dump(), "chat": json.loads(shared_chat.chat)}
-        )
-    else:
-        raise HTTPException(
-            status_code=status.HTTP_401_UNAUTHORIZED,
-            detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
-        )
-
-
-############################
-# DeletedSharedChatById
-############################
-
-
-@router.delete("/{id}/share", response_model=Optional[bool])
-async def delete_shared_chat_by_id(id: str, user=Depends(get_verified_user)):
-    chat = Chats.get_chat_by_id_and_user_id(id, user.id)
-    if chat:
-        if not chat.share_id:
-            return False
-
-        result = Chats.delete_shared_chat_by_chat_id(id)
-        update_result = Chats.update_chat_share_id_by_id(id, None)
-
-        return result and update_result != None
-    else:
-        raise HTTPException(
-            status_code=status.HTTP_401_UNAUTHORIZED,
-            detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
-        )
-
-
-############################
-# GetChatTagsById
-############################
-
-
-@router.get("/{id}/tags", response_model=List[TagModel])
-async def get_chat_tags_by_id(id: str, user=Depends(get_verified_user)):
-    tags = Tags.get_tags_by_chat_id_and_user_id(id, user.id)
-
-    if tags != None:
-        return tags
-    else:
-        raise HTTPException(
-            status_code=status.HTTP_401_UNAUTHORIZED, detail=ERROR_MESSAGES.NOT_FOUND
-        )
-
-
-############################
-# AddChatTagById
-############################
-
-
-@router.post("/{id}/tags", response_model=Optional[ChatIdTagModel])
-async def add_chat_tag_by_id(
-    id: str, form_data: ChatIdTagForm, user=Depends(get_verified_user)
-):
-    tags = Tags.get_tags_by_chat_id_and_user_id(id, user.id)
-
-    if form_data.tag_name not in tags:
-        tag = Tags.add_tag_to_chat(user.id, form_data)
-
-        if tag:
-            return tag
-        else:
-            raise HTTPException(
-                status_code=status.HTTP_401_UNAUTHORIZED,
-                detail=ERROR_MESSAGES.NOT_FOUND,
-            )
-    else:
-        raise HTTPException(
-            status_code=status.HTTP_401_UNAUTHORIZED, detail=ERROR_MESSAGES.DEFAULT()
-        )
-
-
-############################
-# DeleteChatTagById
-############################
-
-
-@router.delete("/{id}/tags", response_model=Optional[bool])
-async def delete_chat_tag_by_id(
-    id: str, form_data: ChatIdTagForm, user=Depends(get_verified_user)
-):
-    result = Tags.delete_tag_by_tag_name_and_chat_id_and_user_id(
-        form_data.tag_name, id, user.id
-    )
-
-    if result:
-        return result
-    else:
-        raise HTTPException(
-            status_code=status.HTTP_401_UNAUTHORIZED, detail=ERROR_MESSAGES.NOT_FOUND
-        )
-
-
-############################
-# DeleteAllChatTagsById
-############################
-
-
-@router.delete("/{id}/tags/all", response_model=Optional[bool])
-async def delete_all_chat_tags_by_id(id: str, user=Depends(get_verified_user)):
-    result = Tags.delete_tags_by_chat_id_and_user_id(id, user.id)
-
-    if result:
-        return result
-    else:
-        raise HTTPException(
-            status_code=status.HTTP_401_UNAUTHORIZED, detail=ERROR_MESSAGES.NOT_FOUND
-        )

backend/apps/webui/routers/configs.py

@@ -1,89 +0,0 @@
-from fastapi import Response, Request
-from fastapi import Depends, FastAPI, HTTPException, status
-from datetime import datetime, timedelta
-from typing import List, Union
-
-from fastapi import APIRouter
-from pydantic import BaseModel
-import time
-import uuid
-
-from config import BannerModel
-
-from apps.webui.models.users import Users
-
-from utils.utils import (
-    get_password_hash,
-    get_verified_user,
-    get_admin_user,
-    create_token,
-)
-from utils.misc import get_gravatar_url, validate_email_format
-from constants import ERROR_MESSAGES
-
-router = APIRouter()
-
-
-class SetDefaultModelsForm(BaseModel):
-    models: str
-
-
-class PromptSuggestion(BaseModel):
-    title: List[str]
-    content: str
-
-
-class SetDefaultSuggestionsForm(BaseModel):
-    suggestions: List[PromptSuggestion]
-
-
-############################
-# SetDefaultModels
-############################
-
-
-@router.post("/default/models", response_model=str)
-async def set_global_default_models(
-    request: Request, form_data: SetDefaultModelsForm, user=Depends(get_admin_user)
-):
-    request.app.state.config.DEFAULT_MODELS = form_data.models
-    return request.app.state.config.DEFAULT_MODELS
-
-
-@router.post("/default/suggestions", response_model=List[PromptSuggestion])
-async def set_global_default_suggestions(
-    request: Request,
-    form_data: SetDefaultSuggestionsForm,
-    user=Depends(get_admin_user),
-):
-    data = form_data.model_dump()
-    request.app.state.config.DEFAULT_PROMPT_SUGGESTIONS = data["suggestions"]
-    return request.app.state.config.DEFAULT_PROMPT_SUGGESTIONS
-
-
-############################
-# SetBanners
-############################
-
-
-class SetBannersForm(BaseModel):
-    banners: List[BannerModel]
-
-
-@router.post("/banners", response_model=List[BannerModel])
-async def set_banners(
-    request: Request,
-    form_data: SetBannersForm,
-    user=Depends(get_admin_user),
-):
-    data = form_data.model_dump()
-    request.app.state.config.BANNERS = data["banners"]
-    return request.app.state.config.BANNERS
-
-
-@router.get("/banners", response_model=List[BannerModel])
-async def get_banners(
-    request: Request,
-    user=Depends(get_verified_user),
-):
-    return request.app.state.config.BANNERS

backend/apps/webui/routers/documents.py

@@ -1,160 +0,0 @@
-from fastapi import Depends, FastAPI, HTTPException, status
-from datetime import datetime, timedelta
-from typing import List, Union, Optional
-
-from fastapi import APIRouter
-from pydantic import BaseModel
-import json
-
-from apps.webui.models.documents import (
-    Documents,
-    DocumentForm,
-    DocumentUpdateForm,
-    DocumentModel,
-    DocumentResponse,
-)
-
-from utils.utils import get_verified_user, get_admin_user
-from constants import ERROR_MESSAGES
-
-router = APIRouter()
-
-############################
-# GetDocuments
-############################
-
-
-@router.get("/", response_model=List[DocumentResponse])
-async def get_documents(user=Depends(get_verified_user)):
-    docs = [
-        DocumentResponse(
-            **{
-                **doc.model_dump(),
-                "content": json.loads(doc.content if doc.content else "{}"),
-            }
-        )
-        for doc in Documents.get_docs()
-    ]
-    return docs
-
-
-############################
-# CreateNewDoc
-############################
-
-
-@router.post("/create", response_model=Optional[DocumentResponse])
-async def create_new_doc(form_data: DocumentForm, user=Depends(get_admin_user)):
-    doc = Documents.get_doc_by_name(form_data.name)
-    if doc == None:
-        doc = Documents.insert_new_doc(user.id, form_data)
-
-        if doc:
-            return DocumentResponse(
-                **{
-                    **doc.model_dump(),
-                    "content": json.loads(doc.content if doc.content else "{}"),
-                }
-            )
-        else:
-            raise HTTPException(
-                status_code=status.HTTP_400_BAD_REQUEST,
-                detail=ERROR_MESSAGES.FILE_EXISTS,
-            )
-    else:
-        raise HTTPException(
-            status_code=status.HTTP_400_BAD_REQUEST,
-            detail=ERROR_MESSAGES.NAME_TAG_TAKEN,
-        )
-
-
-############################
-# GetDocByName
-############################
-
-
-@router.get("/doc", response_model=Optional[DocumentResponse])
-async def get_doc_by_name(name: str, user=Depends(get_verified_user)):
-    doc = Documents.get_doc_by_name(name)
-
-    if doc:
-        return DocumentResponse(
-            **{
-                **doc.model_dump(),
-                "content": json.loads(doc.content if doc.content else "{}"),
-            }
-        )
-    else:
-        raise HTTPException(
-            status_code=status.HTTP_401_UNAUTHORIZED,
-            detail=ERROR_MESSAGES.NOT_FOUND,
-        )
-
-
-############################
-# TagDocByName
-############################
-
-
-class TagItem(BaseModel):
-    name: str
-
-
-class TagDocumentForm(BaseModel):
-    name: str
-    tags: List[dict]
-
-
-@router.post("/doc/tags", response_model=Optional[DocumentResponse])
-async def tag_doc_by_name(form_data: TagDocumentForm, user=Depends(get_verified_user)):
-    doc = Documents.update_doc_content_by_name(form_data.name, {"tags": form_data.tags})
-
-    if doc:
-        return DocumentResponse(
-            **{
-                **doc.model_dump(),
-                "content": json.loads(doc.content if doc.content else "{}"),
-            }
-        )
-    else:
-        raise HTTPException(
-            status_code=status.HTTP_401_UNAUTHORIZED,
-            detail=ERROR_MESSAGES.NOT_FOUND,
-        )
-
-
-############################
-# UpdateDocByName
-############################
-
-
-@router.post("/doc/update", response_model=Optional[DocumentResponse])
-async def update_doc_by_name(
-    name: str,
-    form_data: DocumentUpdateForm,
-    user=Depends(get_admin_user),
-):
-    doc = Documents.update_doc_by_name(name, form_data)
-    if doc:
-        return DocumentResponse(
-            **{
-                **doc.model_dump(),
-                "content": json.loads(doc.content if doc.content else "{}"),
-            }
-        )
-    else:
-        raise HTTPException(
-            status_code=status.HTTP_400_BAD_REQUEST,
-            detail=ERROR_MESSAGES.NAME_TAG_TAKEN,
-        )
-
-
-############################
-# DeleteDocByName
-############################
-
-
-@router.delete("/doc/delete", response_model=bool)
-async def delete_doc_by_name(name: str, user=Depends(get_admin_user)):
-    result = Documents.delete_doc_by_name(name)
-    return result

backend/apps/webui/routers/files.py

@@ -1,241 +0,0 @@
-from fastapi import (
-    Depends,
-    FastAPI,
-    HTTPException,
-    status,
-    Request,
-    UploadFile,
-    File,
-    Form,
-)
-
-
-from datetime import datetime, timedelta
-from typing import List, Union, Optional
-from pathlib import Path
-
-from fastapi import APIRouter
-from fastapi.responses import StreamingResponse, JSONResponse, FileResponse
-
-from pydantic import BaseModel
-import json
-
-from apps.webui.models.files import (
-    Files,
-    FileForm,
-    FileModel,
-    FileModelResponse,
-)
-from utils.utils import get_verified_user, get_admin_user
-from constants import ERROR_MESSAGES
-
-from importlib import util
-import os
-import uuid
-import os, shutil, logging, re
-
-
-from config import SRC_LOG_LEVELS, UPLOAD_DIR
-
-
-log = logging.getLogger(__name__)
-log.setLevel(SRC_LOG_LEVELS["MODELS"])
-
-
-router = APIRouter()
-
-############################
-# Upload File
-############################
-
-
-@router.post("/")
-def upload_file(file: UploadFile = File(...), user=Depends(get_verified_user)):
-    log.info(f"file.content_type: {file.content_type}")
-    try:
-        unsanitized_filename = file.filename
-        filename = os.path.basename(unsanitized_filename)
-
-        # replace filename with uuid
-        id = str(uuid.uuid4())
-        name = filename
-        filename = f"{id}_{filename}"
-        file_path = f"{UPLOAD_DIR}/{filename}"
-
-        contents = file.file.read()
-        with open(file_path, "wb") as f:
-            f.write(contents)
-            f.close()
-
-        file = Files.insert_new_file(
-            user.id,
-            FileForm(
-                **{
-                    "id": id,
-                    "filename": filename,
-                    "meta": {
-                        "name": name,
-                        "content_type": file.content_type,
-                        "size": len(contents),
-                        "path": file_path,
-                    },
-                }
-            ),
-        )
-
-        if file:
-            return file
-        else:
-            raise HTTPException(
-                status_code=status.HTTP_400_BAD_REQUEST,
-                detail=ERROR_MESSAGES.DEFAULT("Error uploading file"),
-            )
-
-    except Exception as e:
-        log.exception(e)
-        raise HTTPException(
-            status_code=status.HTTP_400_BAD_REQUEST,
-            detail=ERROR_MESSAGES.DEFAULT(e),
-        )
-
-
-############################
-# List Files
-############################
-
-
-@router.get("/", response_model=List[FileModel])
-async def list_files(user=Depends(get_verified_user)):
-    files = Files.get_files()
-    return files
-
-
-############################
-# Delete All Files
-############################
-
-
-@router.delete("/all")
-async def delete_all_files(user=Depends(get_admin_user)):
-    result = Files.delete_all_files()
-
-    if result:
-        folder = f"{UPLOAD_DIR}"
-        try:
-            # Check if the directory exists
-            if os.path.exists(folder):
-                # Iterate over all the files and directories in the specified directory
-                for filename in os.listdir(folder):
-                    file_path = os.path.join(folder, filename)
-                    try:
-                        if os.path.isfile(file_path) or os.path.islink(file_path):
-                            os.unlink(file_path)  # Remove the file or link
-                        elif os.path.isdir(file_path):
-                            shutil.rmtree(file_path)  # Remove the directory
-                    except Exception as e:
-                        print(f"Failed to delete {file_path}. Reason: {e}")
-            else:
-                print(f"The directory {folder} does not exist")
-        except Exception as e:
-            print(f"Failed to process the directory {folder}. Reason: {e}")
-
-        return {"message": "All files deleted successfully"}
-    else:
-        raise HTTPException(
-            status_code=status.HTTP_400_BAD_REQUEST,
-            detail=ERROR_MESSAGES.DEFAULT("Error deleting files"),
-        )
-
-
-############################
-# Get File By Id
-############################
-
-
-@router.get("/{id}", response_model=Optional[FileModel])
-async def get_file_by_id(id: str, user=Depends(get_verified_user)):
-    file = Files.get_file_by_id(id)
-
-    if file:
-        return file
-    else:
-        raise HTTPException(
-            status_code=status.HTTP_404_NOT_FOUND,
-            detail=ERROR_MESSAGES.NOT_FOUND,
-        )
-
-
-############################
-# Get File Content By Id
-############################
-
-
-@router.get("/{id}/content", response_model=Optional[FileModel])
-async def get_file_content_by_id(id: str, user=Depends(get_verified_user)):
-    file = Files.get_file_by_id(id)
-
-    if file:
-        file_path = Path(file.meta["path"])
-
-        # Check if the file already exists in the cache
-        if file_path.is_file():
-            print(f"file_path: {file_path}")
-            return FileResponse(file_path)
-        else:
-            raise HTTPException(
-                status_code=status.HTTP_404_NOT_FOUND,
-                detail=ERROR_MESSAGES.NOT_FOUND,
-            )
-    else:
-        raise HTTPException(
-            status_code=status.HTTP_404_NOT_FOUND,
-            detail=ERROR_MESSAGES.NOT_FOUND,
-        )
-
-
-@router.get("/{id}/content/{file_name}", response_model=Optional[FileModel])
-async def get_file_content_by_id(id: str, user=Depends(get_verified_user)):
-    file = Files.get_file_by_id(id)
-
-    if file:
-        file_path = Path(file.meta["path"])
-
-        # Check if the file already exists in the cache
-        if file_path.is_file():
-            print(f"file_path: {file_path}")
-            return FileResponse(file_path)
-        else:
-            raise HTTPException(
-                status_code=status.HTTP_404_NOT_FOUND,
-                detail=ERROR_MESSAGES.NOT_FOUND,
-            )
-    else:
-        raise HTTPException(
-            status_code=status.HTTP_404_NOT_FOUND,
-            detail=ERROR_MESSAGES.NOT_FOUND,
-        )
-
-
-############################
-# Delete File By Id
-############################
-
-
-@router.delete("/{id}")
-async def delete_file_by_id(id: str, user=Depends(get_verified_user)):
-    file = Files.get_file_by_id(id)
-
-    if file:
-        result = Files.delete_file_by_id(id)
-        if result:
-            return {"message": "File deleted successfully"}
-        else:
-            raise HTTPException(
-                status_code=status.HTTP_400_BAD_REQUEST,
-                detail=ERROR_MESSAGES.DEFAULT("Error deleting file"),
-            )
-    else:
-        raise HTTPException(
-            status_code=status.HTTP_404_NOT_FOUND,
-            detail=ERROR_MESSAGES.NOT_FOUND,
-        )

backend/apps/webui/routers/models.py

@@ -1,113 +0,0 @@
-from fastapi import Depends, FastAPI, HTTPException, status, Request
-from datetime import datetime, timedelta
-from typing import List, Union, Optional
-
-from fastapi import APIRouter
-from pydantic import BaseModel
-import json
-
-from apps.webui.models.models import Models, ModelModel, ModelForm, ModelResponse
-
-from utils.utils import get_verified_user, get_admin_user
-from constants import ERROR_MESSAGES
-
-router = APIRouter()
-
-###########################
-# getModels
-###########################
-
-
-@router.get("/", response_model=List[ModelResponse])
-async def get_models(user=Depends(get_verified_user)):
-    return Models.get_all_models()
-
-
-############################
-# AddNewModel
-############################
-
-
-@router.post("/add", response_model=Optional[ModelModel])
-async def add_new_model(
-    request: Request,
-    form_data: ModelForm,
-    user=Depends(get_admin_user),
-):
-    if form_data.id in request.app.state.MODELS:
-        raise HTTPException(
-            status_code=status.HTTP_401_UNAUTHORIZED,
-            detail=ERROR_MESSAGES.MODEL_ID_TAKEN,
-        )
-    else:
-        model = Models.insert_new_model(form_data, user.id)
-
-        if model:
-            return model
-        else:
-            raise HTTPException(
-                status_code=status.HTTP_401_UNAUTHORIZED,
-                detail=ERROR_MESSAGES.DEFAULT(),
-            )
-
-
-############################
-# GetModelById
-############################
-
-
-@router.get("/", response_model=Optional[ModelModel])
-async def get_model_by_id(id: str, user=Depends(get_verified_user)):
-    model = Models.get_model_by_id(id)
-
-    if model:
-        return model
-    else:
-        raise HTTPException(
-            status_code=status.HTTP_401_UNAUTHORIZED,
-            detail=ERROR_MESSAGES.NOT_FOUND,
-        )
-
-
-############################
-# UpdateModelById
-############################
-
-
-@router.post("/update", response_model=Optional[ModelModel])
-async def update_model_by_id(
-    request: Request,
-    id: str,
-    form_data: ModelForm,
-    user=Depends(get_admin_user),
-):
-    model = Models.get_model_by_id(id)
-    if model:
-        model = Models.update_model_by_id(id, form_data)
-        return model
-    else:
-        if form_data.id in request.app.state.MODELS:
-            model = Models.insert_new_model(form_data, user.id)
-            if model:
-                return model
-            else:
-                raise HTTPException(
-                    status_code=status.HTTP_401_UNAUTHORIZED,
-                    detail=ERROR_MESSAGES.DEFAULT(),
-                )
-        else:
-            raise HTTPException(
-                status_code=status.HTTP_401_UNAUTHORIZED,
-                detail=ERROR_MESSAGES.DEFAULT(),
-            )
-
-
-############################
-# DeleteModelById
-############################
-
-
-@router.delete("/delete", response_model=bool)
-async def delete_model_by_id(id: str, user=Depends(get_admin_user)):
-    result = Models.delete_model_by_id(id)
-    return result

backend/apps/webui/routers/prompts.py

@@ -1,96 +0,0 @@
-from fastapi import Depends, FastAPI, HTTPException, status
-from datetime import datetime, timedelta
-from typing import List, Union, Optional
-
-from fastapi import APIRouter
-from pydantic import BaseModel
-import json
-
-from apps.webui.models.prompts import Prompts, PromptForm, PromptModel
-
-from utils.utils import get_verified_user, get_admin_user
-from constants import ERROR_MESSAGES
-
-router = APIRouter()
-
-############################
-# GetPrompts
-############################
-
-
-@router.get("/", response_model=List[PromptModel])
-async def get_prompts(user=Depends(get_verified_user)):
-    return Prompts.get_prompts()
-
-
-############################
-# CreateNewPrompt
-############################
-
-
-@router.post("/create", response_model=Optional[PromptModel])
-async def create_new_prompt(form_data: PromptForm, user=Depends(get_admin_user)):
-    prompt = Prompts.get_prompt_by_command(form_data.command)
-    if prompt == None:
-        prompt = Prompts.insert_new_prompt(user.id, form_data)
-
-        if prompt:
-            return prompt
-        raise HTTPException(
-            status_code=status.HTTP_400_BAD_REQUEST,
-            detail=ERROR_MESSAGES.DEFAULT(),
-        )
-    raise HTTPException(
-        status_code=status.HTTP_400_BAD_REQUEST,
-        detail=ERROR_MESSAGES.COMMAND_TAKEN,
-    )
-
-
-############################
-# GetPromptByCommand
-############################
-
-
-@router.get("/command/{command}", response_model=Optional[PromptModel])
-async def get_prompt_by_command(command: str, user=Depends(get_verified_user)):
-    prompt = Prompts.get_prompt_by_command(f"/{command}")
-
-    if prompt:
-        return prompt
-    else:
-        raise HTTPException(
-            status_code=status.HTTP_401_UNAUTHORIZED,
-            detail=ERROR_MESSAGES.NOT_FOUND,
-        )
-
-
-############################
-# UpdatePromptByCommand
-############################
-
-
-@router.post("/command/{command}/update", response_model=Optional[PromptModel])
-async def update_prompt_by_command(
-    command: str,
-    form_data: PromptForm,
-    user=Depends(get_admin_user),
-):
-    prompt = Prompts.update_prompt_by_command(f"/{command}", form_data)
-    if prompt:
-        return prompt
-    else:
-        raise HTTPException(
-            status_code=status.HTTP_401_UNAUTHORIZED,
-            detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
-        )
-
-
-############################
-# DeletePromptByCommand
-############################
-
-
-@router.delete("/command/{command}/delete", response_model=bool)
-async def delete_prompt_by_command(command: str, user=Depends(get_admin_user)):
-    result = Prompts.delete_prompt_by_command(f"/{command}")
-    return result

backend/apps/webui/routers/utils.py

@@ -1,148 +0,0 @@
-from pathlib import Path
-import site
-
-from fastapi import APIRouter, UploadFile, File, Response
-from fastapi import Depends, HTTPException, status
-from starlette.responses import StreamingResponse, FileResponse
-from pydantic import BaseModel
-
-
-from fpdf import FPDF
-import markdown
-import black
-
-
-from utils.utils import get_admin_user
-from utils.misc import calculate_sha256, get_gravatar_url
-
-from config import OLLAMA_BASE_URLS, DATA_DIR, UPLOAD_DIR, ENABLE_ADMIN_EXPORT
-from constants import ERROR_MESSAGES
-from typing import List
-
-router = APIRouter()
-
-
-@router.get("/gravatar")
-async def get_gravatar(
-    email: str,
-):
-    return get_gravatar_url(email)
-
-
-class CodeFormatRequest(BaseModel):
-    code: str
-
-
-@router.post("/code/format")
-async def format_code(request: CodeFormatRequest):
-    try:
-        formatted_code = black.format_str(request.code, mode=black.Mode())
-        return {"code": formatted_code}
-    except black.NothingChanged:
-        return {"code": request.code}
-    except Exception as e:
-        raise HTTPException(status_code=400, detail=str(e))
-
-
-class MarkdownForm(BaseModel):
-    md: str
-
-
-@router.post("/markdown")
-async def get_html_from_markdown(
-    form_data: MarkdownForm,
-):
-    return {"html": markdown.markdown(form_data.md)}
-
-
-class ChatForm(BaseModel):
-    title: str
-    messages: List[dict]
-
-
-@router.post("/pdf")
-async def download_chat_as_pdf(
-    form_data: ChatForm,
-):
-    pdf = FPDF()
-    pdf.add_page()
-
-    # When running in docker, workdir is /app/backend, so fonts is in /app/backend/static/fonts
-    FONTS_DIR = Path("./static/fonts")
-
-    # Non Docker Installation
-
-    # When running using `pip install` the static directory is in the site packages.
-    if not FONTS_DIR.exists():
-        FONTS_DIR = Path(site.getsitepackages()[0]) / "static/fonts"
-    # When running using `pip install -e .` the static directory is in the site packages.
-    # This path only works if `open-webui serve` is run from the root of this project.
-    if not FONTS_DIR.exists():
-        FONTS_DIR = Path("./backend/static/fonts")
-
-    pdf.add_font("NotoSans", "", f"{FONTS_DIR}/NotoSans-Regular.ttf")
-    pdf.add_font("NotoSans", "b", f"{FONTS_DIR}/NotoSans-Bold.ttf")
-    pdf.add_font("NotoSans", "i", f"{FONTS_DIR}/NotoSans-Italic.ttf")
-    pdf.add_font("NotoSansKR", "", f"{FONTS_DIR}/NotoSansKR-Regular.ttf")
-    pdf.add_font("NotoSansJP", "", f"{FONTS_DIR}/NotoSansJP-Regular.ttf")
-
-    pdf.set_font("NotoSans", size=12)
-    pdf.set_fallback_fonts(["NotoSansKR", "NotoSansJP"])
-
-    pdf.set_auto_page_break(auto=True, margin=15)
-
-    # Adjust the effective page width for multi_cell
-    effective_page_width = (
-        pdf.w - 2 * pdf.l_margin - 10
-    )  # Subtracted an additional 10 for extra padding
-
-    # Add chat messages
-    for message in form_data.messages:
-        role = message["role"]
-        content = message["content"]
-        pdf.set_font("NotoSans", "B", size=14)  # Bold for the role
-        pdf.multi_cell(effective_page_width, 10, f"{role.upper()}", 0, "L")
-        pdf.ln(1)  # Extra space between messages
-
-        pdf.set_font("NotoSans", size=10)  # Regular for content
-        pdf.multi_cell(effective_page_width, 6, content, 0, "L")
-        pdf.ln(1.5)  # Extra space between messages
-
-    # Save the pdf with name .pdf
-    pdf_bytes = pdf.output()
-
-    return Response(
-        content=bytes(pdf_bytes),
-        media_type="application/pdf",
-        headers={"Content-Disposition": f"attachment;filename=chat.pdf"},
-    )
-
-
-@router.get("/db/download")
-async def download_db(user=Depends(get_admin_user)):
-    if not ENABLE_ADMIN_EXPORT:
-        raise HTTPException(
-            status_code=status.HTTP_401_UNAUTHORIZED,
-            detail=ERROR_MESSAGES.ACCESS_PROHIBITED,
-        )
-    from apps.webui.internal.db import engine
-
-    if engine.name != "sqlite":
-        raise HTTPException(
-            status_code=status.HTTP_400_BAD_REQUEST,
-            detail=ERROR_MESSAGES.DB_NOT_SQLITE,
-        )
-    return FileResponse(
-        engine.url.database,
-        media_type="application/octet-stream",
-        filename="webui.db",
-    )
-
-
-@router.get("/litellm/config")
-async def download_litellm_config_yaml(user=Depends(get_admin_user)):
-    return FileResponse(
-        f"{DATA_DIR}/litellm/config.yaml",
-        media_type="application/octet-stream",
-        filename="config.yaml",
-    )

backend/apps/webui/utils.py

@@ -1,90 +0,0 @@
-from importlib import util
-import os
-import re
-
-from config import TOOLS_DIR, FUNCTIONS_DIR
-
-
-def extract_frontmatter(file_path):
-    """
-    Extract frontmatter as a dictionary from the specified file path.
-    """
-    frontmatter = {}
-    frontmatter_started = False
-    frontmatter_ended = False
-    frontmatter_pattern = re.compile(r"^\s*([a-z_]+):\s*(.*)\s*$", re.IGNORECASE)
-
-    try:
-        with open(file_path, "r", encoding="utf-8") as file:
-            first_line = file.readline()
-            if first_line.strip() != '"""':
-                # The file doesn't start with triple quotes
-                return {}
-
-            frontmatter_started = True
-
-            for line in file:
-                if '"""' in line:
-                    if frontmatter_started:
-                        frontmatter_ended = True
-                        break
-
-                if frontmatter_started and not frontmatter_ended:
-                    match = frontmatter_pattern.match(line)
-                    if match:
-                        key, value = match.groups()
-                        frontmatter[key.strip()] = value.strip()
-
-    except FileNotFoundError:
-        print(f"Error: The file {file_path} does not exist.")
-        return {}
-    except Exception as e:
-        print(f"An error occurred: {e}")
-        return {}
-
-    return frontmatter
-
-
-def load_toolkit_module_by_id(toolkit_id):
-    toolkit_path = os.path.join(TOOLS_DIR, f"{toolkit_id}.py")
-    spec = util.spec_from_file_location(toolkit_id, toolkit_path)
-    module = util.module_from_spec(spec)
-    frontmatter = extract_frontmatter(toolkit_path)
-
-    try:
-        spec.loader.exec_module(module)
-        print(f"Loaded module: {module.__name__}")
-        if hasattr(module, "Tools"):
-            return module.Tools(), frontmatter
-        else:
-            raise Exception("No Tools class found")
-    except Exception as e:
-        print(f"Error loading module: {toolkit_id}")
-        # Move the file to the error folder
-        os.rename(toolkit_path, f"{toolkit_path}.error")
-        raise e
-
-
-def load_function_module_by_id(function_id):
-    function_path = os.path.join(FUNCTIONS_DIR, f"{function_id}.py")
-
-    spec = util.spec_from_file_location(function_id, function_path)
-    module = util.module_from_spec(spec)
-    frontmatter = extract_frontmatter(function_path)
-
-    try:
-        spec.loader.exec_module(module)
-        print(f"Loaded module: {module.__name__}")
-        if hasattr(module, "Pipe"):
-            return module.Pipe(), "pipe", frontmatter
-        elif hasattr(module, "Filter"):
-            return module.Filter(), "filter", frontmatter
-        elif hasattr(module, "Action"):
-            return module.Action(), "action", frontmatter
-        else:
-            raise Exception("No Function class found")
-    except Exception as e:
-        print(f"Error loading module: {function_id}")
-        # Move the file to the error folder
-        os.rename(function_path, f"{function_path}.error")
-        raise e

backend/data/config.json

@@ -1,36 +0,0 @@
-{
-	"version": 0,
-	"ui": {
-		"default_locale": "",
-		"prompt_suggestions": [
-			{
-				"title": ["Help me study", "vocabulary for a college entrance exam"],
-				"content": "Help me study vocabulary: write a sentence for me to fill in the blank, and I'll try to pick the correct option."
-			},
-			{
-				"title": ["Give me ideas", "for what to do with my kids' art"],
-				"content": "What are 5 creative things I could do with my kids' art? I don't want to throw them away, but it's also so much clutter."
-			},
-			{
-				"title": ["Tell me a fun fact", "about the Roman Empire"],
-				"content": "Tell me a random fun fact about the Roman Empire"
-			},
-			{
-				"title": ["Show me a code snippet", "of a website's sticky header"],
-				"content": "Show me a code snippet of a website's sticky header in CSS and JavaScript."
-			},
-			{
-				"title": ["Explain options trading", "if I'm familiar with buying and selling stocks"],
-				"content": "Explain options trading in simple terms if I'm familiar with buying and selling stocks."
-			},
-			{
-				"title": ["Overcome procrastination", "give me tips"],
-				"content": "Could you start by asking me about instances when I procrastinate the most and then give me some suggestions to overcome it?"
-			},
-			{
-				"title": ["Grammar check", "rewrite it for better readability "],
-				"content": "Check the following sentence for grammar and clarity: \"[sentence]\". Rewrite it for better readability while maintaining its original meaning."
-			}
-		]
-	}
-}

backend/data/litellm/config.yaml

@@ -1,4 +0,0 @@
-general_settings: {}
-litellm_settings: {}
-model_list: []
-router_settings: {}

backend/data/readme.txt

@@ -1 +1 @@
-dir for backend files (db, documents, etc.)
+docker dir for backend files (db, documents, etc.)

backend/dev.sh

@@ -1,2 +1,2 @@
 PORT="${PORT:-8080}"
-uvicorn main:app --port $PORT --host 0.0.0.0 --forwarded-allow-ips '*' --reload
+uvicorn open_webui.main:app --port $PORT --host 0.0.0.0 --forwarded-allow-ips '*' --reload

backend/open_webui/__init__.py

@@ -9,8 +9,6 @@ import uvicorn
 app = typer.Typer()
 
 KEY_FILE = Path.cwd() / ".webui_secret_key"
-if (frontend_build_dir := Path(__file__).parent / "frontend").exists():
-    os.environ["FRONTEND_BUILD_DIR"] = str(frontend_build_dir)
 
 
 @app.command()
@@ -18,6 +16,7 @@ def serve(
     host: str = "0.0.0.0",
     port: int = 8080,
 ):
+    os.environ["FROM_INIT_PY"] = "true"
     if os.getenv("WEBUI_SECRET_KEY") is None:
         typer.echo(
             "Loading WEBUI_SECRET_KEY from file, not provided as an environment variable."
@@ -40,9 +39,23 @@ def serve(
                 "/usr/local/lib/python3.11/site-packages/nvidia/cudnn/lib",
             ]
         )
-    import main  # we need set environment variables before importing main
+        try:
+            import torch
 
-    uvicorn.run(main.app, host=host, port=port, forwarded_allow_ips="*")
+            assert torch.cuda.is_available(), "CUDA not available"
+            typer.echo("CUDA seems to be working")
+        except Exception as e:
+            typer.echo(
+                "Error when testing CUDA but USE_CUDA_DOCKER is true. "
+                "Resetting USE_CUDA_DOCKER to false and removing "
+                f"LD_LIBRARY_PATH modifications: {e}"
+            )
+            os.environ["USE_CUDA_DOCKER"] = "false"
+            os.environ["LD_LIBRARY_PATH"] = ":".join(LD_LIBRARY_PATH)
+
+    import open_webui.main  # we need set environment variables before importing main
+
+    uvicorn.run(open_webui.main.app, host=host, port=port, forwarded_allow_ips="*")
 
 
 @app.command()
@@ -52,7 +65,11 @@ def dev(
     reload: bool = True,
 ):
     uvicorn.run(
-        "main:app", host=host, port=port, reload=reload, forwarded_allow_ips="*"
+        "open_webui.main:app",
+        host=host,
+        port=port,
+        reload=reload,
+        forwarded_allow_ips="*",
     )
 
 

backend/open_webui/apps/audio/main.py

@@ -0,0 +1,703 @@
+import hashlib
+import json
+import logging
+import os
+import uuid
+from functools import lru_cache
+from pathlib import Path
+from pydub import AudioSegment
+from pydub.silence import split_on_silence
+
+import aiohttp
+import aiofiles
+import requests
+from open_webui.config import (
+    AUDIO_STT_ENGINE,
+    AUDIO_STT_MODEL,
+    AUDIO_STT_OPENAI_API_BASE_URL,
+    AUDIO_STT_OPENAI_API_KEY,
+    AUDIO_TTS_API_KEY,
+    AUDIO_TTS_ENGINE,
+    AUDIO_TTS_MODEL,
+    AUDIO_TTS_OPENAI_API_BASE_URL,
+    AUDIO_TTS_OPENAI_API_KEY,
+    AUDIO_TTS_SPLIT_ON,
+    AUDIO_TTS_VOICE,
+    AUDIO_TTS_AZURE_SPEECH_REGION,
+    AUDIO_TTS_AZURE_SPEECH_OUTPUT_FORMAT,
+    CACHE_DIR,
+    CORS_ALLOW_ORIGIN,
+    WHISPER_MODEL,
+    WHISPER_MODEL_AUTO_UPDATE,
+    WHISPER_MODEL_DIR,
+    AppConfig,
+)
+
+from open_webui.constants import ERROR_MESSAGES
+from open_webui.env import (
+    ENV,
+    SRC_LOG_LEVELS,
+    DEVICE_TYPE,
+    ENABLE_FORWARD_USER_INFO_HEADERS,
+)
+
+from fastapi import Depends, FastAPI, File, HTTPException, Request, UploadFile, status
+from fastapi.middleware.cors import CORSMiddleware
+from fastapi.responses import FileResponse
+from pydantic import BaseModel
+from open_webui.utils.utils import get_admin_user, get_verified_user
+
+# Constants
+MAX_FILE_SIZE_MB = 25
+MAX_FILE_SIZE = MAX_FILE_SIZE_MB * 1024 * 1024  # Convert MB to bytes
+
+
+log = logging.getLogger(__name__)
+log.setLevel(SRC_LOG_LEVELS["AUDIO"])
+
+app = FastAPI(
+    docs_url="/docs" if ENV == "dev" else None,
+    openapi_url="/openapi.json" if ENV == "dev" else None,
+    redoc_url=None,
+)
+
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=CORS_ALLOW_ORIGIN,
+    allow_credentials=True,
+    allow_methods=["*"],
+    allow_headers=["*"],
+)
+
+app.state.config = AppConfig()
+
+app.state.config.STT_OPENAI_API_BASE_URL = AUDIO_STT_OPENAI_API_BASE_URL
+app.state.config.STT_OPENAI_API_KEY = AUDIO_STT_OPENAI_API_KEY
+app.state.config.STT_ENGINE = AUDIO_STT_ENGINE
+app.state.config.STT_MODEL = AUDIO_STT_MODEL
+
+app.state.config.WHISPER_MODEL = WHISPER_MODEL
+app.state.faster_whisper_model = None
+
+app.state.config.TTS_OPENAI_API_BASE_URL = AUDIO_TTS_OPENAI_API_BASE_URL
+app.state.config.TTS_OPENAI_API_KEY = AUDIO_TTS_OPENAI_API_KEY
+app.state.config.TTS_ENGINE = AUDIO_TTS_ENGINE
+app.state.config.TTS_MODEL = AUDIO_TTS_MODEL
+app.state.config.TTS_VOICE = AUDIO_TTS_VOICE
+app.state.config.TTS_API_KEY = AUDIO_TTS_API_KEY
+app.state.config.TTS_SPLIT_ON = AUDIO_TTS_SPLIT_ON
+
+
+app.state.speech_synthesiser = None
+app.state.speech_speaker_embeddings_dataset = None
+
+app.state.config.TTS_AZURE_SPEECH_REGION = AUDIO_TTS_AZURE_SPEECH_REGION
+app.state.config.TTS_AZURE_SPEECH_OUTPUT_FORMAT = AUDIO_TTS_AZURE_SPEECH_OUTPUT_FORMAT
+
+# setting device type for whisper model
+whisper_device_type = DEVICE_TYPE if DEVICE_TYPE and DEVICE_TYPE == "cuda" else "cpu"
+log.info(f"whisper_device_type: {whisper_device_type}")
+
+SPEECH_CACHE_DIR = Path(CACHE_DIR).joinpath("./audio/speech/")
+SPEECH_CACHE_DIR.mkdir(parents=True, exist_ok=True)
+
+
+def set_faster_whisper_model(model: str, auto_update: bool = False):
+    if model and app.state.config.STT_ENGINE == "":
+        from faster_whisper import WhisperModel
+
+        faster_whisper_kwargs = {
+            "model_size_or_path": model,
+            "device": whisper_device_type,
+            "compute_type": "int8",
+            "download_root": WHISPER_MODEL_DIR,
+            "local_files_only": not auto_update,
+        }
+
+        try:
+            app.state.faster_whisper_model = WhisperModel(**faster_whisper_kwargs)
+        except Exception:
+            log.warning(
+                "WhisperModel initialization failed, attempting download with local_files_only=False"
+            )
+            faster_whisper_kwargs["local_files_only"] = False
+            app.state.faster_whisper_model = WhisperModel(**faster_whisper_kwargs)
+
+    else:
+        app.state.faster_whisper_model = None
+
+
+class TTSConfigForm(BaseModel):
+    OPENAI_API_BASE_URL: str
+    OPENAI_API_KEY: str
+    API_KEY: str
+    ENGINE: str
+    MODEL: str
+    VOICE: str
+    SPLIT_ON: str
+    AZURE_SPEECH_REGION: str
+    AZURE_SPEECH_OUTPUT_FORMAT: str
+
+
+class STTConfigForm(BaseModel):
+    OPENAI_API_BASE_URL: str
+    OPENAI_API_KEY: str
+    ENGINE: str
+    MODEL: str
+    WHISPER_MODEL: str
+
+
+class AudioConfigUpdateForm(BaseModel):
+    tts: TTSConfigForm
+    stt: STTConfigForm
+
+
+from pydub import AudioSegment
+from pydub.utils import mediainfo
+
+
+def is_mp4_audio(file_path):
+    """Check if the given file is an MP4 audio file."""
+    if not os.path.isfile(file_path):
+        print(f"File not found: {file_path}")
+        return False
+
+    info = mediainfo(file_path)
+    if (
+        info.get("codec_name") == "aac"
+        and info.get("codec_type") == "audio"
+        and info.get("codec_tag_string") == "mp4a"
+    ):
+        return True
+    return False
+
+
+def convert_mp4_to_wav(file_path, output_path):
+    """Convert MP4 audio file to WAV format."""
+    audio = AudioSegment.from_file(file_path, format="mp4")
+    audio.export(output_path, format="wav")
+    print(f"Converted {file_path} to {output_path}")
+
+
+@app.get("/config")
+async def get_audio_config(user=Depends(get_admin_user)):
+    return {
+        "tts": {
+            "OPENAI_API_BASE_URL": app.state.config.TTS_OPENAI_API_BASE_URL,
+            "OPENAI_API_KEY": app.state.config.TTS_OPENAI_API_KEY,
+            "API_KEY": app.state.config.TTS_API_KEY,
+            "ENGINE": app.state.config.TTS_ENGINE,
+            "MODEL": app.state.config.TTS_MODEL,
+            "VOICE": app.state.config.TTS_VOICE,
+            "SPLIT_ON": app.state.config.TTS_SPLIT_ON,
+            "AZURE_SPEECH_REGION": app.state.config.TTS_AZURE_SPEECH_REGION,
+            "AZURE_SPEECH_OUTPUT_FORMAT": app.state.config.TTS_AZURE_SPEECH_OUTPUT_FORMAT,
+        },
+        "stt": {
+            "OPENAI_API_BASE_URL": app.state.config.STT_OPENAI_API_BASE_URL,
+            "OPENAI_API_KEY": app.state.config.STT_OPENAI_API_KEY,
+            "ENGINE": app.state.config.STT_ENGINE,
+            "MODEL": app.state.config.STT_MODEL,
+            "WHISPER_MODEL": app.state.config.WHISPER_MODEL,
+        },
+    }
+
+
+@app.post("/config/update")
+async def update_audio_config(
+    form_data: AudioConfigUpdateForm, user=Depends(get_admin_user)
+):
+    app.state.config.TTS_OPENAI_API_BASE_URL = form_data.tts.OPENAI_API_BASE_URL
+    app.state.config.TTS_OPENAI_API_KEY = form_data.tts.OPENAI_API_KEY
+    app.state.config.TTS_API_KEY = form_data.tts.API_KEY
+    app.state.config.TTS_ENGINE = form_data.tts.ENGINE
+    app.state.config.TTS_MODEL = form_data.tts.MODEL
+    app.state.config.TTS_VOICE = form_data.tts.VOICE
+    app.state.config.TTS_SPLIT_ON = form_data.tts.SPLIT_ON
+    app.state.config.TTS_AZURE_SPEECH_REGION = form_data.tts.AZURE_SPEECH_REGION
+    app.state.config.TTS_AZURE_SPEECH_OUTPUT_FORMAT = (
+        form_data.tts.AZURE_SPEECH_OUTPUT_FORMAT
+    )
+
+    app.state.config.STT_OPENAI_API_BASE_URL = form_data.stt.OPENAI_API_BASE_URL
+    app.state.config.STT_OPENAI_API_KEY = form_data.stt.OPENAI_API_KEY
+    app.state.config.STT_ENGINE = form_data.stt.ENGINE
+    app.state.config.STT_MODEL = form_data.stt.MODEL
+    app.state.config.WHISPER_MODEL = form_data.stt.WHISPER_MODEL
+    set_faster_whisper_model(form_data.stt.WHISPER_MODEL, WHISPER_MODEL_AUTO_UPDATE)
+
+    return {
+        "tts": {
+            "OPENAI_API_BASE_URL": app.state.config.TTS_OPENAI_API_BASE_URL,
+            "OPENAI_API_KEY": app.state.config.TTS_OPENAI_API_KEY,
+            "API_KEY": app.state.config.TTS_API_KEY,
+            "ENGINE": app.state.config.TTS_ENGINE,
+            "MODEL": app.state.config.TTS_MODEL,
+            "VOICE": app.state.config.TTS_VOICE,
+            "SPLIT_ON": app.state.config.TTS_SPLIT_ON,
+            "AZURE_SPEECH_REGION": app.state.config.TTS_AZURE_SPEECH_REGION,
+            "AZURE_SPEECH_OUTPUT_FORMAT": app.state.config.TTS_AZURE_SPEECH_OUTPUT_FORMAT,
+        },
+        "stt": {
+            "OPENAI_API_BASE_URL": app.state.config.STT_OPENAI_API_BASE_URL,
+            "OPENAI_API_KEY": app.state.config.STT_OPENAI_API_KEY,
+            "ENGINE": app.state.config.STT_ENGINE,
+            "MODEL": app.state.config.STT_MODEL,
+            "WHISPER_MODEL": app.state.config.WHISPER_MODEL,
+        },
+    }
+
+
+def load_speech_pipeline():
+    from transformers import pipeline
+    from datasets import load_dataset
+
+    if app.state.speech_synthesiser is None:
+        app.state.speech_synthesiser = pipeline(
+            "text-to-speech", "microsoft/speecht5_tts"
+        )
+
+    if app.state.speech_speaker_embeddings_dataset is None:
+        app.state.speech_speaker_embeddings_dataset = load_dataset(
+            "Matthijs/cmu-arctic-xvectors", split="validation"
+        )
+
+
+@app.post("/speech")
+async def speech(request: Request, user=Depends(get_verified_user)):
+    body = await request.body()
+    name = hashlib.sha256(body).hexdigest()
+
+    file_path = SPEECH_CACHE_DIR.joinpath(f"{name}.mp3")
+    file_body_path = SPEECH_CACHE_DIR.joinpath(f"{name}.json")
+
+    # Check if the file already exists in the cache
+    if file_path.is_file():
+        return FileResponse(file_path)
+
+    if app.state.config.TTS_ENGINE == "openai":
+        headers = {}
+        headers["Authorization"] = f"Bearer {app.state.config.TTS_OPENAI_API_KEY}"
+        headers["Content-Type"] = "application/json"
+
+        if ENABLE_FORWARD_USER_INFO_HEADERS:
+            headers["X-OpenWebUI-User-Name"] = user.name
+            headers["X-OpenWebUI-User-Id"] = user.id
+            headers["X-OpenWebUI-User-Email"] = user.email
+            headers["X-OpenWebUI-User-Role"] = user.role
+
+        try:
+            body = body.decode("utf-8")
+            body = json.loads(body)
+            body["model"] = app.state.config.TTS_MODEL
+            body = json.dumps(body).encode("utf-8")
+        except Exception:
+            pass
+
+        try:
+            async with aiohttp.ClientSession() as session:
+                async with session.post(
+                    url=f"{app.state.config.TTS_OPENAI_API_BASE_URL}/audio/speech",
+                    data=body,
+                    headers=headers,
+                ) as r:
+                    r.raise_for_status()
+                    async with aiofiles.open(file_path, "wb") as f:
+                        await f.write(await r.read())
+
+                    async with aiofiles.open(file_body_path, "w") as f:
+                        await f.write(json.dumps(json.loads(body.decode("utf-8"))))
+
+            return FileResponse(file_path)
+
+        except Exception as e:
+            log.exception(e)
+            error_detail = "Open WebUI: Server Connection Error"
+            try:
+                if r.status != 200:
+                    res = await r.json()
+                    if "error" in res:
+                        error_detail = f"External: {res['error']['message']}"
+            except Exception:
+                error_detail = f"External: {e}"
+
+            raise HTTPException(
+                status_code=getattr(r, "status", 500),
+                detail=error_detail,
+            )
+
+    elif app.state.config.TTS_ENGINE == "elevenlabs":
+        try:
+            payload = json.loads(body.decode("utf-8"))
+        except Exception as e:
+            log.exception(e)
+            raise HTTPException(status_code=400, detail="Invalid JSON payload")
+
+        voice_id = payload.get("voice", "")
+        if voice_id not in get_available_voices():
+            raise HTTPException(
+                status_code=400,
+                detail="Invalid voice id",
+            )
+
+        url = f"https://api.elevenlabs.io/v1/text-to-speech/{voice_id}"
+        headers = {
+            "Accept": "audio/mpeg",
+            "Content-Type": "application/json",
+            "xi-api-key": app.state.config.TTS_API_KEY,
+        }
+        data = {
+            "text": payload["input"],
+            "model_id": app.state.config.TTS_MODEL,
+            "voice_settings": {"stability": 0.5, "similarity_boost": 0.5},
+        }
+
+        try:
+            async with aiohttp.ClientSession() as session:
+                async with session.post(url, json=data, headers=headers) as r:
+                    r.raise_for_status()
+                    async with aiofiles.open(file_path, "wb") as f:
+                        await f.write(await r.read())
+
+                    async with aiofiles.open(file_body_path, "w") as f:
+                        await f.write(json.dumps(json.loads(body.decode("utf-8"))))
+
+            return FileResponse(file_path)
+
+        except Exception as e:
+            log.exception(e)
+            error_detail = "Open WebUI: Server Connection Error"
+            try:
+                if r.status != 200:
+                    res = await r.json()
+                    if "error" in res:
+                        error_detail = f"External: {res['error']['message']}"
+            except Exception:
+                error_detail = f"External: {e}"
+
+            raise HTTPException(
+                status_code=getattr(r, "status", 500),
+                detail=error_detail,
+            )
+
+    elif app.state.config.TTS_ENGINE == "azure":
+        try:
+            payload = json.loads(body.decode("utf-8"))
+        except Exception as e:
+            log.exception(e)
+            raise HTTPException(status_code=400, detail="Invalid JSON payload")
+
+        region = app.state.config.TTS_AZURE_SPEECH_REGION
+        language = app.state.config.TTS_VOICE
+        locale = "-".join(app.state.config.TTS_VOICE.split("-")[:1])
+        output_format = app.state.config.TTS_AZURE_SPEECH_OUTPUT_FORMAT
+        url = f"https://{region}.tts.speech.microsoft.com/cognitiveservices/v1"
+
+        headers = {
+            "Ocp-Apim-Subscription-Key": app.state.config.TTS_API_KEY,
+            "Content-Type": "application/ssml+xml",
+            "X-Microsoft-OutputFormat": output_format,
+        }
+
+        data = f"""<speak version="1.0" xmlns="http://www.w3.org/2001/10/synthesis" xml:lang="{locale}">
+                <voice name="{language}">{payload["input"]}</voice>
+            </speak>"""
+
+        try:
+            async with aiohttp.ClientSession() as session:
+                async with session.post(url, headers=headers, data=data) as response:
+                    if response.status == 200:
+                        async with aiofiles.open(file_path, "wb") as f:
+                            await f.write(await response.read())
+                        return FileResponse(file_path)
+                    else:
+                        error_msg = f"Error synthesizing speech - {response.reason}"
+                        log.error(error_msg)
+                        raise HTTPException(status_code=500, detail=error_msg)
+        except Exception as e:
+            log.exception(e)
+            raise HTTPException(status_code=500, detail=str(e))
+    elif app.state.config.TTS_ENGINE == "transformers":
+        payload = None
+        try:
+            payload = json.loads(body.decode("utf-8"))
+        except Exception as e:
+            log.exception(e)
+            raise HTTPException(status_code=400, detail="Invalid JSON payload")
+
+        import torch
+        import soundfile as sf
+
+        load_speech_pipeline()
+
+        embeddings_dataset = app.state.speech_speaker_embeddings_dataset
+
+        speaker_index = 6799
+        try:
+            speaker_index = embeddings_dataset["filename"].index(
+                app.state.config.TTS_MODEL
+            )
+        except Exception:
+            pass
+
+        speaker_embedding = torch.tensor(
+            embeddings_dataset[speaker_index]["xvector"]
+        ).unsqueeze(0)
+
+        speech = app.state.speech_synthesiser(
+            payload["input"],
+            forward_params={"speaker_embeddings": speaker_embedding},
+        )
+
+        sf.write(file_path, speech["audio"], samplerate=speech["sampling_rate"])
+        with open(file_body_path, "w") as f:
+            json.dump(json.loads(body.decode("utf-8")), f)
+
+        return FileResponse(file_path)
+
+
+def transcribe(file_path):
+    print("transcribe", file_path)
+    filename = os.path.basename(file_path)
+    file_dir = os.path.dirname(file_path)
+    id = filename.split(".")[0]
+
+    if app.state.config.STT_ENGINE == "":
+        if app.state.faster_whisper_model is None:
+            set_faster_whisper_model(app.state.config.WHISPER_MODEL)
+
+        model = app.state.faster_whisper_model
+        segments, info = model.transcribe(file_path, beam_size=5)
+        log.info(
+            "Detected language '%s' with probability %f"
+            % (info.language, info.language_probability)
+        )
+
+        transcript = "".join([segment.text for segment in list(segments)])
+        data = {"text": transcript.strip()}
+
+        # save the transcript to a json file
+        transcript_file = f"{file_dir}/{id}.json"
+        with open(transcript_file, "w") as f:
+            json.dump(data, f)
+
+        log.debug(data)
+        return data
+    elif app.state.config.STT_ENGINE == "openai":
+        if is_mp4_audio(file_path):
+            print("is_mp4_audio")
+            os.rename(file_path, file_path.replace(".wav", ".mp4"))
+            # Convert MP4 audio file to WAV format
+            convert_mp4_to_wav(file_path.replace(".wav", ".mp4"), file_path)
+
+        headers = {"Authorization": f"Bearer {app.state.config.STT_OPENAI_API_KEY}"}
+
+        files = {"file": (filename, open(file_path, "rb"))}
+        data = {"model": app.state.config.STT_MODEL}
+
+        log.debug(files, data)
+
+        r = None
+        try:
+            r = requests.post(
+                url=f"{app.state.config.STT_OPENAI_API_BASE_URL}/audio/transcriptions",
+                headers=headers,
+                files=files,
+                data=data,
+            )
+
+            r.raise_for_status()
+
+            data = r.json()
+
+            # save the transcript to a json file
+            transcript_file = f"{file_dir}/{id}.json"
+            with open(transcript_file, "w") as f:
+                json.dump(data, f)
+
+            print(data)
+            return data
+        except Exception as e:
+            log.exception(e)
+            error_detail = "Open WebUI: Server Connection Error"
+            if r is not None:
+                try:
+                    res = r.json()
+                    if "error" in res:
+                        error_detail = f"External: {res['error']['message']}"
+                except Exception:
+                    error_detail = f"External: {e}"
+
+            raise Exception(error_detail)
+
+
+@app.post("/transcriptions")
+def transcription(
+    file: UploadFile = File(...),
+    user=Depends(get_verified_user),
+):
+    log.info(f"file.content_type: {file.content_type}")
+
+    if file.content_type not in ["audio/mpeg", "audio/wav", "audio/ogg", "audio/x-m4a"]:
+        raise HTTPException(
+            status_code=status.HTTP_400_BAD_REQUEST,
+            detail=ERROR_MESSAGES.FILE_NOT_SUPPORTED,
+        )
+
+    try:
+        ext = file.filename.split(".")[-1]
+        id = uuid.uuid4()
+
+        filename = f"{id}.{ext}"
+        contents = file.file.read()
+
+        file_dir = f"{CACHE_DIR}/audio/transcriptions"
+        os.makedirs(file_dir, exist_ok=True)
+        file_path = f"{file_dir}/{filename}"
+
+        with open(file_path, "wb") as f:
+            f.write(contents)
+
+        try:
+            if os.path.getsize(file_path) > MAX_FILE_SIZE:  # file is bigger than 25MB
+                log.debug(f"File size is larger than {MAX_FILE_SIZE_MB}MB")
+                audio = AudioSegment.from_file(file_path)
+                audio = audio.set_frame_rate(16000).set_channels(1)  # Compress audio
+                compressed_path = f"{file_dir}/{id}_compressed.opus"
+                audio.export(compressed_path, format="opus", bitrate="32k")
+                log.debug(f"Compressed audio to {compressed_path}")
+                file_path = compressed_path
+
+                if (
+                    os.path.getsize(file_path) > MAX_FILE_SIZE
+                ):  # Still larger than 25MB after compression
+                    log.debug(
+                        f"Compressed file size is still larger than {MAX_FILE_SIZE_MB}MB: {os.path.getsize(file_path)}"
+                    )
+                    raise HTTPException(
+                        status_code=status.HTTP_400_BAD_REQUEST,
+                        detail=ERROR_MESSAGES.FILE_TOO_LARGE(
+                            size=f"{MAX_FILE_SIZE_MB}MB"
+                        ),
+                    )
+
+                data = transcribe(file_path)
+            else:
+                data = transcribe(file_path)
+
+            file_path = file_path.split("/")[-1]
+            return {**data, "filename": file_path}
+        except Exception as e:
+            log.exception(e)
+            raise HTTPException(
+                status_code=status.HTTP_400_BAD_REQUEST,
+                detail=ERROR_MESSAGES.DEFAULT(e),
+            )
+
+    except Exception as e:
+        log.exception(e)
+
+        raise HTTPException(
+            status_code=status.HTTP_400_BAD_REQUEST,
+            detail=ERROR_MESSAGES.DEFAULT(e),
+        )
+
+
+def get_available_models() -> list[dict]:
+    if app.state.config.TTS_ENGINE == "openai":
+        return [{"id": "tts-1"}, {"id": "tts-1-hd"}]
+    elif app.state.config.TTS_ENGINE == "elevenlabs":
+        headers = {
+            "xi-api-key": app.state.config.TTS_API_KEY,
+            "Content-Type": "application/json",
+        }
+
+        try:
+            response = requests.get(
+                "https://api.elevenlabs.io/v1/models", headers=headers, timeout=5
+            )
+            response.raise_for_status()
+            models = response.json()
+            return [
+                {"name": model["name"], "id": model["model_id"]} for model in models
+            ]
+        except requests.RequestException as e:
+            log.error(f"Error fetching voices: {str(e)}")
+    return []
+
+
+@app.get("/models")
+async def get_models(user=Depends(get_verified_user)):
+    return {"models": get_available_models()}
+
+
+def get_available_voices() -> dict:
+    """Returns {voice_id: voice_name} dict"""
+    ret = {}
+    if app.state.config.TTS_ENGINE == "openai":
+        ret = {
+            "alloy": "alloy",
+            "echo": "echo",
+            "fable": "fable",
+            "onyx": "onyx",
+            "nova": "nova",
+            "shimmer": "shimmer",
+        }
+    elif app.state.config.TTS_ENGINE == "elevenlabs":
+        try:
+            ret = get_elevenlabs_voices()
+        except Exception:
+            # Avoided @lru_cache with exception
+            pass
+    elif app.state.config.TTS_ENGINE == "azure":
+        try:
+            region = app.state.config.TTS_AZURE_SPEECH_REGION
+            url = f"https://{region}.tts.speech.microsoft.com/cognitiveservices/voices/list"
+            headers = {"Ocp-Apim-Subscription-Key": app.state.config.TTS_API_KEY}
+
+            response = requests.get(url, headers=headers)
+            response.raise_for_status()
+            voices = response.json()
+            for voice in voices:
+                ret[voice["ShortName"]] = (
+                    f"{voice['DisplayName']} ({voice['ShortName']})"
+                )
+        except requests.RequestException as e:
+            log.error(f"Error fetching voices: {str(e)}")
+
+    return ret
+
+
+@lru_cache
+def get_elevenlabs_voices() -> dict:
+    """
+    Note, set the following in your .env file to use Elevenlabs:
+    AUDIO_TTS_ENGINE=elevenlabs
+    AUDIO_TTS_API_KEY=sk_...  # Your Elevenlabs API key
+    AUDIO_TTS_VOICE=EXAVITQu4vr4xnSDxMaL  # From https://api.elevenlabs.io/v1/voices
+    AUDIO_TTS_MODEL=eleven_multilingual_v2
+    """
+    headers = {
+        "xi-api-key": app.state.config.TTS_API_KEY,
+        "Content-Type": "application/json",
+    }
+    try:
+        # TODO: Add retries
+        response = requests.get("https://api.elevenlabs.io/v1/voices", headers=headers)
+        response.raise_for_status()
+        voices_data = response.json()
+
+        voices = {}
+        for voice in voices_data.get("voices", []):
+            voices[voice["voice_id"]] = voice["name"]
+    except requests.RequestException as e:
+        # Avoid @lru_cache with exception
+        log.error(f"Error fetching voices: {str(e)}")
+        raise RuntimeError(f"Error fetching voices: {str(e)}")
+
+    return voices
+
+
+@app.get("/voices")
+async def get_voices(user=Depends(get_verified_user)):
+    return {"voices": [{"id": k, "name": v} for k, v in get_available_voices().items()]}

backend/open_webui/apps/images/main.py

@@ -1,57 +1,46 @@
-import re
-import requests
-import base64
-from fastapi import (
-    FastAPI,
-    Request,
-    Depends,
-    HTTPException,
-    status,
-    UploadFile,
-    File,
-    Form,
-)
-from fastapi.middleware.cors import CORSMiddleware
-
-from constants import ERROR_MESSAGES
-from utils.utils import (
-    get_verified_user,
-    get_admin_user,
-)
-
-from apps.images.utils.comfyui import ImageGenerationPayload, comfyui_generate_image
-from utils.misc import calculate_sha256
-from typing import Optional
-from pydantic import BaseModel
-from pathlib import Path
-import mimetypes
-import uuid
+import asyncio
 import base64
 import json
 import logging
+import mimetypes
+import re
+import uuid
+from pathlib import Path
+from typing import Optional
 
-from config import (
-    SRC_LOG_LEVELS,
-    CACHE_DIR,
-    IMAGE_GENERATION_ENGINE,
-    ENABLE_IMAGE_GENERATION,
-    AUTOMATIC1111_BASE_URL,
+import requests
+from open_webui.apps.images.utils.comfyui import (
+    ComfyUIGenerateImageForm,
+    ComfyUIWorkflow,
+    comfyui_generate_image,
+)
+from open_webui.config import (
     AUTOMATIC1111_API_AUTH,
+    AUTOMATIC1111_BASE_URL,
+    AUTOMATIC1111_CFG_SCALE,
+    AUTOMATIC1111_SAMPLER,
+    AUTOMATIC1111_SCHEDULER,
+    CACHE_DIR,
     COMFYUI_BASE_URL,
-    COMFYUI_CFG_SCALE,
-    COMFYUI_SAMPLER,
-    COMFYUI_SCHEDULER,
-    COMFYUI_SD3,
-    COMFYUI_FLUX,
-    COMFYUI_FLUX_WEIGHT_DTYPE,
-    COMFYUI_FLUX_FP8_CLIP,
-    IMAGES_OPENAI_API_BASE_URL,
-    IMAGES_OPENAI_API_KEY,
+    COMFYUI_WORKFLOW,
+    COMFYUI_WORKFLOW_NODES,
+    CORS_ALLOW_ORIGIN,
+    ENABLE_IMAGE_GENERATION,
+    IMAGE_GENERATION_ENGINE,
     IMAGE_GENERATION_MODEL,
     IMAGE_SIZE,
     IMAGE_STEPS,
+    IMAGES_OPENAI_API_BASE_URL,
+    IMAGES_OPENAI_API_KEY,
     AppConfig,
 )
+from open_webui.constants import ERROR_MESSAGES
+from open_webui.env import ENV, SRC_LOG_LEVELS, ENABLE_FORWARD_USER_INFO_HEADERS
+
+from fastapi import Depends, FastAPI, HTTPException, Request
+from fastapi.middleware.cors import CORSMiddleware
+from pydantic import BaseModel
+from open_webui.utils.utils import get_admin_user, get_verified_user
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["IMAGES"])
@@ -59,10 +48,15 @@ log.setLevel(SRC_LOG_LEVELS["IMAGES"])
 IMAGE_CACHE_DIR = Path(CACHE_DIR).joinpath("./image/generations/")
 IMAGE_CACHE_DIR.mkdir(parents=True, exist_ok=True)
 
-app = FastAPI()
+app = FastAPI(
+    docs_url="/docs" if ENV == "dev" else None,
+    openapi_url="/openapi.json" if ENV == "dev" else None,
+    redoc_url=None,
+)
+
 app.add_middleware(
     CORSMiddleware,
-    allow_origins=["*"],
+    allow_origins=CORS_ALLOW_ORIGIN,
     allow_credentials=True,
     allow_methods=["*"],
     allow_headers=["*"],
@@ -80,21 +74,127 @@ app.state.config.MODEL = IMAGE_GENERATION_MODEL
 
 app.state.config.AUTOMATIC1111_BASE_URL = AUTOMATIC1111_BASE_URL
 app.state.config.AUTOMATIC1111_API_AUTH = AUTOMATIC1111_API_AUTH
+app.state.config.AUTOMATIC1111_CFG_SCALE = AUTOMATIC1111_CFG_SCALE
+app.state.config.AUTOMATIC1111_SAMPLER = AUTOMATIC1111_SAMPLER
+app.state.config.AUTOMATIC1111_SCHEDULER = AUTOMATIC1111_SCHEDULER
 app.state.config.COMFYUI_BASE_URL = COMFYUI_BASE_URL
+app.state.config.COMFYUI_WORKFLOW = COMFYUI_WORKFLOW
+app.state.config.COMFYUI_WORKFLOW_NODES = COMFYUI_WORKFLOW_NODES
 
 app.state.config.IMAGE_SIZE = IMAGE_SIZE
 app.state.config.IMAGE_STEPS = IMAGE_STEPS
-app.state.config.COMFYUI_CFG_SCALE = COMFYUI_CFG_SCALE
-app.state.config.COMFYUI_SAMPLER = COMFYUI_SAMPLER
-app.state.config.COMFYUI_SCHEDULER = COMFYUI_SCHEDULER
-app.state.config.COMFYUI_SD3 = COMFYUI_SD3
-app.state.config.COMFYUI_FLUX = COMFYUI_FLUX
-app.state.config.COMFYUI_FLUX_WEIGHT_DTYPE = COMFYUI_FLUX_WEIGHT_DTYPE
-app.state.config.COMFYUI_FLUX_FP8_CLIP = COMFYUI_FLUX_FP8_CLIP
+
+
+@app.get("/config")
+async def get_config(request: Request, user=Depends(get_admin_user)):
+    return {
+        "enabled": app.state.config.ENABLED,
+        "engine": app.state.config.ENGINE,
+        "openai": {
+            "OPENAI_API_BASE_URL": app.state.config.OPENAI_API_BASE_URL,
+            "OPENAI_API_KEY": app.state.config.OPENAI_API_KEY,
+        },
+        "automatic1111": {
+            "AUTOMATIC1111_BASE_URL": app.state.config.AUTOMATIC1111_BASE_URL,
+            "AUTOMATIC1111_API_AUTH": app.state.config.AUTOMATIC1111_API_AUTH,
+            "AUTOMATIC1111_CFG_SCALE": app.state.config.AUTOMATIC1111_CFG_SCALE,
+            "AUTOMATIC1111_SAMPLER": app.state.config.AUTOMATIC1111_SAMPLER,
+            "AUTOMATIC1111_SCHEDULER": app.state.config.AUTOMATIC1111_SCHEDULER,
+        },
+        "comfyui": {
+            "COMFYUI_BASE_URL": app.state.config.COMFYUI_BASE_URL,
+            "COMFYUI_WORKFLOW": app.state.config.COMFYUI_WORKFLOW,
+            "COMFYUI_WORKFLOW_NODES": app.state.config.COMFYUI_WORKFLOW_NODES,
+        },
+    }
+
+
+class OpenAIConfigForm(BaseModel):
+    OPENAI_API_BASE_URL: str
+    OPENAI_API_KEY: str
+
+
+class Automatic1111ConfigForm(BaseModel):
+    AUTOMATIC1111_BASE_URL: str
+    AUTOMATIC1111_API_AUTH: str
+    AUTOMATIC1111_CFG_SCALE: Optional[str]
+    AUTOMATIC1111_SAMPLER: Optional[str]
+    AUTOMATIC1111_SCHEDULER: Optional[str]
+
+
+class ComfyUIConfigForm(BaseModel):
+    COMFYUI_BASE_URL: str
+    COMFYUI_WORKFLOW: str
+    COMFYUI_WORKFLOW_NODES: list[dict]
+
+
+class ConfigForm(BaseModel):
+    enabled: bool
+    engine: str
+    openai: OpenAIConfigForm
+    automatic1111: Automatic1111ConfigForm
+    comfyui: ComfyUIConfigForm
+
+
+@app.post("/config/update")
+async def update_config(form_data: ConfigForm, user=Depends(get_admin_user)):
+    app.state.config.ENGINE = form_data.engine
+    app.state.config.ENABLED = form_data.enabled
+
+    app.state.config.OPENAI_API_BASE_URL = form_data.openai.OPENAI_API_BASE_URL
+    app.state.config.OPENAI_API_KEY = form_data.openai.OPENAI_API_KEY
+
+    app.state.config.AUTOMATIC1111_BASE_URL = (
+        form_data.automatic1111.AUTOMATIC1111_BASE_URL
+    )
+    app.state.config.AUTOMATIC1111_API_AUTH = (
+        form_data.automatic1111.AUTOMATIC1111_API_AUTH
+    )
+
+    app.state.config.AUTOMATIC1111_CFG_SCALE = (
+        float(form_data.automatic1111.AUTOMATIC1111_CFG_SCALE)
+        if form_data.automatic1111.AUTOMATIC1111_CFG_SCALE
+        else None
+    )
+    app.state.config.AUTOMATIC1111_SAMPLER = (
+        form_data.automatic1111.AUTOMATIC1111_SAMPLER
+        if form_data.automatic1111.AUTOMATIC1111_SAMPLER
+        else None
+    )
+    app.state.config.AUTOMATIC1111_SCHEDULER = (
+        form_data.automatic1111.AUTOMATIC1111_SCHEDULER
+        if form_data.automatic1111.AUTOMATIC1111_SCHEDULER
+        else None
+    )
+
+    app.state.config.COMFYUI_BASE_URL = form_data.comfyui.COMFYUI_BASE_URL.strip("/")
+    app.state.config.COMFYUI_WORKFLOW = form_data.comfyui.COMFYUI_WORKFLOW
+    app.state.config.COMFYUI_WORKFLOW_NODES = form_data.comfyui.COMFYUI_WORKFLOW_NODES
+
+    return {
+        "enabled": app.state.config.ENABLED,
+        "engine": app.state.config.ENGINE,
+        "openai": {
+            "OPENAI_API_BASE_URL": app.state.config.OPENAI_API_BASE_URL,
+            "OPENAI_API_KEY": app.state.config.OPENAI_API_KEY,
+        },
+        "automatic1111": {
+            "AUTOMATIC1111_BASE_URL": app.state.config.AUTOMATIC1111_BASE_URL,
+            "AUTOMATIC1111_API_AUTH": app.state.config.AUTOMATIC1111_API_AUTH,
+            "AUTOMATIC1111_CFG_SCALE": app.state.config.AUTOMATIC1111_CFG_SCALE,
+            "AUTOMATIC1111_SAMPLER": app.state.config.AUTOMATIC1111_SAMPLER,
+            "AUTOMATIC1111_SCHEDULER": app.state.config.AUTOMATIC1111_SCHEDULER,
+        },
+        "comfyui": {
+            "COMFYUI_BASE_URL": app.state.config.COMFYUI_BASE_URL,
+            "COMFYUI_WORKFLOW": app.state.config.COMFYUI_WORKFLOW,
+            "COMFYUI_WORKFLOW_NODES": app.state.config.COMFYUI_WORKFLOW_NODES,
+        },
+    }
 
 
 def get_automatic1111_api_auth():
-    if app.state.config.AUTOMATIC1111_API_AUTH == None:
+    if app.state.config.AUTOMATIC1111_API_AUTH is None:
         return ""
     else:
         auth1111_byte_string = app.state.config.AUTOMATIC1111_API_AUTH.encode("utf-8")
@@ -103,164 +203,112 @@ def get_automatic1111_api_auth():
         return f"Basic {auth1111_base64_encoded_string}"
 
 
-@app.get("/config")
-async def get_config(request: Request, user=Depends(get_admin_user)):
-    return {
-        "engine": app.state.config.ENGINE,
-        "enabled": app.state.config.ENABLED,
-    }
-
-
-class ConfigUpdateForm(BaseModel):
-    engine: str
-    enabled: bool
-
-
-@app.post("/config/update")
-async def update_config(form_data: ConfigUpdateForm, user=Depends(get_admin_user)):
-    app.state.config.ENGINE = form_data.engine
-    app.state.config.ENABLED = form_data.enabled
-    return {
-        "engine": app.state.config.ENGINE,
-        "enabled": app.state.config.ENABLED,
-    }
-
-
-class EngineUrlUpdateForm(BaseModel):
-    AUTOMATIC1111_BASE_URL: Optional[str] = None
-    AUTOMATIC1111_API_AUTH: Optional[str] = None
-    COMFYUI_BASE_URL: Optional[str] = None
-
-
-@app.get("/url")
-async def get_engine_url(user=Depends(get_admin_user)):
-    return {
-        "AUTOMATIC1111_BASE_URL": app.state.config.AUTOMATIC1111_BASE_URL,
-        "AUTOMATIC1111_API_AUTH": app.state.config.AUTOMATIC1111_API_AUTH,
-        "COMFYUI_BASE_URL": app.state.config.COMFYUI_BASE_URL,
-    }
-
-
-@app.post("/url/update")
-async def update_engine_url(
-    form_data: EngineUrlUpdateForm, user=Depends(get_admin_user)
-):
-    if form_data.AUTOMATIC1111_BASE_URL == None:
-        app.state.config.AUTOMATIC1111_BASE_URL = AUTOMATIC1111_BASE_URL
-    else:
-        url = form_data.AUTOMATIC1111_BASE_URL.strip("/")
+@app.get("/config/url/verify")
+async def verify_url(user=Depends(get_admin_user)):
+    if app.state.config.ENGINE == "automatic1111":
         try:
-            r = requests.head(url)
-            app.state.config.AUTOMATIC1111_BASE_URL = url
+            r = requests.get(
+                url=f"{app.state.config.AUTOMATIC1111_BASE_URL}/sdapi/v1/options",
+                headers={"authorization": get_automatic1111_api_auth()},
+            )
+            r.raise_for_status()
+            return True
+        except Exception:
+            app.state.config.ENABLED = False
+            raise HTTPException(status_code=400, detail=ERROR_MESSAGES.INVALID_URL)
+    elif app.state.config.ENGINE == "comfyui":
+        try:
+            r = requests.get(url=f"{app.state.config.COMFYUI_BASE_URL}/object_info")
+            r.raise_for_status()
+            return True
+        except Exception:
+            app.state.config.ENABLED = False
+            raise HTTPException(status_code=400, detail=ERROR_MESSAGES.INVALID_URL)
+    else:
+        return True
+
+
+def set_image_model(model: str):
+    log.info(f"Setting image model to {model}")
+    app.state.config.MODEL = model
+    if app.state.config.ENGINE in ["", "automatic1111"]:
+        api_auth = get_automatic1111_api_auth()
+        r = requests.get(
+            url=f"{app.state.config.AUTOMATIC1111_BASE_URL}/sdapi/v1/options",
+            headers={"authorization": api_auth},
+        )
+        options = r.json()
+        if model != options["sd_model_checkpoint"]:
+            options["sd_model_checkpoint"] = model
+            r = requests.post(
+                url=f"{app.state.config.AUTOMATIC1111_BASE_URL}/sdapi/v1/options",
+                json=options,
+                headers={"authorization": api_auth},
+            )
+    return app.state.config.MODEL
+
+
+def get_image_model():
+    if app.state.config.ENGINE == "openai":
+        return app.state.config.MODEL if app.state.config.MODEL else "dall-e-2"
+    elif app.state.config.ENGINE == "comfyui":
+        return app.state.config.MODEL if app.state.config.MODEL else ""
+    elif app.state.config.ENGINE == "automatic1111" or app.state.config.ENGINE == "":
+        try:
+            r = requests.get(
+                url=f"{app.state.config.AUTOMATIC1111_BASE_URL}/sdapi/v1/options",
+                headers={"authorization": get_automatic1111_api_auth()},
+            )
+            options = r.json()
+            return options["sd_model_checkpoint"]
         except Exception as e:
+            app.state.config.ENABLED = False
             raise HTTPException(status_code=400, detail=ERROR_MESSAGES.DEFAULT(e))
 
-    if form_data.COMFYUI_BASE_URL == None:
-        app.state.config.COMFYUI_BASE_URL = COMFYUI_BASE_URL
-    else:
-        url = form_data.COMFYUI_BASE_URL.strip("/")
 
-        try:
-            r = requests.head(url)
-            app.state.config.COMFYUI_BASE_URL = url
-        except Exception as e:
-            raise HTTPException(status_code=400, detail=ERROR_MESSAGES.DEFAULT(e))
+class ImageConfigForm(BaseModel):
+    MODEL: str
+    IMAGE_SIZE: str
+    IMAGE_STEPS: int
 
-    if form_data.AUTOMATIC1111_API_AUTH == None:
-        app.state.config.AUTOMATIC1111_API_AUTH = AUTOMATIC1111_API_AUTH
-    else:
-        app.state.config.AUTOMATIC1111_API_AUTH = form_data.AUTOMATIC1111_API_AUTH
 
+@app.get("/image/config")
+async def get_image_config(user=Depends(get_admin_user)):
     return {
-        "AUTOMATIC1111_BASE_URL": app.state.config.AUTOMATIC1111_BASE_URL,
-        "AUTOMATIC1111_API_AUTH": app.state.config.AUTOMATIC1111_API_AUTH,
-        "COMFYUI_BASE_URL": app.state.config.COMFYUI_BASE_URL,
-        "status": True,
+        "MODEL": app.state.config.MODEL,
+        "IMAGE_SIZE": app.state.config.IMAGE_SIZE,
+        "IMAGE_STEPS": app.state.config.IMAGE_STEPS,
     }
 
 
-class OpenAIConfigUpdateForm(BaseModel):
-    url: str
-    key: str
+@app.post("/image/config/update")
+async def update_image_config(form_data: ImageConfigForm, user=Depends(get_admin_user)):
 
+    set_image_model(form_data.MODEL)
 
-@app.get("/openai/config")
-async def get_openai_config(user=Depends(get_admin_user)):
-    return {
-        "OPENAI_API_BASE_URL": app.state.config.OPENAI_API_BASE_URL,
-        "OPENAI_API_KEY": app.state.config.OPENAI_API_KEY,
-    }
-
-
-@app.post("/openai/config/update")
-async def update_openai_config(
-    form_data: OpenAIConfigUpdateForm, user=Depends(get_admin_user)
-):
-    if form_data.key == "":
-        raise HTTPException(status_code=400, detail=ERROR_MESSAGES.API_KEY_NOT_FOUND)
-
-    app.state.config.OPENAI_API_BASE_URL = form_data.url
-    app.state.config.OPENAI_API_KEY = form_data.key
-
-    return {
-        "status": True,
-        "OPENAI_API_BASE_URL": app.state.config.OPENAI_API_BASE_URL,
-        "OPENAI_API_KEY": app.state.config.OPENAI_API_KEY,
-    }
-
-
-class ImageSizeUpdateForm(BaseModel):
-    size: str
-
-
-@app.get("/size")
-async def get_image_size(user=Depends(get_admin_user)):
-    return {"IMAGE_SIZE": app.state.config.IMAGE_SIZE}
-
-
-@app.post("/size/update")
-async def update_image_size(
-    form_data: ImageSizeUpdateForm, user=Depends(get_admin_user)
-):
-    pattern = r"^\d+x\d+$"  # Regular expression pattern
-    if re.match(pattern, form_data.size):
-        app.state.config.IMAGE_SIZE = form_data.size
-        return {
-            "IMAGE_SIZE": app.state.config.IMAGE_SIZE,
-            "status": True,
-        }
+    pattern = r"^\d+x\d+$"
+    if re.match(pattern, form_data.IMAGE_SIZE):
+        app.state.config.IMAGE_SIZE = form_data.IMAGE_SIZE
     else:
         raise HTTPException(
             status_code=400,
             detail=ERROR_MESSAGES.INCORRECT_FORMAT("  (e.g., 512x512)."),
         )
 
-
-class ImageStepsUpdateForm(BaseModel):
-    steps: int
-
-
-@app.get("/steps")
-async def get_image_size(user=Depends(get_admin_user)):
-    return {"IMAGE_STEPS": app.state.config.IMAGE_STEPS}
-
-
-@app.post("/steps/update")
-async def update_image_size(
-    form_data: ImageStepsUpdateForm, user=Depends(get_admin_user)
-):
-    if form_data.steps >= 0:
-        app.state.config.IMAGE_STEPS = form_data.steps
-        return {
-            "IMAGE_STEPS": app.state.config.IMAGE_STEPS,
-            "status": True,
-        }
+    if form_data.IMAGE_STEPS >= 0:
+        app.state.config.IMAGE_STEPS = form_data.IMAGE_STEPS
     else:
         raise HTTPException(
             status_code=400,
             detail=ERROR_MESSAGES.INCORRECT_FORMAT("  (e.g., 50)."),
         )
 
+    return {
+        "MODEL": app.state.config.MODEL,
+        "IMAGE_SIZE": app.state.config.IMAGE_SIZE,
+        "IMAGE_STEPS": app.state.config.IMAGE_STEPS,
+    }
+
 
 @app.get("/models")
 def get_models(user=Depends(get_verified_user)):
@@ -271,18 +319,51 @@ def get_models(user=Depends(get_verified_user)):
                 {"id": "dall-e-3", "name": "DALL·E 3"},
             ]
         elif app.state.config.ENGINE == "comfyui":
-
+            # TODO - get models from comfyui
             r = requests.get(url=f"{app.state.config.COMFYUI_BASE_URL}/object_info")
             info = r.json()
 
-            return list(
-                map(
-                    lambda model: {"id": model, "name": model},
-                    info["CheckpointLoaderSimple"]["input"]["required"]["ckpt_name"][0],
-                )
-            )
+            workflow = json.loads(app.state.config.COMFYUI_WORKFLOW)
+            model_node_id = None
 
-        else:
+            for node in app.state.config.COMFYUI_WORKFLOW_NODES:
+                if node["type"] == "model":
+                    if node["node_ids"]:
+                        model_node_id = node["node_ids"][0]
+                    break
+
+            if model_node_id:
+                model_list_key = None
+
+                print(workflow[model_node_id]["class_type"])
+                for key in info[workflow[model_node_id]["class_type"]]["input"][
+                    "required"
+                ]:
+                    if "_name" in key:
+                        model_list_key = key
+                        break
+
+                if model_list_key:
+                    return list(
+                        map(
+                            lambda model: {"id": model, "name": model},
+                            info[workflow[model_node_id]["class_type"]]["input"][
+                                "required"
+                            ][model_list_key][0],
+                        )
+                    )
+            else:
+                return list(
+                    map(
+                        lambda model: {"id": model, "name": model},
+                        info["CheckpointLoaderSimple"]["input"]["required"][
+                            "ckpt_name"
+                        ][0],
+                    )
+                )
+        elif (
+            app.state.config.ENGINE == "automatic1111" or app.state.config.ENGINE == ""
+        ):
             r = requests.get(
                 url=f"{app.state.config.AUTOMATIC1111_BASE_URL}/sdapi/v1/sd-models",
                 headers={"authorization": get_automatic1111_api_auth()},
@@ -299,69 +380,11 @@ def get_models(user=Depends(get_verified_user)):
         raise HTTPException(status_code=400, detail=ERROR_MESSAGES.DEFAULT(e))
 
 
-@app.get("/models/default")
-async def get_default_model(user=Depends(get_admin_user)):
-    try:
-        if app.state.config.ENGINE == "openai":
-            return {
-                "model": (
-                    app.state.config.MODEL if app.state.config.MODEL else "dall-e-2"
-                )
-            }
-        elif app.state.config.ENGINE == "comfyui":
-            return {"model": (app.state.config.MODEL if app.state.config.MODEL else "")}
-        else:
-            r = requests.get(
-                url=f"{app.state.config.AUTOMATIC1111_BASE_URL}/sdapi/v1/options",
-                headers={"authorization": get_automatic1111_api_auth()},
-            )
-            options = r.json()
-            return {"model": options["sd_model_checkpoint"]}
-    except Exception as e:
-        app.state.config.ENABLED = False
-        raise HTTPException(status_code=400, detail=ERROR_MESSAGES.DEFAULT(e))
-
-
-class UpdateModelForm(BaseModel):
-    model: str
-
-
-def set_model_handler(model: str):
-    if app.state.config.ENGINE in ["openai", "comfyui"]:
-        app.state.config.MODEL = model
-        return app.state.config.MODEL
-    else:
-        api_auth = get_automatic1111_api_auth()
-        r = requests.get(
-            url=f"{app.state.config.AUTOMATIC1111_BASE_URL}/sdapi/v1/options",
-            headers={"authorization": api_auth},
-        )
-        options = r.json()
-
-        if model != options["sd_model_checkpoint"]:
-            options["sd_model_checkpoint"] = model
-            r = requests.post(
-                url=f"{app.state.config.AUTOMATIC1111_BASE_URL}/sdapi/v1/options",
-                json=options,
-                headers={"authorization": api_auth},
-            )
-
-        return options
-
-
-@app.post("/models/default/update")
-def update_default_model(
-    form_data: UpdateModelForm,
-    user=Depends(get_verified_user),
-):
-    return set_model_handler(form_data.model)
-
-
 class GenerateImageForm(BaseModel):
     model: Optional[str] = None
     prompt: str
-    n: int = 1
     size: Optional[str] = None
+    n: int = 1
     negative_prompt: Optional[str] = None
 
 
@@ -403,7 +426,6 @@ def save_url_image(url):
         r = requests.get(url)
         r.raise_for_status()
         if r.headers["content-type"].split("/")[0] == "image":
-
             mime_type = r.headers["content-type"]
             image_format = mimetypes.guess_extension(mime_type)
 
@@ -418,7 +440,7 @@ def save_url_image(url):
                     image_file.write(chunk)
             return image_filename
         else:
-            log.error(f"Url does not point to an image.")
+            log.error("Url does not point to an image.")
             return None
 
     except Exception as e:
@@ -436,11 +458,16 @@ async def image_generations(
     r = None
     try:
         if app.state.config.ENGINE == "openai":
-
             headers = {}
             headers["Authorization"] = f"Bearer {app.state.config.OPENAI_API_KEY}"
             headers["Content-Type"] = "application/json"
 
+            if ENABLE_FORWARD_USER_INFO_HEADERS:
+                headers["X-OpenWebUI-User-Name"] = user.name
+                headers["X-OpenWebUI-User-Id"] = user.id
+                headers["X-OpenWebUI-User-Email"] = user.email
+                headers["X-OpenWebUI-User-Role"] = user.role
+
             data = {
                 "model": (
                     app.state.config.MODEL
@@ -455,7 +482,9 @@ async def image_generations(
                 "response_format": "b64_json",
             }
 
-            r = requests.post(
+            # Use asyncio.to_thread for the requests.post call
+            r = await asyncio.to_thread(
+                requests.post,
                 url=f"{app.state.config.OPENAI_API_BASE_URL}/images/generations",
                 json=data,
                 headers=headers,
@@ -477,7 +506,6 @@ async def image_generations(
             return images
 
         elif app.state.config.ENGINE == "comfyui":
-
             data = {
                 "prompt": form_data.prompt,
                 "width": width,
@@ -491,32 +519,20 @@ async def image_generations(
             if form_data.negative_prompt is not None:
                 data["negative_prompt"] = form_data.negative_prompt
 
-            if app.state.config.COMFYUI_CFG_SCALE:
-                data["cfg_scale"] = app.state.config.COMFYUI_CFG_SCALE
-
-            if app.state.config.COMFYUI_SAMPLER is not None:
-                data["sampler"] = app.state.config.COMFYUI_SAMPLER
-
-            if app.state.config.COMFYUI_SCHEDULER is not None:
-                data["scheduler"] = app.state.config.COMFYUI_SCHEDULER
-
-            if app.state.config.COMFYUI_SD3 is not None:
-                data["sd3"] = app.state.config.COMFYUI_SD3
-
-            if app.state.config.COMFYUI_FLUX is not None:
-                data["flux"] = app.state.config.COMFYUI_FLUX
-
-            if app.state.config.COMFYUI_FLUX_WEIGHT_DTYPE is not None:
-                data["flux_weight_dtype"] = app.state.config.COMFYUI_FLUX_WEIGHT_DTYPE
-
-            if app.state.config.COMFYUI_FLUX_FP8_CLIP is not None:
-                data["flux_fp8_clip"] = app.state.config.COMFYUI_FLUX_FP8_CLIP
-
-            data = ImageGenerationPayload(**data)
-
-            res = comfyui_generate_image(
+            form_data = ComfyUIGenerateImageForm(
+                **{
+                    "workflow": ComfyUIWorkflow(
+                        **{
+                            "workflow": app.state.config.COMFYUI_WORKFLOW,
+                            "nodes": app.state.config.COMFYUI_WORKFLOW_NODES,
+                        }
+                    ),
+                    **data,
+                }
+            )
+            res = await comfyui_generate_image(
                 app.state.config.MODEL,
-                data,
+                form_data,
                 user.id,
                 app.state.config.COMFYUI_BASE_URL,
             )
@@ -530,13 +546,15 @@ async def image_generations(
                 file_body_path = IMAGE_CACHE_DIR.joinpath(f"{image_filename}.json")
 
                 with open(file_body_path, "w") as f:
-                    json.dump(data.model_dump(exclude_none=True), f)
+                    json.dump(form_data.model_dump(exclude_none=True), f)
 
             log.debug(f"images: {images}")
             return images
-        else:
+        elif (
+            app.state.config.ENGINE == "automatic1111" or app.state.config.ENGINE == ""
+        ):
             if form_data.model:
-                set_model_handler(form_data.model)
+                set_image_model(form_data.model)
 
             data = {
                 "prompt": form_data.prompt,
@@ -551,14 +569,24 @@ async def image_generations(
             if form_data.negative_prompt is not None:
                 data["negative_prompt"] = form_data.negative_prompt
 
-            r = requests.post(
+            if app.state.config.AUTOMATIC1111_CFG_SCALE:
+                data["cfg_scale"] = app.state.config.AUTOMATIC1111_CFG_SCALE
+
+            if app.state.config.AUTOMATIC1111_SAMPLER:
+                data["sampler_name"] = app.state.config.AUTOMATIC1111_SAMPLER
+
+            if app.state.config.AUTOMATIC1111_SCHEDULER:
+                data["scheduler"] = app.state.config.AUTOMATIC1111_SCHEDULER
+
+            # Use asyncio.to_thread for the requests.post call
+            r = await asyncio.to_thread(
+                requests.post,
                 url=f"{app.state.config.AUTOMATIC1111_BASE_URL}/sdapi/v1/txt2img",
                 json=data,
                 headers={"authorization": get_automatic1111_api_auth()},
             )
 
             res = r.json()
-
             log.debug(f"res: {res}")
 
             images = []
@@ -572,10 +600,8 @@ async def image_generations(
                     json.dump({**data, "info": res["info"]}, f)
 
             return images
-
     except Exception as e:
         error = e
-
         if r != None:
             data = r.json()
             if "error" in data:

backend/open_webui/apps/images/utils/comfyui.py

@@ -0,0 +1,186 @@
+import asyncio
+import json
+import logging
+import random
+import urllib.parse
+import urllib.request
+from typing import Optional
+
+import websocket  # NOTE: websocket-client (https://github.com/websocket-client/websocket-client)
+from open_webui.env import SRC_LOG_LEVELS
+from pydantic import BaseModel
+
+log = logging.getLogger(__name__)
+log.setLevel(SRC_LOG_LEVELS["COMFYUI"])
+
+default_headers = {"User-Agent": "Mozilla/5.0"}
+
+
+def queue_prompt(prompt, client_id, base_url):
+    log.info("queue_prompt")
+    p = {"prompt": prompt, "client_id": client_id}
+    data = json.dumps(p).encode("utf-8")
+    log.debug(f"queue_prompt data: {data}")
+    try:
+        req = urllib.request.Request(
+            f"{base_url}/prompt", data=data, headers=default_headers
+        )
+        response = urllib.request.urlopen(req).read()
+        return json.loads(response)
+    except Exception as e:
+        log.exception(f"Error while queuing prompt: {e}")
+        raise e
+
+
+def get_image(filename, subfolder, folder_type, base_url):
+    log.info("get_image")
+    data = {"filename": filename, "subfolder": subfolder, "type": folder_type}
+    url_values = urllib.parse.urlencode(data)
+    req = urllib.request.Request(
+        f"{base_url}/view?{url_values}", headers=default_headers
+    )
+    with urllib.request.urlopen(req) as response:
+        return response.read()
+
+
+def get_image_url(filename, subfolder, folder_type, base_url):
+    log.info("get_image")
+    data = {"filename": filename, "subfolder": subfolder, "type": folder_type}
+    url_values = urllib.parse.urlencode(data)
+    return f"{base_url}/view?{url_values}"
+
+
+def get_history(prompt_id, base_url):
+    log.info("get_history")
+
+    req = urllib.request.Request(
+        f"{base_url}/history/{prompt_id}", headers=default_headers
+    )
+    with urllib.request.urlopen(req) as response:
+        return json.loads(response.read())
+
+
+def get_images(ws, prompt, client_id, base_url):
+    prompt_id = queue_prompt(prompt, client_id, base_url)["prompt_id"]
+    output_images = []
+    while True:
+        out = ws.recv()
+        if isinstance(out, str):
+            message = json.loads(out)
+            if message["type"] == "executing":
+                data = message["data"]
+                if data["node"] is None and data["prompt_id"] == prompt_id:
+                    break  # Execution is done
+        else:
+            continue  # previews are binary data
+
+    history = get_history(prompt_id, base_url)[prompt_id]
+    for o in history["outputs"]:
+        for node_id in history["outputs"]:
+            node_output = history["outputs"][node_id]
+            if "images" in node_output:
+                for image in node_output["images"]:
+                    url = get_image_url(
+                        image["filename"], image["subfolder"], image["type"], base_url
+                    )
+                    output_images.append({"url": url})
+    return {"data": output_images}
+
+
+class ComfyUINodeInput(BaseModel):
+    type: Optional[str] = None
+    node_ids: list[str] = []
+    key: Optional[str] = "text"
+    value: Optional[str] = None
+
+
+class ComfyUIWorkflow(BaseModel):
+    workflow: str
+    nodes: list[ComfyUINodeInput]
+
+
+class ComfyUIGenerateImageForm(BaseModel):
+    workflow: ComfyUIWorkflow
+
+    prompt: str
+    negative_prompt: Optional[str] = None
+    width: int
+    height: int
+    n: int = 1
+
+    steps: Optional[int] = None
+    seed: Optional[int] = None
+
+
+async def comfyui_generate_image(
+    model: str, payload: ComfyUIGenerateImageForm, client_id, base_url
+):
+    ws_url = base_url.replace("http://", "ws://").replace("https://", "wss://")
+    workflow = json.loads(payload.workflow.workflow)
+
+    for node in payload.workflow.nodes:
+        if node.type:
+            if node.type == "model":
+                for node_id in node.node_ids:
+                    workflow[node_id]["inputs"][node.key] = model
+            elif node.type == "prompt":
+                for node_id in node.node_ids:
+                    workflow[node_id]["inputs"][
+                        node.key if node.key else "text"
+                    ] = payload.prompt
+            elif node.type == "negative_prompt":
+                for node_id in node.node_ids:
+                    workflow[node_id]["inputs"][
+                        node.key if node.key else "text"
+                    ] = payload.negative_prompt
+            elif node.type == "width":
+                for node_id in node.node_ids:
+                    workflow[node_id]["inputs"][
+                        node.key if node.key else "width"
+                    ] = payload.width
+            elif node.type == "height":
+                for node_id in node.node_ids:
+                    workflow[node_id]["inputs"][
+                        node.key if node.key else "height"
+                    ] = payload.height
+            elif node.type == "n":
+                for node_id in node.node_ids:
+                    workflow[node_id]["inputs"][
+                        node.key if node.key else "batch_size"
+                    ] = payload.n
+            elif node.type == "steps":
+                for node_id in node.node_ids:
+                    workflow[node_id]["inputs"][
+                        node.key if node.key else "steps"
+                    ] = payload.steps
+            elif node.type == "seed":
+                seed = (
+                    payload.seed
+                    if payload.seed
+                    else random.randint(0, 18446744073709551614)
+                )
+                for node_id in node.node_ids:
+                    workflow[node_id]["inputs"][node.key] = seed
+        else:
+            for node_id in node.node_ids:
+                workflow[node_id]["inputs"][node.key] = node.value
+
+    try:
+        ws = websocket.WebSocket()
+        ws.connect(f"{ws_url}/ws?clientId={client_id}")
+        log.info("WebSocket connection established.")
+    except Exception as e:
+        log.exception(f"Failed to connect to WebSocket server: {e}")
+        return None
+
+    try:
+        log.info("Sending workflow to WebSocket server.")
+        log.info(f"Workflow: {workflow}")
+        images = await asyncio.to_thread(get_images, ws, workflow, client_id, base_url)
+    except Exception as e:
+        log.exception(f"Error while receiving images: {e}")
+        images = None
+
+    ws.close()
+
+    return images

backend/open_webui/apps/openai/main.py

@@ -0,0 +1,718 @@
+import asyncio
+import hashlib
+import json
+import logging
+from pathlib import Path
+from typing import Literal, Optional, overload
+
+import aiohttp
+from aiocache import cached
+import requests
+
+
+from open_webui.apps.webui.models.models import Models
+from open_webui.config import (
+    CACHE_DIR,
+    CORS_ALLOW_ORIGIN,
+    ENABLE_OPENAI_API,
+    OPENAI_API_BASE_URLS,
+    OPENAI_API_KEYS,
+    OPENAI_API_CONFIGS,
+    AppConfig,
+)
+from open_webui.env import (
+    AIOHTTP_CLIENT_TIMEOUT,
+    AIOHTTP_CLIENT_TIMEOUT_OPENAI_MODEL_LIST,
+    ENABLE_FORWARD_USER_INFO_HEADERS,
+)
+
+from open_webui.constants import ERROR_MESSAGES
+from open_webui.env import ENV, SRC_LOG_LEVELS
+from fastapi import Depends, FastAPI, HTTPException, Request
+from fastapi.middleware.cors import CORSMiddleware
+from fastapi.responses import FileResponse, StreamingResponse
+from pydantic import BaseModel
+from starlette.background import BackgroundTask
+
+from open_webui.utils.payload import (
+    apply_model_params_to_body_openai,
+    apply_model_system_prompt_to_body,
+)
+
+from open_webui.utils.utils import get_admin_user, get_verified_user
+from open_webui.utils.access_control import has_access
+
+
+log = logging.getLogger(__name__)
+log.setLevel(SRC_LOG_LEVELS["OPENAI"])
+
+
+app = FastAPI(
+    docs_url="/docs" if ENV == "dev" else None,
+    openapi_url="/openapi.json" if ENV == "dev" else None,
+    redoc_url=None,
+)
+
+
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=CORS_ALLOW_ORIGIN,
+    allow_credentials=True,
+    allow_methods=["*"],
+    allow_headers=["*"],
+)
+
+app.state.config = AppConfig()
+
+app.state.config.ENABLE_OPENAI_API = ENABLE_OPENAI_API
+app.state.config.OPENAI_API_BASE_URLS = OPENAI_API_BASE_URLS
+app.state.config.OPENAI_API_KEYS = OPENAI_API_KEYS
+app.state.config.OPENAI_API_CONFIGS = OPENAI_API_CONFIGS
+
+
+@app.get("/config")
+async def get_config(user=Depends(get_admin_user)):
+    return {
+        "ENABLE_OPENAI_API": app.state.config.ENABLE_OPENAI_API,
+        "OPENAI_API_BASE_URLS": app.state.config.OPENAI_API_BASE_URLS,
+        "OPENAI_API_KEYS": app.state.config.OPENAI_API_KEYS,
+        "OPENAI_API_CONFIGS": app.state.config.OPENAI_API_CONFIGS,
+    }
+
+
+class OpenAIConfigForm(BaseModel):
+    ENABLE_OPENAI_API: Optional[bool] = None
+    OPENAI_API_BASE_URLS: list[str]
+    OPENAI_API_KEYS: list[str]
+    OPENAI_API_CONFIGS: dict
+
+
+@app.post("/config/update")
+async def update_config(form_data: OpenAIConfigForm, user=Depends(get_admin_user)):
+    app.state.config.ENABLE_OPENAI_API = form_data.ENABLE_OPENAI_API
+
+    app.state.config.OPENAI_API_BASE_URLS = form_data.OPENAI_API_BASE_URLS
+    app.state.config.OPENAI_API_KEYS = form_data.OPENAI_API_KEYS
+
+    # Check if API KEYS length is same than API URLS length
+    if len(app.state.config.OPENAI_API_KEYS) != len(
+        app.state.config.OPENAI_API_BASE_URLS
+    ):
+        if len(app.state.config.OPENAI_API_KEYS) > len(
+            app.state.config.OPENAI_API_BASE_URLS
+        ):
+            app.state.config.OPENAI_API_KEYS = app.state.config.OPENAI_API_KEYS[
+                : len(app.state.config.OPENAI_API_BASE_URLS)
+            ]
+        else:
+            app.state.config.OPENAI_API_KEYS += [""] * (
+                len(app.state.config.OPENAI_API_BASE_URLS)
+                - len(app.state.config.OPENAI_API_KEYS)
+            )
+
+    app.state.config.OPENAI_API_CONFIGS = form_data.OPENAI_API_CONFIGS
+
+    # Remove any extra configs
+    config_urls = app.state.config.OPENAI_API_CONFIGS.keys()
+    for idx, url in enumerate(app.state.config.OPENAI_API_BASE_URLS):
+        if url not in config_urls:
+            app.state.config.OPENAI_API_CONFIGS.pop(url, None)
+
+    return {
+        "ENABLE_OPENAI_API": app.state.config.ENABLE_OPENAI_API,
+        "OPENAI_API_BASE_URLS": app.state.config.OPENAI_API_BASE_URLS,
+        "OPENAI_API_KEYS": app.state.config.OPENAI_API_KEYS,
+        "OPENAI_API_CONFIGS": app.state.config.OPENAI_API_CONFIGS,
+    }
+
+
+@app.post("/audio/speech")
+async def speech(request: Request, user=Depends(get_verified_user)):
+    idx = None
+    try:
+        idx = app.state.config.OPENAI_API_BASE_URLS.index("https://api.openai.com/v1")
+        body = await request.body()
+        name = hashlib.sha256(body).hexdigest()
+
+        SPEECH_CACHE_DIR = Path(CACHE_DIR).joinpath("./audio/speech/")
+        SPEECH_CACHE_DIR.mkdir(parents=True, exist_ok=True)
+        file_path = SPEECH_CACHE_DIR.joinpath(f"{name}.mp3")
+        file_body_path = SPEECH_CACHE_DIR.joinpath(f"{name}.json")
+
+        # Check if the file already exists in the cache
+        if file_path.is_file():
+            return FileResponse(file_path)
+
+        headers = {}
+        headers["Authorization"] = f"Bearer {app.state.config.OPENAI_API_KEYS[idx]}"
+        headers["Content-Type"] = "application/json"
+        if "openrouter.ai" in app.state.config.OPENAI_API_BASE_URLS[idx]:
+            headers["HTTP-Referer"] = "https://openwebui.com/"
+            headers["X-Title"] = "Open WebUI"
+        if ENABLE_FORWARD_USER_INFO_HEADERS:
+            headers["X-OpenWebUI-User-Name"] = user.name
+            headers["X-OpenWebUI-User-Id"] = user.id
+            headers["X-OpenWebUI-User-Email"] = user.email
+            headers["X-OpenWebUI-User-Role"] = user.role
+        r = None
+        try:
+            r = requests.post(
+                url=f"{app.state.config.OPENAI_API_BASE_URLS[idx]}/audio/speech",
+                data=body,
+                headers=headers,
+                stream=True,
+            )
+
+            r.raise_for_status()
+
+            # Save the streaming content to a file
+            with open(file_path, "wb") as f:
+                for chunk in r.iter_content(chunk_size=8192):
+                    f.write(chunk)
+
+            with open(file_body_path, "w") as f:
+                json.dump(json.loads(body.decode("utf-8")), f)
+
+            # Return the saved file
+            return FileResponse(file_path)
+
+        except Exception as e:
+            log.exception(e)
+            error_detail = "Open WebUI: Server Connection Error"
+            if r is not None:
+                try:
+                    res = r.json()
+                    if "error" in res:
+                        error_detail = f"External: {res['error']}"
+                except Exception:
+                    error_detail = f"External: {e}"
+
+            raise HTTPException(
+                status_code=r.status_code if r else 500, detail=error_detail
+            )
+
+    except ValueError:
+        raise HTTPException(status_code=401, detail=ERROR_MESSAGES.OPENAI_NOT_FOUND)
+
+
+async def aiohttp_get(url, key=None):
+    timeout = aiohttp.ClientTimeout(total=AIOHTTP_CLIENT_TIMEOUT_OPENAI_MODEL_LIST)
+    try:
+        headers = {"Authorization": f"Bearer {key}"} if key else {}
+        async with aiohttp.ClientSession(timeout=timeout, trust_env=True) as session:
+            async with session.get(url, headers=headers) as response:
+                return await response.json()
+    except Exception as e:
+        # Handle connection error here
+        log.error(f"Connection error: {e}")
+        return None
+
+
+async def cleanup_response(
+    response: Optional[aiohttp.ClientResponse],
+    session: Optional[aiohttp.ClientSession],
+):
+    if response:
+        response.close()
+    if session:
+        await session.close()
+
+
+def merge_models_lists(model_lists):
+    log.debug(f"merge_models_lists {model_lists}")
+    merged_list = []
+
+    for idx, models in enumerate(model_lists):
+        if models is not None and "error" not in models:
+            merged_list.extend(
+                [
+                    {
+                        **model,
+                        "name": model.get("name", model["id"]),
+                        "owned_by": "openai",
+                        "openai": model,
+                        "urlIdx": idx,
+                    }
+                    for model in models
+                    if "api.openai.com"
+                    not in app.state.config.OPENAI_API_BASE_URLS[idx]
+                    or not any(
+                        name in model["id"]
+                        for name in [
+                            "babbage",
+                            "dall-e",
+                            "davinci",
+                            "embedding",
+                            "tts",
+                            "whisper",
+                        ]
+                    )
+                ]
+            )
+
+    return merged_list
+
+
+async def get_all_models_responses() -> list:
+    if not app.state.config.ENABLE_OPENAI_API:
+        return []
+
+    # Check if API KEYS length is same than API URLS length
+    num_urls = len(app.state.config.OPENAI_API_BASE_URLS)
+    num_keys = len(app.state.config.OPENAI_API_KEYS)
+
+    if num_keys != num_urls:
+        # if there are more keys than urls, remove the extra keys
+        if num_keys > num_urls:
+            new_keys = app.state.config.OPENAI_API_KEYS[:num_urls]
+            app.state.config.OPENAI_API_KEYS = new_keys
+        # if there are more urls than keys, add empty keys
+        else:
+            app.state.config.OPENAI_API_KEYS += [""] * (num_urls - num_keys)
+
+    tasks = []
+    for idx, url in enumerate(app.state.config.OPENAI_API_BASE_URLS):
+        if url not in app.state.config.OPENAI_API_CONFIGS:
+            tasks.append(
+                aiohttp_get(f"{url}/models", app.state.config.OPENAI_API_KEYS[idx])
+            )
+        else:
+            api_config = app.state.config.OPENAI_API_CONFIGS.get(url, {})
+
+            enable = api_config.get("enable", True)
+            model_ids = api_config.get("model_ids", [])
+
+            if enable:
+                if len(model_ids) == 0:
+                    tasks.append(
+                        aiohttp_get(
+                            f"{url}/models", app.state.config.OPENAI_API_KEYS[idx]
+                        )
+                    )
+                else:
+                    model_list = {
+                        "object": "list",
+                        "data": [
+                            {
+                                "id": model_id,
+                                "name": model_id,
+                                "owned_by": "openai",
+                                "openai": {"id": model_id},
+                                "urlIdx": idx,
+                            }
+                            for model_id in model_ids
+                        ],
+                    }
+
+                    tasks.append(asyncio.ensure_future(asyncio.sleep(0, model_list)))
+            else:
+                tasks.append(asyncio.ensure_future(asyncio.sleep(0, None)))
+
+    responses = await asyncio.gather(*tasks)
+
+    for idx, response in enumerate(responses):
+        if response:
+            url = app.state.config.OPENAI_API_BASE_URLS[idx]
+            api_config = app.state.config.OPENAI_API_CONFIGS.get(url, {})
+
+            prefix_id = api_config.get("prefix_id", None)
+
+            if prefix_id:
+                for model in (
+                    response if isinstance(response, list) else response.get("data", [])
+                ):
+                    model["id"] = f"{prefix_id}.{model['id']}"
+
+    log.debug(f"get_all_models:responses() {responses}")
+
+    return responses
+
+
+@cached(ttl=3)
+async def get_all_models() -> dict[str, list]:
+    log.info("get_all_models()")
+
+    if not app.state.config.ENABLE_OPENAI_API:
+        return {"data": []}
+
+    responses = await get_all_models_responses()
+
+    def extract_data(response):
+        if response and "data" in response:
+            return response["data"]
+        if isinstance(response, list):
+            return response
+        return None
+
+    models = {"data": merge_models_lists(map(extract_data, responses))}
+    log.debug(f"models: {models}")
+
+    return models
+
+
+@app.get("/models")
+@app.get("/models/{url_idx}")
+async def get_models(url_idx: Optional[int] = None, user=Depends(get_verified_user)):
+    models = {
+        "data": [],
+    }
+
+    if url_idx is None:
+        models = await get_all_models()
+    else:
+        url = app.state.config.OPENAI_API_BASE_URLS[url_idx]
+        key = app.state.config.OPENAI_API_KEYS[url_idx]
+
+        headers = {}
+        headers["Authorization"] = f"Bearer {key}"
+        headers["Content-Type"] = "application/json"
+
+        if ENABLE_FORWARD_USER_INFO_HEADERS:
+            headers["X-OpenWebUI-User-Name"] = user.name
+            headers["X-OpenWebUI-User-Id"] = user.id
+            headers["X-OpenWebUI-User-Email"] = user.email
+            headers["X-OpenWebUI-User-Role"] = user.role
+
+        r = None
+
+        timeout = aiohttp.ClientTimeout(total=AIOHTTP_CLIENT_TIMEOUT_OPENAI_MODEL_LIST)
+        async with aiohttp.ClientSession(timeout=timeout) as session:
+            try:
+                async with session.get(f"{url}/models", headers=headers) as r:
+                    if r.status != 200:
+                        # Extract response error details if available
+                        error_detail = f"HTTP Error: {r.status}"
+                        res = await r.json()
+                        if "error" in res:
+                            error_detail = f"External Error: {res['error']}"
+                        raise Exception(error_detail)
+
+                    response_data = await r.json()
+
+                    # Check if we're calling OpenAI API based on the URL
+                    if "api.openai.com" in url:
+                        # Filter models according to the specified conditions
+                        response_data["data"] = [
+                            model
+                            for model in response_data.get("data", [])
+                            if not any(
+                                name in model["id"]
+                                for name in [
+                                    "babbage",
+                                    "dall-e",
+                                    "davinci",
+                                    "embedding",
+                                    "tts",
+                                    "whisper",
+                                ]
+                            )
+                        ]
+
+                    models = response_data
+            except aiohttp.ClientError as e:
+                # ClientError covers all aiohttp requests issues
+                log.exception(f"Client error: {str(e)}")
+                # Handle aiohttp-specific connection issues, timeout etc.
+                raise HTTPException(
+                    status_code=500, detail="Open WebUI: Server Connection Error"
+                )
+            except Exception as e:
+                log.exception(f"Unexpected error: {e}")
+                # Generic error handler in case parsing JSON or other steps fail
+                error_detail = f"Unexpected error: {str(e)}"
+                raise HTTPException(status_code=500, detail=error_detail)
+
+    if user.role == "user":
+        # Filter models based on user access control
+        filtered_models = []
+        for model in models.get("data", []):
+            model_info = Models.get_model_by_id(model["id"])
+            if model_info:
+                if user.id == model_info.user_id or has_access(
+                    user.id, type="read", access_control=model_info.access_control
+                ):
+                    filtered_models.append(model)
+        models["data"] = filtered_models
+
+    return models
+
+
+class ConnectionVerificationForm(BaseModel):
+    url: str
+    key: str
+
+
+@app.post("/verify")
+async def verify_connection(
+    form_data: ConnectionVerificationForm, user=Depends(get_admin_user)
+):
+    url = form_data.url
+    key = form_data.key
+
+    headers = {}
+    headers["Authorization"] = f"Bearer {key}"
+    headers["Content-Type"] = "application/json"
+
+    timeout = aiohttp.ClientTimeout(total=AIOHTTP_CLIENT_TIMEOUT_OPENAI_MODEL_LIST)
+    async with aiohttp.ClientSession(timeout=timeout) as session:
+        try:
+            async with session.get(f"{url}/models", headers=headers) as r:
+                if r.status != 200:
+                    # Extract response error details if available
+                    error_detail = f"HTTP Error: {r.status}"
+                    res = await r.json()
+                    if "error" in res:
+                        error_detail = f"External Error: {res['error']}"
+                    raise Exception(error_detail)
+
+                response_data = await r.json()
+                return response_data
+
+        except aiohttp.ClientError as e:
+            # ClientError covers all aiohttp requests issues
+            log.exception(f"Client error: {str(e)}")
+            # Handle aiohttp-specific connection issues, timeout etc.
+            raise HTTPException(
+                status_code=500, detail="Open WebUI: Server Connection Error"
+            )
+        except Exception as e:
+            log.exception(f"Unexpected error: {e}")
+            # Generic error handler in case parsing JSON or other steps fail
+            error_detail = f"Unexpected error: {str(e)}"
+            raise HTTPException(status_code=500, detail=error_detail)
+
+
+@app.post("/chat/completions")
+async def generate_chat_completion(
+    form_data: dict,
+    user=Depends(get_verified_user),
+    bypass_filter: Optional[bool] = False,
+):
+    idx = 0
+    payload = {**form_data}
+
+    if "metadata" in payload:
+        del payload["metadata"]
+
+    model_id = form_data.get("model")
+    model_info = Models.get_model_by_id(model_id)
+
+    # Check model info and override the payload
+    if model_info:
+        if model_info.base_model_id:
+            payload["model"] = model_info.base_model_id
+
+        params = model_info.params.model_dump()
+        payload = apply_model_params_to_body_openai(params, payload)
+        payload = apply_model_system_prompt_to_body(params, payload, user)
+
+        # Check if user has access to the model
+        if not bypass_filter and user.role == "user":
+            if not (
+                user.id == model_info.user_id
+                or has_access(
+                    user.id, type="read", access_control=model_info.access_control
+                )
+            ):
+                raise HTTPException(
+                    status_code=403,
+                    detail="Model not found",
+                )
+    elif not bypass_filter:
+        if user.role != "admin":
+            raise HTTPException(
+                status_code=403,
+                detail="Model not found",
+            )
+
+    # Attemp to get urlIdx from the model
+    models = await get_all_models()
+
+    # Find the model from the list
+    model = next(
+        (model for model in models["data"] if model["id"] == payload.get("model")),
+        None,
+    )
+
+    if model:
+        idx = model["urlIdx"]
+    else:
+        raise HTTPException(
+            status_code=404,
+            detail="Model not found",
+        )
+
+    # Get the API config for the model
+    api_config = app.state.config.OPENAI_API_CONFIGS.get(
+        app.state.config.OPENAI_API_BASE_URLS[idx], {}
+    )
+    prefix_id = api_config.get("prefix_id", None)
+
+    if prefix_id:
+        payload["model"] = payload["model"].replace(f"{prefix_id}.", "")
+
+    # Add user info to the payload if the model is a pipeline
+    if "pipeline" in model and model.get("pipeline"):
+        payload["user"] = {
+            "name": user.name,
+            "id": user.id,
+            "email": user.email,
+            "role": user.role,
+        }
+
+    url = app.state.config.OPENAI_API_BASE_URLS[idx]
+    key = app.state.config.OPENAI_API_KEYS[idx]
+
+    # Fix: O1 does not support the "max_tokens" parameter, Modify "max_tokens" to "max_completion_tokens"
+    is_o1 = payload["model"].lower().startswith("o1-")
+    # Change max_completion_tokens to max_tokens (Backward compatible)
+    if "api.openai.com" not in url and not is_o1:
+        if "max_completion_tokens" in payload:
+            # Remove "max_completion_tokens" from the payload
+            payload["max_tokens"] = payload["max_completion_tokens"]
+            del payload["max_completion_tokens"]
+    else:
+        if is_o1 and "max_tokens" in payload:
+            payload["max_completion_tokens"] = payload["max_tokens"]
+            del payload["max_tokens"]
+        if "max_tokens" in payload and "max_completion_tokens" in payload:
+            del payload["max_tokens"]
+
+    # Fix: O1 does not support the "system" parameter, Modify "system" to "user"
+    if is_o1 and payload["messages"][0]["role"] == "system":
+        payload["messages"][0]["role"] = "user"
+
+    # Convert the modified body back to JSON
+    payload = json.dumps(payload)
+
+    headers = {}
+    headers["Authorization"] = f"Bearer {key}"
+    headers["Content-Type"] = "application/json"
+    if "openrouter.ai" in app.state.config.OPENAI_API_BASE_URLS[idx]:
+        headers["HTTP-Referer"] = "https://openwebui.com/"
+        headers["X-Title"] = "Open WebUI"
+    if ENABLE_FORWARD_USER_INFO_HEADERS:
+        headers["X-OpenWebUI-User-Name"] = user.name
+        headers["X-OpenWebUI-User-Id"] = user.id
+        headers["X-OpenWebUI-User-Email"] = user.email
+        headers["X-OpenWebUI-User-Role"] = user.role
+
+    r = None
+    session = None
+    streaming = False
+    response = None
+
+    try:
+        session = aiohttp.ClientSession(
+            trust_env=True, timeout=aiohttp.ClientTimeout(total=AIOHTTP_CLIENT_TIMEOUT)
+        )
+        r = await session.request(
+            method="POST",
+            url=f"{url}/chat/completions",
+            data=payload,
+            headers=headers,
+        )
+
+        # Check if response is SSE
+        if "text/event-stream" in r.headers.get("Content-Type", ""):
+            streaming = True
+            return StreamingResponse(
+                r.content,
+                status_code=r.status,
+                headers=dict(r.headers),
+                background=BackgroundTask(
+                    cleanup_response, response=r, session=session
+                ),
+            )
+        else:
+            try:
+                response = await r.json()
+            except Exception as e:
+                log.error(e)
+                response = await r.text()
+
+            r.raise_for_status()
+            return response
+    except Exception as e:
+        log.exception(e)
+        error_detail = "Open WebUI: Server Connection Error"
+        if isinstance(response, dict):
+            if "error" in response:
+                error_detail = f"{response['error']['message'] if 'message' in response['error'] else response['error']}"
+        elif isinstance(response, str):
+            error_detail = response
+
+        raise HTTPException(status_code=r.status if r else 500, detail=error_detail)
+    finally:
+        if not streaming and session:
+            if r:
+                r.close()
+            await session.close()
+
+
+@app.api_route("/{path:path}", methods=["GET", "POST", "PUT", "DELETE"])
+async def proxy(path: str, request: Request, user=Depends(get_verified_user)):
+    idx = 0
+
+    body = await request.body()
+
+    url = app.state.config.OPENAI_API_BASE_URLS[idx]
+    key = app.state.config.OPENAI_API_KEYS[idx]
+
+    target_url = f"{url}/{path}"
+
+    headers = {}
+    headers["Authorization"] = f"Bearer {key}"
+    headers["Content-Type"] = "application/json"
+    if ENABLE_FORWARD_USER_INFO_HEADERS:
+        headers["X-OpenWebUI-User-Name"] = user.name
+        headers["X-OpenWebUI-User-Id"] = user.id
+        headers["X-OpenWebUI-User-Email"] = user.email
+        headers["X-OpenWebUI-User-Role"] = user.role
+
+    r = None
+    session = None
+    streaming = False
+
+    try:
+        session = aiohttp.ClientSession(trust_env=True)
+        r = await session.request(
+            method=request.method,
+            url=target_url,
+            data=body,
+            headers=headers,
+        )
+
+        r.raise_for_status()
+
+        # Check if response is SSE
+        if "text/event-stream" in r.headers.get("Content-Type", ""):
+            streaming = True
+            return StreamingResponse(
+                r.content,
+                status_code=r.status,
+                headers=dict(r.headers),
+                background=BackgroundTask(
+                    cleanup_response, response=r, session=session
+                ),
+            )
+        else:
+            response_data = await r.json()
+            return response_data
+    except Exception as e:
+        log.exception(e)
+        error_detail = "Open WebUI: Server Connection Error"
+        if r is not None:
+            try:
+                res = await r.json()
+                print(res)
+                if "error" in res:
+                    error_detail = f"External: {res['error']['message'] if 'message' in res['error'] else res['error']}"
+            except Exception:
+                error_detail = f"External: {e}"
+        raise HTTPException(status_code=r.status if r else 500, detail=error_detail)
+    finally:
+        if not streaming and session:
+            if r:
+                r.close()
+            await session.close()

backend/open_webui/apps/retrieval/loaders/main.py

@@ -0,0 +1,190 @@
+import requests
+import logging
+import ftfy
+
+from langchain_community.document_loaders import (
+    BSHTMLLoader,
+    CSVLoader,
+    Docx2txtLoader,
+    OutlookMessageLoader,
+    PyPDFLoader,
+    TextLoader,
+    UnstructuredEPubLoader,
+    UnstructuredExcelLoader,
+    UnstructuredMarkdownLoader,
+    UnstructuredPowerPointLoader,
+    UnstructuredRSTLoader,
+    UnstructuredXMLLoader,
+    YoutubeLoader,
+)
+from langchain_core.documents import Document
+from open_webui.env import SRC_LOG_LEVELS
+
+log = logging.getLogger(__name__)
+log.setLevel(SRC_LOG_LEVELS["RAG"])
+
+known_source_ext = [
+    "go",
+    "py",
+    "java",
+    "sh",
+    "bat",
+    "ps1",
+    "cmd",
+    "js",
+    "ts",
+    "css",
+    "cpp",
+    "hpp",
+    "h",
+    "c",
+    "cs",
+    "sql",
+    "log",
+    "ini",
+    "pl",
+    "pm",
+    "r",
+    "dart",
+    "dockerfile",
+    "env",
+    "php",
+    "hs",
+    "hsc",
+    "lua",
+    "nginxconf",
+    "conf",
+    "m",
+    "mm",
+    "plsql",
+    "perl",
+    "rb",
+    "rs",
+    "db2",
+    "scala",
+    "bash",
+    "swift",
+    "vue",
+    "svelte",
+    "msg",
+    "ex",
+    "exs",
+    "erl",
+    "tsx",
+    "jsx",
+    "hs",
+    "lhs",
+]
+
+
+class TikaLoader:
+    def __init__(self, url, file_path, mime_type=None):
+        self.url = url
+        self.file_path = file_path
+        self.mime_type = mime_type
+
+    def load(self) -> list[Document]:
+        with open(self.file_path, "rb") as f:
+            data = f.read()
+
+        if self.mime_type is not None:
+            headers = {"Content-Type": self.mime_type}
+        else:
+            headers = {}
+
+        endpoint = self.url
+        if not endpoint.endswith("/"):
+            endpoint += "/"
+        endpoint += "tika/text"
+
+        r = requests.put(endpoint, data=data, headers=headers)
+
+        if r.ok:
+            raw_metadata = r.json()
+            text = raw_metadata.get("X-TIKA:content", "<No text content found>")
+
+            if "Content-Type" in raw_metadata:
+                headers["Content-Type"] = raw_metadata["Content-Type"]
+
+            log.info("Tika extracted text: %s", text)
+
+            return [Document(page_content=text, metadata=headers)]
+        else:
+            raise Exception(f"Error calling Tika: {r.reason}")
+
+
+class Loader:
+    def __init__(self, engine: str = "", **kwargs):
+        self.engine = engine
+        self.kwargs = kwargs
+
+    def load(
+        self, filename: str, file_content_type: str, file_path: str
+    ) -> list[Document]:
+        loader = self._get_loader(filename, file_content_type, file_path)
+        docs = loader.load()
+
+        return [
+            Document(
+                page_content=ftfy.fix_text(doc.page_content), metadata=doc.metadata
+            )
+            for doc in docs
+        ]
+
+    def _get_loader(self, filename: str, file_content_type: str, file_path: str):
+        file_ext = filename.split(".")[-1].lower()
+
+        if self.engine == "tika" and self.kwargs.get("TIKA_SERVER_URL"):
+            if file_ext in known_source_ext or (
+                file_content_type and file_content_type.find("text/") >= 0
+            ):
+                loader = TextLoader(file_path, autodetect_encoding=True)
+            else:
+                loader = TikaLoader(
+                    url=self.kwargs.get("TIKA_SERVER_URL"),
+                    file_path=file_path,
+                    mime_type=file_content_type,
+                )
+        else:
+            if file_ext == "pdf":
+                loader = PyPDFLoader(
+                    file_path, extract_images=self.kwargs.get("PDF_EXTRACT_IMAGES")
+                )
+            elif file_ext == "csv":
+                loader = CSVLoader(file_path)
+            elif file_ext == "rst":
+                loader = UnstructuredRSTLoader(file_path, mode="elements")
+            elif file_ext == "xml":
+                loader = UnstructuredXMLLoader(file_path)
+            elif file_ext in ["htm", "html"]:
+                loader = BSHTMLLoader(file_path, open_encoding="unicode_escape")
+            elif file_ext == "md":
+                loader = TextLoader(file_path, autodetect_encoding=True)
+            elif file_content_type == "application/epub+zip":
+                loader = UnstructuredEPubLoader(file_path)
+            elif (
+                file_content_type
+                == "application/vnd.openxmlformats-officedocument.wordprocessingml.document"
+                or file_ext == "docx"
+            ):
+                loader = Docx2txtLoader(file_path)
+            elif file_content_type in [
+                "application/vnd.ms-excel",
+                "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet",
+            ] or file_ext in ["xls", "xlsx"]:
+                loader = UnstructuredExcelLoader(file_path)
+            elif file_content_type in [
+                "application/vnd.ms-powerpoint",
+                "application/vnd.openxmlformats-officedocument.presentationml.presentation",
+            ] or file_ext in ["ppt", "pptx"]:
+                loader = UnstructuredPowerPointLoader(file_path)
+            elif file_ext == "msg":
+                loader = OutlookMessageLoader(file_path)
+            elif file_ext in known_source_ext or (
+                file_content_type and file_content_type.find("text/") >= 0
+            ):
+                loader = TextLoader(file_path, autodetect_encoding=True)
+            else:
+                loader = TextLoader(file_path, autodetect_encoding=True)
+
+        return loader

backend/open_webui/apps/retrieval/loaders/youtube.py

@@ -0,0 +1,117 @@
+import logging
+
+from typing import Any, Dict, Generator, List, Optional, Sequence, Union
+from urllib.parse import parse_qs, urlparse
+from langchain_core.documents import Document
+from open_webui.env import SRC_LOG_LEVELS
+
+log = logging.getLogger(__name__)
+log.setLevel(SRC_LOG_LEVELS["RAG"])
+
+ALLOWED_SCHEMES = {"http", "https"}
+ALLOWED_NETLOCS = {
+    "youtu.be",
+    "m.youtube.com",
+    "youtube.com",
+    "www.youtube.com",
+    "www.youtube-nocookie.com",
+    "vid.plus",
+}
+
+
+def _parse_video_id(url: str) -> Optional[str]:
+    """Parse a YouTube URL and return the video ID if valid, otherwise None."""
+    parsed_url = urlparse(url)
+
+    if parsed_url.scheme not in ALLOWED_SCHEMES:
+        return None
+
+    if parsed_url.netloc not in ALLOWED_NETLOCS:
+        return None
+
+    path = parsed_url.path
+
+    if path.endswith("/watch"):
+        query = parsed_url.query
+        parsed_query = parse_qs(query)
+        if "v" in parsed_query:
+            ids = parsed_query["v"]
+            video_id = ids if isinstance(ids, str) else ids[0]
+        else:
+            return None
+    else:
+        path = parsed_url.path.lstrip("/")
+        video_id = path.split("/")[-1]
+
+    if len(video_id) != 11:  # Video IDs are 11 characters long
+        return None
+
+    return video_id
+
+
+class YoutubeLoader:
+    """Load `YouTube` video transcripts."""
+
+    def __init__(
+        self,
+        video_id: str,
+        language: Union[str, Sequence[str]] = "en",
+        proxy_url: Optional[str] = None,
+    ):
+        """Initialize with YouTube video ID."""
+        _video_id = _parse_video_id(video_id)
+        self.video_id = _video_id if _video_id is not None else video_id
+        self._metadata = {"source": video_id}
+        self.language = language
+        self.proxy_url = proxy_url
+        if isinstance(language, str):
+            self.language = [language]
+        else:
+            self.language = language
+
+    def load(self) -> List[Document]:
+        """Load YouTube transcripts into `Document` objects."""
+        try:
+            from youtube_transcript_api import (
+                NoTranscriptFound,
+                TranscriptsDisabled,
+                YouTubeTranscriptApi,
+            )
+        except ImportError:
+            raise ImportError(
+                'Could not import "youtube_transcript_api" Python package. '
+                "Please install it with `pip install youtube-transcript-api`."
+            )
+
+        if self.proxy_url:
+            youtube_proxies = {
+                "http": self.proxy_url,
+                "https": self.proxy_url,
+            }
+            # Don't log complete URL because it might contain secrets
+            log.debug(f"Using proxy URL: {self.proxy_url[:14]}...")
+        else:
+            youtube_proxies = None
+
+        try:
+            transcript_list = YouTubeTranscriptApi.list_transcripts(
+                self.video_id, proxies=youtube_proxies
+            )
+        except Exception as e:
+            log.exception("Loading YouTube transcript failed")
+            return []
+
+        try:
+            transcript = transcript_list.find_transcript(self.language)
+        except NoTranscriptFound:
+            transcript = transcript_list.find_transcript(["en"])
+
+        transcript_pieces: List[Dict[str, Any]] = transcript.fetch()
+
+        transcript = " ".join(
+            map(
+                lambda transcript_piece: transcript_piece["text"].strip(" "),
+                transcript_pieces,
+            )
+        )
+        return [Document(page_content=transcript, metadata=self._metadata)]

backend/open_webui/apps/retrieval/models/colbert.py

@@ -0,0 +1,81 @@
+import os
+import torch
+import numpy as np
+from colbert.infra import ColBERTConfig
+from colbert.modeling.checkpoint import Checkpoint
+
+
+class ColBERT:
+    def __init__(self, name, **kwargs) -> None:
+        print("ColBERT: Loading model", name)
+        self.device = "cuda" if torch.cuda.is_available() else "cpu"
+
+        DOCKER = kwargs.get("env") == "docker"
+        if DOCKER:
+            # This is a workaround for the issue with the docker container
+            # where the torch extension is not loaded properly
+            # and the following error is thrown:
+            # /root/.cache/torch_extensions/py311_cpu/segmented_maxsim_cpp/segmented_maxsim_cpp.so: cannot open shared object file: No such file or directory
+
+            lock_file = (
+                "/root/.cache/torch_extensions/py311_cpu/segmented_maxsim_cpp/lock"
+            )
+            if os.path.exists(lock_file):
+                os.remove(lock_file)
+
+        self.ckpt = Checkpoint(
+            name,
+            colbert_config=ColBERTConfig(model_name=name),
+        ).to(self.device)
+        pass
+
+    def calculate_similarity_scores(self, query_embeddings, document_embeddings):
+
+        query_embeddings = query_embeddings.to(self.device)
+        document_embeddings = document_embeddings.to(self.device)
+
+        # Validate dimensions to ensure compatibility
+        if query_embeddings.dim() != 3:
+            raise ValueError(
+                f"Expected query embeddings to have 3 dimensions, but got {query_embeddings.dim()}."
+            )
+        if document_embeddings.dim() != 3:
+            raise ValueError(
+                f"Expected document embeddings to have 3 dimensions, but got {document_embeddings.dim()}."
+            )
+        if query_embeddings.size(0) not in [1, document_embeddings.size(0)]:
+            raise ValueError(
+                "There should be either one query or queries equal to the number of documents."
+            )
+
+        # Transpose the query embeddings to align for matrix multiplication
+        transposed_query_embeddings = query_embeddings.permute(0, 2, 1)
+        # Compute similarity scores using batch matrix multiplication
+        computed_scores = torch.matmul(document_embeddings, transposed_query_embeddings)
+        # Apply max pooling to extract the highest semantic similarity across each document's sequence
+        maximum_scores = torch.max(computed_scores, dim=1).values
+
+        # Sum up the maximum scores across features to get the overall document relevance scores
+        final_scores = maximum_scores.sum(dim=1)
+
+        normalized_scores = torch.softmax(final_scores, dim=0)
+
+        return normalized_scores.detach().cpu().numpy().astype(np.float32)
+
+    def predict(self, sentences):
+
+        query = sentences[0][0]
+        docs = [i[1] for i in sentences]
+
+        # Embedding the documents
+        embedded_docs = self.ckpt.docFromText(docs, bsize=32)[0]
+        # Embedding the queries
+        embedded_queries = self.ckpt.queryFromText([query], bsize=32)
+        embedded_query = embedded_queries[0]
+
+        # Calculate retrieval scores for the query against all documents
+        scores = self.calculate_similarity_scores(
+            embedded_query.unsqueeze(0), embedded_docs
+        )
+
+        return scores

backend/open_webui/apps/retrieval/utils.py

@@ -1,50 +1,79 @@
-import os
 import logging
+import os
+import uuid
+from typing import Optional, Union
+
+import asyncio
 import requests
 
-from typing import List, Union
-
-from apps.ollama.main import (
-    generate_ollama_embeddings,
-    GenerateEmbeddingsForm,
-)
-
 from huggingface_hub import snapshot_download
-
-from langchain_core.documents import Document
+from langchain.retrievers import ContextualCompressionRetriever, EnsembleRetriever
 from langchain_community.retrievers import BM25Retriever
-from langchain.retrievers import (
-    ContextualCompressionRetriever,
-    EnsembleRetriever,
-)
+from langchain_core.documents import Document
 
-from typing import Optional
+from open_webui.apps.retrieval.vector.connector import VECTOR_DB_CLIENT
+from open_webui.utils.misc import get_last_user_message
 
-from utils.misc import get_last_user_message, add_or_update_system_message
-from config import SRC_LOG_LEVELS, CHROMA_CLIENT
+from open_webui.env import SRC_LOG_LEVELS
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["RAG"])
 
 
+from typing import Any
+
+from langchain_core.callbacks import CallbackManagerForRetrieverRun
+from langchain_core.retrievers import BaseRetriever
+
+
+class VectorSearchRetriever(BaseRetriever):
+    collection_name: Any
+    embedding_function: Any
+    top_k: int
+
+    def _get_relevant_documents(
+        self,
+        query: str,
+        *,
+        run_manager: CallbackManagerForRetrieverRun,
+    ) -> list[Document]:
+        result = VECTOR_DB_CLIENT.search(
+            collection_name=self.collection_name,
+            vectors=[self.embedding_function(query)],
+            limit=self.top_k,
+        )
+
+        ids = result.ids[0]
+        metadatas = result.metadatas[0]
+        documents = result.documents[0]
+
+        results = []
+        for idx in range(len(ids)):
+            results.append(
+                Document(
+                    metadata=metadatas[idx],
+                    page_content=documents[idx],
+                )
+            )
+        return results
+
+
 def query_doc(
     collection_name: str,
-    query: str,
-    embedding_function,
+    query_embedding: list[float],
     k: int,
 ):
     try:
-        collection = CHROMA_CLIENT.get_collection(name=collection_name)
-        query_embeddings = embedding_function(query)
-
-        result = collection.query(
-            query_embeddings=[query_embeddings],
-            n_results=k,
+        result = VECTOR_DB_CLIENT.search(
+            collection_name=collection_name,
+            vectors=[query_embedding],
+            limit=k,
         )
 
-        log.info(f"query_doc:result {result}")
+        log.info(f"query_doc:result {result.ids} {result.metadatas}")
         return result
     except Exception as e:
+        print(e)
         raise e
 
 
@@ -55,27 +84,25 @@ def query_doc_with_hybrid_search(
     k: int,
     reranking_function,
     r: float,
-):
+) -> dict:
     try:
-        collection = CHROMA_CLIENT.get_collection(name=collection_name)
-        documents = collection.get()  # get all documents
+        result = VECTOR_DB_CLIENT.get(collection_name=collection_name)
 
         bm25_retriever = BM25Retriever.from_texts(
-            texts=documents.get("documents"),
-            metadatas=documents.get("metadatas"),
+            texts=result.documents[0],
+            metadatas=result.metadatas[0],
         )
         bm25_retriever.k = k
 
-        chroma_retriever = ChromaRetriever(
-            collection=collection,
+        vector_search_retriever = VectorSearchRetriever(
+            collection_name=collection_name,
             embedding_function=embedding_function,
-            top_n=k,
+            top_k=k,
         )
 
         ensemble_retriever = EnsembleRetriever(
-            retrievers=[bm25_retriever, chroma_retriever], weights=[0.5, 0.5]
+            retrievers=[bm25_retriever, vector_search_retriever], weights=[0.5, 0.5]
         )
-
         compressor = RerankCompressor(
             embedding_function=embedding_function,
             top_n=k,
@@ -94,13 +121,18 @@ def query_doc_with_hybrid_search(
             "metadatas": [[d.metadata for d in result]],
         }
 
-        log.info(f"query_doc_with_hybrid_search:result {result}")
+        log.info(
+            "query_doc_with_hybrid_search:result "
+            + f'{result["metadatas"]} {result["distances"]}'
+        )
         return result
     except Exception as e:
         raise e
 
 
-def merge_and_sort_query_results(query_results, k, reverse=False):
+def merge_and_sort_query_results(
+    query_results: list[dict], k: int, reverse: bool = False
+) -> list[dict]:
     # Initialize lists to store combined data
     combined_distances = []
     combined_documents = []
@@ -142,182 +174,197 @@ def merge_and_sort_query_results(query_results, k, reverse=False):
 
 
 def query_collection(
-    collection_names: List[str],
-    query: str,
+    collection_names: list[str],
+    queries: list[str],
     embedding_function,
     k: int,
-):
+) -> dict:
     results = []
-    for collection_name in collection_names:
-        try:
-            result = query_doc(
-                collection_name=collection_name,
-                query=query,
-                k=k,
-                embedding_function=embedding_function,
-            )
-            results.append(result)
-        except:
-            pass
+    for query in queries:
+        query_embedding = embedding_function(query)
+        for collection_name in collection_names:
+            if collection_name:
+                try:
+                    result = query_doc(
+                        collection_name=collection_name,
+                        k=k,
+                        query_embedding=query_embedding,
+                    )
+                    if result is not None:
+                        results.append(result.model_dump())
+                except Exception as e:
+                    log.exception(f"Error when querying the collection: {e}")
+            else:
+                pass
+
     return merge_and_sort_query_results(results, k=k)
 
 
 def query_collection_with_hybrid_search(
-    collection_names: List[str],
-    query: str,
+    collection_names: list[str],
+    queries: list[str],
     embedding_function,
     k: int,
     reranking_function,
     r: float,
-):
+) -> dict:
     results = []
+    error = False
     for collection_name in collection_names:
         try:
-            result = query_doc_with_hybrid_search(
-                collection_name=collection_name,
-                query=query,
-                embedding_function=embedding_function,
-                k=k,
-                reranking_function=reranking_function,
-                r=r,
+            for query in queries:
+                result = query_doc_with_hybrid_search(
+                    collection_name=collection_name,
+                    query=query,
+                    embedding_function=embedding_function,
+                    k=k,
+                    reranking_function=reranking_function,
+                    r=r,
+                )
+                results.append(result)
+        except Exception as e:
+            log.exception(
+                "Error when querying the collection with " f"hybrid_search: {e}"
             )
-            results.append(result)
-        except:
-            pass
+            error = True
+
+    if error:
+        raise Exception(
+            "Hybrid search failed for all collections. Using Non hybrid search as fallback."
+        )
+
     return merge_and_sort_query_results(results, k=k, reverse=True)
 
 
-def rag_template(template: str, context: str, query: str):
-    template = template.replace("[context]", context)
-    template = template.replace("[query]", query)
-    return template
-
-
 def get_embedding_function(
     embedding_engine,
     embedding_model,
     embedding_function,
-    openai_key,
-    openai_url,
-    batch_size,
+    url,
+    key,
+    embedding_batch_size,
 ):
     if embedding_engine == "":
         return lambda query: embedding_function.encode(query).tolist()
     elif embedding_engine in ["ollama", "openai"]:
-        if embedding_engine == "ollama":
-            func = lambda query: generate_ollama_embeddings(
-                GenerateEmbeddingsForm(
-                    **{
-                        "model": embedding_model,
-                        "prompt": query,
-                    }
-                )
-            )
-        elif embedding_engine == "openai":
-            func = lambda query: generate_openai_embeddings(
-                model=embedding_model,
-                text=query,
-                key=openai_key,
-                url=openai_url,
-            )
+        func = lambda query: generate_embeddings(
+            engine=embedding_engine,
+            model=embedding_model,
+            text=query,
+            url=url,
+            key=key,
+        )
 
-        def generate_multiple(query, f):
+        def generate_multiple(query, func):
             if isinstance(query, list):
-                if embedding_engine == "openai":
-                    embeddings = []
-                    for i in range(0, len(query), batch_size):
-                        embeddings.extend(f(query[i : i + batch_size]))
-                    return embeddings
-                else:
-                    return [f(q) for q in query]
+                embeddings = []
+                for i in range(0, len(query), embedding_batch_size):
+                    embeddings.extend(func(query[i : i + embedding_batch_size]))
+                return embeddings
             else:
-                return f(query)
+                return func(query)
 
         return lambda query: generate_multiple(query, func)
 
 
-def get_rag_context(
+def get_sources_from_files(
     files,
-    messages,
+    queries,
     embedding_function,
     k,
     reranking_function,
     r,
     hybrid_search,
 ):
-    log.debug(f"files: {files} {messages} {embedding_function} {reranking_function}")
-    query = get_last_user_message(messages)
+    log.debug(f"files: {files} {queries} {embedding_function} {reranking_function}")
 
     extracted_collections = []
     relevant_contexts = []
 
     for file in files:
-        context = None
-
-        collection_names = (
-            file["collection_names"]
-            if file["type"] == "collection"
-            else [file["collection_name"]]
-        )
-
-        collection_names = set(collection_names).difference(extracted_collections)
-        if not collection_names:
-            log.debug(f"skipping {file} as it has already been extracted")
-            continue
-
-        try:
-            if file["type"] == "text":
-                context = file["content"]
-            else:
-                if hybrid_search:
-                    context = query_collection_with_hybrid_search(
-                        collection_names=collection_names,
-                        query=query,
-                        embedding_function=embedding_function,
-                        k=k,
-                        reranking_function=reranking_function,
-                        r=r,
-                    )
-                else:
-                    context = query_collection(
-                        collection_names=collection_names,
-                        query=query,
-                        embedding_function=embedding_function,
-                        k=k,
-                    )
-        except Exception as e:
-            log.exception(e)
+        if file.get("context") == "full":
+            context = {
+                "documents": [[file.get("file").get("data", {}).get("content")]],
+                "metadatas": [[{"file_id": file.get("id"), "name": file.get("name")}]],
+            }
+        else:
             context = None
 
+            collection_names = []
+            if file.get("type") == "collection":
+                if file.get("legacy"):
+                    collection_names = file.get("collection_names", [])
+                else:
+                    collection_names.append(file["id"])
+            elif file.get("collection_name"):
+                collection_names.append(file["collection_name"])
+            elif file.get("id"):
+                if file.get("legacy"):
+                    collection_names.append(f"{file['id']}")
+                else:
+                    collection_names.append(f"file-{file['id']}")
+
+            collection_names = set(collection_names).difference(extracted_collections)
+            if not collection_names:
+                log.debug(f"skipping {file} as it has already been extracted")
+                continue
+
+            try:
+                context = None
+                if file.get("type") == "text":
+                    context = file["content"]
+                else:
+                    if hybrid_search:
+                        try:
+                            context = query_collection_with_hybrid_search(
+                                collection_names=collection_names,
+                                queries=queries,
+                                embedding_function=embedding_function,
+                                k=k,
+                                reranking_function=reranking_function,
+                                r=r,
+                            )
+                        except Exception as e:
+                            log.debug(
+                                "Error when using hybrid search, using"
+                                " non hybrid search as fallback."
+                            )
+
+                    if (not hybrid_search) or (context is None):
+                        context = query_collection(
+                            collection_names=collection_names,
+                            queries=queries,
+                            embedding_function=embedding_function,
+                            k=k,
+                        )
+            except Exception as e:
+                log.exception(e)
+
+            extracted_collections.extend(collection_names)
+
         if context:
-            relevant_contexts.append({**context, "source": file})
-
-        extracted_collections.extend(collection_names)
-
-    contexts = []
-    citations = []
+            if "data" in file:
+                del file["data"]
+            relevant_contexts.append({**context, "file": file})
 
+    sources = []
     for context in relevant_contexts:
         try:
             if "documents" in context:
-                contexts.append(
-                    "\n\n".join(
-                        [text for text in context["documents"][0] if text is not None]
-                    )
-                )
-
                 if "metadatas" in context:
-                    citations.append(
-                        {
-                            "source": context["source"],
-                            "document": context["documents"][0],
-                            "metadata": context["metadatas"][0],
-                        }
-                    )
+                    source = {
+                        "source": context["file"],
+                        "document": context["documents"][0],
+                        "metadata": context["metadatas"][0],
+                    }
+                    if "distances" in context and context["distances"]:
+                        source["distances"] = context["distances"][0]
+
+                    sources.append(source)
         except Exception as e:
             log.exception(e)
 
-    return contexts, citations
+    return sources
 
 
 def get_model_path(model: str, update_model: bool = False):
@@ -358,22 +405,8 @@ def get_model_path(model: str, update_model: bool = False):
         return model
 
 
-def generate_openai_embeddings(
-    model: str,
-    text: Union[str, list[str]],
-    key: str,
-    url: str = "https://api.openai.com/v1",
-):
-    if isinstance(text, list):
-        embeddings = generate_openai_batch_embeddings(model, text, key, url)
-    else:
-        embeddings = generate_openai_batch_embeddings(model, [text], key, url)
-
-    return embeddings[0] if isinstance(text, str) else embeddings
-
-
 def generate_openai_batch_embeddings(
-    model: str, texts: list[str], key: str, url: str = "https://api.openai.com/v1"
+    model: str, texts: list[str], url: str = "https://api.openai.com/v1", key: str = ""
 ) -> Optional[list[list[float]]]:
     try:
         r = requests.post(
@@ -395,52 +428,58 @@ def generate_openai_batch_embeddings(
         return None
 
 
-from typing import Any
-
-from langchain_core.retrievers import BaseRetriever
-from langchain_core.callbacks import CallbackManagerForRetrieverRun
-
-
-class ChromaRetriever(BaseRetriever):
-    collection: Any
-    embedding_function: Any
-    top_n: int
-
-    def _get_relevant_documents(
-        self,
-        query: str,
-        *,
-        run_manager: CallbackManagerForRetrieverRun,
-    ) -> List[Document]:
-        query_embeddings = self.embedding_function(query)
-
-        results = self.collection.query(
-            query_embeddings=[query_embeddings],
-            n_results=self.top_n,
+def generate_ollama_batch_embeddings(
+    model: str, texts: list[str], url: str, key: str = ""
+) -> Optional[list[list[float]]]:
+    try:
+        r = requests.post(
+            f"{url}/api/embed",
+            headers={
+                "Content-Type": "application/json",
+                "Authorization": f"Bearer {key}",
+            },
+            json={"input": texts, "model": model},
         )
+        r.raise_for_status()
+        data = r.json()
 
-        ids = results["ids"][0]
-        metadatas = results["metadatas"][0]
-        documents = results["documents"][0]
+        if "embeddings" in data:
+            return data["embeddings"]
+        else:
+            raise "Something went wrong :/"
+    except Exception as e:
+        print(e)
+        return None
 
-        results = []
-        for idx in range(len(ids)):
-            results.append(
-                Document(
-                    metadata=metadatas[idx],
-                    page_content=documents[idx],
-                )
+
+def generate_embeddings(engine: str, model: str, text: Union[str, list[str]], **kwargs):
+    url = kwargs.get("url", "")
+    key = kwargs.get("key", "")
+
+    if engine == "ollama":
+        if isinstance(text, list):
+            embeddings = generate_ollama_batch_embeddings(
+                **{"model": model, "texts": text, "url": url, "key": key}
             )
-        return results
+        else:
+            embeddings = generate_ollama_batch_embeddings(
+                **{"model": model, "texts": [text], "url": url, "key": key}
+            )
+        return embeddings[0] if isinstance(text, str) else embeddings
+    elif engine == "openai":
+        if isinstance(text, list):
+            embeddings = generate_openai_batch_embeddings(model, text, url, key)
+        else:
+            embeddings = generate_openai_batch_embeddings(model, [text], url, key)
+
+        return embeddings[0] if isinstance(text, str) else embeddings
 
 
 import operator
-
 from typing import Optional, Sequence
 
-from langchain_core.documents import BaseDocumentCompressor, Document
 from langchain_core.callbacks import Callbacks
-from langchain_core.pydantic_v1 import Extra
+from langchain_core.documents import BaseDocumentCompressor, Document
 
 
 class RerankCompressor(BaseDocumentCompressor):
@@ -450,7 +489,7 @@ class RerankCompressor(BaseDocumentCompressor):
     r_score: float
 
     class Config:
-        extra = Extra.forbid
+        extra = "forbid"
         arbitrary_types_allowed = True
 
     def compress_documents(

backend/open_webui/apps/retrieval/vector/connector.py

@@ -0,0 +1,22 @@
+from open_webui.config import VECTOR_DB
+
+if VECTOR_DB == "milvus":
+    from open_webui.apps.retrieval.vector.dbs.milvus import MilvusClient
+
+    VECTOR_DB_CLIENT = MilvusClient()
+elif VECTOR_DB == "qdrant":
+    from open_webui.apps.retrieval.vector.dbs.qdrant import QdrantClient
+
+    VECTOR_DB_CLIENT = QdrantClient()
+elif VECTOR_DB == "opensearch":
+    from open_webui.apps.retrieval.vector.dbs.opensearch import OpenSearchClient
+
+    VECTOR_DB_CLIENT = OpenSearchClient()
+elif VECTOR_DB == "pgvector":
+    from open_webui.apps.retrieval.vector.dbs.pgvector import PgvectorClient
+
+    VECTOR_DB_CLIENT = PgvectorClient()
+else:
+    from open_webui.apps.retrieval.vector.dbs.chroma import ChromaClient
+
+    VECTOR_DB_CLIENT = ChromaClient()

backend/open_webui/apps/retrieval/vector/dbs/chroma.py

@@ -0,0 +1,174 @@
+import chromadb
+from chromadb import Settings
+from chromadb.utils.batch_utils import create_batches
+
+from typing import Optional
+
+from open_webui.apps.retrieval.vector.main import VectorItem, SearchResult, GetResult
+from open_webui.config import (
+    CHROMA_DATA_PATH,
+    CHROMA_HTTP_HOST,
+    CHROMA_HTTP_PORT,
+    CHROMA_HTTP_HEADERS,
+    CHROMA_HTTP_SSL,
+    CHROMA_TENANT,
+    CHROMA_DATABASE,
+    CHROMA_CLIENT_AUTH_PROVIDER,
+    CHROMA_CLIENT_AUTH_CREDENTIALS,
+)
+
+
+class ChromaClient:
+    def __init__(self):
+        settings_dict = {
+            "allow_reset": True,
+            "anonymized_telemetry": False,
+        }
+        if CHROMA_CLIENT_AUTH_PROVIDER is not None:
+            settings_dict["chroma_client_auth_provider"] = CHROMA_CLIENT_AUTH_PROVIDER
+        if CHROMA_CLIENT_AUTH_CREDENTIALS is not None:
+            settings_dict["chroma_client_auth_credentials"] = (
+                CHROMA_CLIENT_AUTH_CREDENTIALS
+            )
+
+        if CHROMA_HTTP_HOST != "":
+            self.client = chromadb.HttpClient(
+                host=CHROMA_HTTP_HOST,
+                port=CHROMA_HTTP_PORT,
+                headers=CHROMA_HTTP_HEADERS,
+                ssl=CHROMA_HTTP_SSL,
+                tenant=CHROMA_TENANT,
+                database=CHROMA_DATABASE,
+                settings=Settings(**settings_dict),
+            )
+        else:
+            self.client = chromadb.PersistentClient(
+                path=CHROMA_DATA_PATH,
+                settings=Settings(**settings_dict),
+                tenant=CHROMA_TENANT,
+                database=CHROMA_DATABASE,
+            )
+
+    def has_collection(self, collection_name: str) -> bool:
+        # Check if the collection exists based on the collection name.
+        collections = self.client.list_collections()
+        return collection_name in [collection.name for collection in collections]
+
+    def delete_collection(self, collection_name: str):
+        # Delete the collection based on the collection name.
+        return self.client.delete_collection(name=collection_name)
+
+    def search(
+        self, collection_name: str, vectors: list[list[float | int]], limit: int
+    ) -> Optional[SearchResult]:
+        # Search for the nearest neighbor items based on the vectors and return 'limit' number of results.
+        try:
+            collection = self.client.get_collection(name=collection_name)
+            if collection:
+                result = collection.query(
+                    query_embeddings=vectors,
+                    n_results=limit,
+                )
+
+                return SearchResult(
+                    **{
+                        "ids": result["ids"],
+                        "distances": result["distances"],
+                        "documents": result["documents"],
+                        "metadatas": result["metadatas"],
+                    }
+                )
+            return None
+        except Exception as e:
+            return None
+
+    def query(
+        self, collection_name: str, filter: dict, limit: Optional[int] = None
+    ) -> Optional[GetResult]:
+        # Query the items from the collection based on the filter.
+        try:
+            collection = self.client.get_collection(name=collection_name)
+            if collection:
+                result = collection.get(
+                    where=filter,
+                    limit=limit,
+                )
+
+                return GetResult(
+                    **{
+                        "ids": [result["ids"]],
+                        "documents": [result["documents"]],
+                        "metadatas": [result["metadatas"]],
+                    }
+                )
+            return None
+        except Exception as e:
+            print(e)
+            return None
+
+    def get(self, collection_name: str) -> Optional[GetResult]:
+        # Get all the items in the collection.
+        collection = self.client.get_collection(name=collection_name)
+        if collection:
+            result = collection.get()
+            return GetResult(
+                **{
+                    "ids": [result["ids"]],
+                    "documents": [result["documents"]],
+                    "metadatas": [result["metadatas"]],
+                }
+            )
+        return None
+
+    def insert(self, collection_name: str, items: list[VectorItem]):
+        # Insert the items into the collection, if the collection does not exist, it will be created.
+        collection = self.client.get_or_create_collection(
+            name=collection_name, metadata={"hnsw:space": "cosine"}
+        )
+
+        ids = [item["id"] for item in items]
+        documents = [item["text"] for item in items]
+        embeddings = [item["vector"] for item in items]
+        metadatas = [item["metadata"] for item in items]
+
+        for batch in create_batches(
+            api=self.client,
+            documents=documents,
+            embeddings=embeddings,
+            ids=ids,
+            metadatas=metadatas,
+        ):
+            collection.add(*batch)
+
+    def upsert(self, collection_name: str, items: list[VectorItem]):
+        # Update the items in the collection, if the items are not present, insert them. If the collection does not exist, it will be created.
+        collection = self.client.get_or_create_collection(
+            name=collection_name, metadata={"hnsw:space": "cosine"}
+        )
+
+        ids = [item["id"] for item in items]
+        documents = [item["text"] for item in items]
+        embeddings = [item["vector"] for item in items]
+        metadatas = [item["metadata"] for item in items]
+
+        collection.upsert(
+            ids=ids, documents=documents, embeddings=embeddings, metadatas=metadatas
+        )
+
+    def delete(
+        self,
+        collection_name: str,
+        ids: Optional[list[str]] = None,
+        filter: Optional[dict] = None,
+    ):
+        # Delete the items from the collection based on the ids.
+        collection = self.client.get_collection(name=collection_name)
+        if collection:
+            if ids:
+                collection.delete(ids=ids)
+            elif filter:
+                collection.delete(where=filter)
+
+    def reset(self):
+        # Resets the database. This will delete all collections and item entries.
+        return self.client.reset()

backend/open_webui/apps/retrieval/vector/dbs/milvus.py

@@ -0,0 +1,286 @@
+from pymilvus import MilvusClient as Client
+from pymilvus import FieldSchema, DataType
+import json
+
+from typing import Optional
+
+from open_webui.apps.retrieval.vector.main import VectorItem, SearchResult, GetResult
+from open_webui.config import (
+    MILVUS_URI,
+)
+
+
+class MilvusClient:
+    def __init__(self):
+        self.collection_prefix = "open_webui"
+        self.client = Client(uri=MILVUS_URI)
+
+    def _result_to_get_result(self, result) -> GetResult:
+        ids = []
+        documents = []
+        metadatas = []
+
+        for match in result:
+            _ids = []
+            _documents = []
+            _metadatas = []
+            for item in match:
+                _ids.append(item.get("id"))
+                _documents.append(item.get("data", {}).get("text"))
+                _metadatas.append(item.get("metadata"))
+
+            ids.append(_ids)
+            documents.append(_documents)
+            metadatas.append(_metadatas)
+
+        return GetResult(
+            **{
+                "ids": ids,
+                "documents": documents,
+                "metadatas": metadatas,
+            }
+        )
+
+    def _result_to_search_result(self, result) -> SearchResult:
+        ids = []
+        distances = []
+        documents = []
+        metadatas = []
+
+        for match in result:
+            _ids = []
+            _distances = []
+            _documents = []
+            _metadatas = []
+
+            for item in match:
+                _ids.append(item.get("id"))
+                _distances.append(item.get("distance"))
+                _documents.append(item.get("entity", {}).get("data", {}).get("text"))
+                _metadatas.append(item.get("entity", {}).get("metadata"))
+
+            ids.append(_ids)
+            distances.append(_distances)
+            documents.append(_documents)
+            metadatas.append(_metadatas)
+
+        return SearchResult(
+            **{
+                "ids": ids,
+                "distances": distances,
+                "documents": documents,
+                "metadatas": metadatas,
+            }
+        )
+
+    def _create_collection(self, collection_name: str, dimension: int):
+        schema = self.client.create_schema(
+            auto_id=False,
+            enable_dynamic_field=True,
+        )
+        schema.add_field(
+            field_name="id",
+            datatype=DataType.VARCHAR,
+            is_primary=True,
+            max_length=65535,
+        )
+        schema.add_field(
+            field_name="vector",
+            datatype=DataType.FLOAT_VECTOR,
+            dim=dimension,
+            description="vector",
+        )
+        schema.add_field(field_name="data", datatype=DataType.JSON, description="data")
+        schema.add_field(
+            field_name="metadata", datatype=DataType.JSON, description="metadata"
+        )
+
+        index_params = self.client.prepare_index_params()
+        index_params.add_index(
+            field_name="vector",
+            index_type="HNSW",
+            metric_type="COSINE",
+            params={"M": 16, "efConstruction": 100},
+        )
+
+        self.client.create_collection(
+            collection_name=f"{self.collection_prefix}_{collection_name}",
+            schema=schema,
+            index_params=index_params,
+        )
+
+    def has_collection(self, collection_name: str) -> bool:
+        # Check if the collection exists based on the collection name.
+        collection_name = collection_name.replace("-", "_")
+        return self.client.has_collection(
+            collection_name=f"{self.collection_prefix}_{collection_name}"
+        )
+
+    def delete_collection(self, collection_name: str):
+        # Delete the collection based on the collection name.
+        collection_name = collection_name.replace("-", "_")
+        return self.client.drop_collection(
+            collection_name=f"{self.collection_prefix}_{collection_name}"
+        )
+
+    def search(
+        self, collection_name: str, vectors: list[list[float | int]], limit: int
+    ) -> Optional[SearchResult]:
+        # Search for the nearest neighbor items based on the vectors and return 'limit' number of results.
+        collection_name = collection_name.replace("-", "_")
+        result = self.client.search(
+            collection_name=f"{self.collection_prefix}_{collection_name}",
+            data=vectors,
+            limit=limit,
+            output_fields=["data", "metadata"],
+        )
+
+        return self._result_to_search_result(result)
+
+    def query(self, collection_name: str, filter: dict, limit: Optional[int] = None):
+        # Construct the filter string for querying
+        collection_name = collection_name.replace("-", "_")
+        if not self.has_collection(collection_name):
+            return None
+
+        filter_string = " && ".join(
+            [
+                f'metadata["{key}"] == {json.dumps(value)}'
+                for key, value in filter.items()
+            ]
+        )
+
+        max_limit = 16383  # The maximum number of records per request
+        all_results = []
+
+        if limit is None:
+            limit = float("inf")  # Use infinity as a placeholder for no limit
+
+        # Initialize offset and remaining to handle pagination
+        offset = 0
+        remaining = limit
+
+        try:
+            # Loop until there are no more items to fetch or the desired limit is reached
+            while remaining > 0:
+                print("remaining", remaining)
+                current_fetch = min(
+                    max_limit, remaining
+                )  # Determine how many items to fetch in this iteration
+
+                results = self.client.query(
+                    collection_name=f"{self.collection_prefix}_{collection_name}",
+                    filter=filter_string,
+                    output_fields=["*"],
+                    limit=current_fetch,
+                    offset=offset,
+                )
+
+                if not results:
+                    break
+
+                all_results.extend(results)
+                results_count = len(results)
+                remaining -= (
+                    results_count  # Decrease remaining by the number of items fetched
+                )
+                offset += results_count
+
+                # Break the loop if the results returned are less than the requested fetch count
+                if results_count < current_fetch:
+                    break
+
+            print(all_results)
+            return self._result_to_get_result([all_results])
+        except Exception as e:
+            print(e)
+            return None
+
+    def get(self, collection_name: str) -> Optional[GetResult]:
+        # Get all the items in the collection.
+        collection_name = collection_name.replace("-", "_")
+        result = self.client.query(
+            collection_name=f"{self.collection_prefix}_{collection_name}",
+            filter='id != ""',
+        )
+        return self._result_to_get_result([result])
+
+    def insert(self, collection_name: str, items: list[VectorItem]):
+        # Insert the items into the collection, if the collection does not exist, it will be created.
+        collection_name = collection_name.replace("-", "_")
+        if not self.client.has_collection(
+            collection_name=f"{self.collection_prefix}_{collection_name}"
+        ):
+            self._create_collection(
+                collection_name=collection_name, dimension=len(items[0]["vector"])
+            )
+
+        return self.client.insert(
+            collection_name=f"{self.collection_prefix}_{collection_name}",
+            data=[
+                {
+                    "id": item["id"],
+                    "vector": item["vector"],
+                    "data": {"text": item["text"]},
+                    "metadata": item["metadata"],
+                }
+                for item in items
+            ],
+        )
+
+    def upsert(self, collection_name: str, items: list[VectorItem]):
+        # Update the items in the collection, if the items are not present, insert them. If the collection does not exist, it will be created.
+        collection_name = collection_name.replace("-", "_")
+        if not self.client.has_collection(
+            collection_name=f"{self.collection_prefix}_{collection_name}"
+        ):
+            self._create_collection(
+                collection_name=collection_name, dimension=len(items[0]["vector"])
+            )
+
+        return self.client.upsert(
+            collection_name=f"{self.collection_prefix}_{collection_name}",
+            data=[
+                {
+                    "id": item["id"],
+                    "vector": item["vector"],
+                    "data": {"text": item["text"]},
+                    "metadata": item["metadata"],
+                }
+                for item in items
+            ],
+        )
+
+    def delete(
+        self,
+        collection_name: str,
+        ids: Optional[list[str]] = None,
+        filter: Optional[dict] = None,
+    ):
+        # Delete the items from the collection based on the ids.
+        collection_name = collection_name.replace("-", "_")
+        if ids:
+            return self.client.delete(
+                collection_name=f"{self.collection_prefix}_{collection_name}",
+                ids=ids,
+            )
+        elif filter:
+            # Convert the filter dictionary to a string using JSON_CONTAINS.
+            filter_string = " && ".join(
+                [
+                    f'metadata["{key}"] == {json.dumps(value)}'
+                    for key, value in filter.items()
+                ]
+            )
+
+            return self.client.delete(
+                collection_name=f"{self.collection_prefix}_{collection_name}",
+                filter=filter_string,
+            )
+
+    def reset(self):
+        # Resets the database. This will delete all collections and item entries.
+        collection_names = self.client.list_collections()
+        for collection_name in collection_names:
+            if collection_name.startswith(self.collection_prefix):
+                self.client.drop_collection(collection_name=collection_name)

backend/open_webui/apps/retrieval/vector/dbs/opensearch.py

@@ -0,0 +1,178 @@
+from opensearchpy import OpenSearch
+from typing import Optional
+
+from open_webui.apps.retrieval.vector.main import VectorItem, SearchResult, GetResult
+from open_webui.config import (
+    OPENSEARCH_URI,
+    OPENSEARCH_SSL,
+    OPENSEARCH_CERT_VERIFY,
+    OPENSEARCH_USERNAME,
+    OPENSEARCH_PASSWORD,
+)
+
+
+class OpenSearchClient:
+    def __init__(self):
+        self.index_prefix = "open_webui"
+        self.client = OpenSearch(
+            hosts=[OPENSEARCH_URI],
+            use_ssl=OPENSEARCH_SSL,
+            verify_certs=OPENSEARCH_CERT_VERIFY,
+            http_auth=(OPENSEARCH_USERNAME, OPENSEARCH_PASSWORD),
+        )
+
+    def _result_to_get_result(self, result) -> GetResult:
+        ids = []
+        documents = []
+        metadatas = []
+
+        for hit in result["hits"]["hits"]:
+            ids.append(hit["_id"])
+            documents.append(hit["_source"].get("text"))
+            metadatas.append(hit["_source"].get("metadata"))
+
+        return GetResult(ids=ids, documents=documents, metadatas=metadatas)
+
+    def _result_to_search_result(self, result) -> SearchResult:
+        ids = []
+        distances = []
+        documents = []
+        metadatas = []
+
+        for hit in result["hits"]["hits"]:
+            ids.append(hit["_id"])
+            distances.append(hit["_score"])
+            documents.append(hit["_source"].get("text"))
+            metadatas.append(hit["_source"].get("metadata"))
+
+        return SearchResult(
+            ids=ids, distances=distances, documents=documents, metadatas=metadatas
+        )
+
+    def _create_index(self, index_name: str, dimension: int):
+        body = {
+            "mappings": {
+                "properties": {
+                    "id": {"type": "keyword"},
+                    "vector": {
+                        "type": "dense_vector",
+                        "dims": dimension,  # Adjust based on your vector dimensions
+                        "index": true,
+                        "similarity": "faiss",
+                        "method": {
+                            "name": "hnsw",
+                            "space_type": "ip",  # Use inner product to approximate cosine similarity
+                            "engine": "faiss",
+                            "ef_construction": 128,
+                            "m": 16,
+                        },
+                    },
+                    "text": {"type": "text"},
+                    "metadata": {"type": "object"},
+                }
+            }
+        }
+        self.client.indices.create(index=f"{self.index_prefix}_{index_name}", body=body)
+
+    def _create_batches(self, items: list[VectorItem], batch_size=100):
+        for i in range(0, len(items), batch_size):
+            yield items[i : i + batch_size]
+
+    def has_collection(self, index_name: str) -> bool:
+        # has_collection here means has index.
+        # We are simply adapting to the norms of the other DBs.
+        return self.client.indices.exists(index=f"{self.index_prefix}_{index_name}")
+
+    def delete_colleciton(self, index_name: str):
+        # delete_collection here means delete index.
+        # We are simply adapting to the norms of the other DBs.
+        self.client.indices.delete(index=f"{self.index_prefix}_{index_name}")
+
+    def search(
+        self, index_name: str, vectors: list[list[float]], limit: int
+    ) -> Optional[SearchResult]:
+        query = {
+            "size": limit,
+            "_source": ["text", "metadata"],
+            "query": {
+                "script_score": {
+                    "query": {"match_all": {}},
+                    "script": {
+                        "source": "cosineSimilarity(params.vector, 'vector') + 1.0",
+                        "params": {
+                            "vector": vectors[0]
+                        },  # Assuming single query vector
+                    },
+                }
+            },
+        }
+
+        result = self.client.search(
+            index=f"{self.index_prefix}_{index_name}", body=query
+        )
+
+        return self._result_to_search_result(result)
+
+    def get_or_create_index(self, index_name: str, dimension: int):
+        if not self.has_index(index_name):
+            self._create_index(index_name, dimension)
+
+    def get(self, index_name: str) -> Optional[GetResult]:
+        query = {"query": {"match_all": {}}, "_source": ["text", "metadata"]}
+
+        result = self.client.search(
+            index=f"{self.index_prefix}_{index_name}", body=query
+        )
+        return self._result_to_get_result(result)
+
+    def insert(self, index_name: str, items: list[VectorItem]):
+        if not self.has_index(index_name):
+            self._create_index(index_name, dimension=len(items[0]["vector"]))
+
+        for batch in self._create_batches(items):
+            actions = [
+                {
+                    "index": {
+                        "_id": item["id"],
+                        "_source": {
+                            "vector": item["vector"],
+                            "text": item["text"],
+                            "metadata": item["metadata"],
+                        },
+                    }
+                }
+                for item in batch
+            ]
+            self.client.bulk(actions)
+
+    def upsert(self, index_name: str, items: list[VectorItem]):
+        if not self.has_index(index_name):
+            self._create_index(index_name, dimension=len(items[0]["vector"]))
+
+        for batch in self._create_batches(items):
+            actions = [
+                {
+                    "index": {
+                        "_id": item["id"],
+                        "_source": {
+                            "vector": item["vector"],
+                            "text": item["text"],
+                            "metadata": item["metadata"],
+                        },
+                    }
+                }
+                for item in batch
+            ]
+            self.client.bulk(actions)
+
+    def delete(self, index_name: str, ids: list[str]):
+        actions = [
+            {"delete": {"_index": f"{self.index_prefix}_{index_name}", "_id": id}}
+            for id in ids
+        ]
+        self.client.bulk(body=actions)
+
+    def reset(self):
+        indices = self.client.indices.get(index=f"{self.index_prefix}_*")
+        for index in indices:
+            self.client.indices.delete(index=index)

backend/open_webui/apps/retrieval/vector/dbs/pgvector.py

@@ -0,0 +1,354 @@
+from typing import Optional, List, Dict, Any
+from sqlalchemy import (
+    cast,
+    column,
+    create_engine,
+    Column,
+    Integer,
+    select,
+    text,
+    Text,
+    values,
+)
+from sqlalchemy.sql import true
+from sqlalchemy.pool import NullPool
+
+from sqlalchemy.orm import declarative_base, scoped_session, sessionmaker
+from sqlalchemy.dialects.postgresql import JSONB, array
+from pgvector.sqlalchemy import Vector
+from sqlalchemy.ext.mutable import MutableDict
+
+from open_webui.apps.retrieval.vector.main import VectorItem, SearchResult, GetResult
+from open_webui.config import PGVECTOR_DB_URL
+
+VECTOR_LENGTH = 1536
+Base = declarative_base()
+
+
+class DocumentChunk(Base):
+    __tablename__ = "document_chunk"
+
+    id = Column(Text, primary_key=True)
+    vector = Column(Vector(dim=VECTOR_LENGTH), nullable=True)
+    collection_name = Column(Text, nullable=False)
+    text = Column(Text, nullable=True)
+    vmetadata = Column(MutableDict.as_mutable(JSONB), nullable=True)
+
+
+class PgvectorClient:
+    def __init__(self) -> None:
+
+        # if no pgvector uri, use the existing database connection
+        if not PGVECTOR_DB_URL:
+            from open_webui.apps.webui.internal.db import Session
+
+            self.session = Session
+        else:
+            engine = create_engine(
+                PGVECTOR_DB_URL, pool_pre_ping=True, poolclass=NullPool
+            )
+            SessionLocal = sessionmaker(
+                autocommit=False, autoflush=False, bind=engine, expire_on_commit=False
+            )
+            self.session = scoped_session(SessionLocal)
+
+        try:
+            # Ensure the pgvector extension is available
+            self.session.execute(text("CREATE EXTENSION IF NOT EXISTS vector;"))
+
+            # Create the tables if they do not exist
+            # Base.metadata.create_all requires a bind (engine or connection)
+            # Get the connection from the session
+            connection = self.session.connection()
+            Base.metadata.create_all(bind=connection)
+
+            # Create an index on the vector column if it doesn't exist
+            self.session.execute(
+                text(
+                    "CREATE INDEX IF NOT EXISTS idx_document_chunk_vector "
+                    "ON document_chunk USING ivfflat (vector vector_cosine_ops) WITH (lists = 100);"
+                )
+            )
+            self.session.execute(
+                text(
+                    "CREATE INDEX IF NOT EXISTS idx_document_chunk_collection_name "
+                    "ON document_chunk (collection_name);"
+                )
+            )
+            self.session.commit()
+            print("Initialization complete.")
+        except Exception as e:
+            self.session.rollback()
+            print(f"Error during initialization: {e}")
+            raise
+
+    def adjust_vector_length(self, vector: List[float]) -> List[float]:
+        # Adjust vector to have length VECTOR_LENGTH
+        current_length = len(vector)
+        if current_length < VECTOR_LENGTH:
+            # Pad the vector with zeros
+            vector += [0.0] * (VECTOR_LENGTH - current_length)
+        elif current_length > VECTOR_LENGTH:
+            raise Exception(
+                f"Vector length {current_length} not supported. Max length must be <= {VECTOR_LENGTH}"
+            )
+        return vector
+
+    def insert(self, collection_name: str, items: List[VectorItem]) -> None:
+        try:
+            new_items = []
+            for item in items:
+                vector = self.adjust_vector_length(item["vector"])
+                new_chunk = DocumentChunk(
+                    id=item["id"],
+                    vector=vector,
+                    collection_name=collection_name,
+                    text=item["text"],
+                    vmetadata=item["metadata"],
+                )
+                new_items.append(new_chunk)
+            self.session.bulk_save_objects(new_items)
+            self.session.commit()
+            print(
+                f"Inserted {len(new_items)} items into collection '{collection_name}'."
+            )
+        except Exception as e:
+            self.session.rollback()
+            print(f"Error during insert: {e}")
+            raise
+
+    def upsert(self, collection_name: str, items: List[VectorItem]) -> None:
+        try:
+            for item in items:
+                vector = self.adjust_vector_length(item["vector"])
+                existing = (
+                    self.session.query(DocumentChunk)
+                    .filter(DocumentChunk.id == item["id"])
+                    .first()
+                )
+                if existing:
+                    existing.vector = vector
+                    existing.text = item["text"]
+                    existing.vmetadata = item["metadata"]
+                    existing.collection_name = (
+                        collection_name  # Update collection_name if necessary
+                    )
+                else:
+                    new_chunk = DocumentChunk(
+                        id=item["id"],
+                        vector=vector,
+                        collection_name=collection_name,
+                        text=item["text"],
+                        vmetadata=item["metadata"],
+                    )
+                    self.session.add(new_chunk)
+            self.session.commit()
+            print(f"Upserted {len(items)} items into collection '{collection_name}'.")
+        except Exception as e:
+            self.session.rollback()
+            print(f"Error during upsert: {e}")
+            raise
+
+    def search(
+        self,
+        collection_name: str,
+        vectors: List[List[float]],
+        limit: Optional[int] = None,
+    ) -> Optional[SearchResult]:
+        try:
+            if not vectors:
+                return None
+
+            # Adjust query vectors to VECTOR_LENGTH
+            vectors = [self.adjust_vector_length(vector) for vector in vectors]
+            num_queries = len(vectors)
+
+            def vector_expr(vector):
+                return cast(array(vector), Vector(VECTOR_LENGTH))
+
+            # Create the values for query vectors
+            qid_col = column("qid", Integer)
+            q_vector_col = column("q_vector", Vector(VECTOR_LENGTH))
+            query_vectors = (
+                values(qid_col, q_vector_col)
+                .data(
+                    [(idx, vector_expr(vector)) for idx, vector in enumerate(vectors)]
+                )
+                .alias("query_vectors")
+            )
+
+            # Build the lateral subquery for each query vector
+            subq = (
+                select(
+                    DocumentChunk.id,
+                    DocumentChunk.text,
+                    DocumentChunk.vmetadata,
+                    (
+                        DocumentChunk.vector.cosine_distance(query_vectors.c.q_vector)
+                    ).label("distance"),
+                )
+                .where(DocumentChunk.collection_name == collection_name)
+                .order_by(
+                    (DocumentChunk.vector.cosine_distance(query_vectors.c.q_vector))
+                )
+            )
+            if limit is not None:
+                subq = subq.limit(limit)
+            subq = subq.lateral("result")
+
+            # Build the main query by joining query_vectors and the lateral subquery
+            stmt = (
+                select(
+                    query_vectors.c.qid,
+                    subq.c.id,
+                    subq.c.text,
+                    subq.c.vmetadata,
+                    subq.c.distance,
+                )
+                .select_from(query_vectors)
+                .join(subq, true())
+                .order_by(query_vectors.c.qid, subq.c.distance)
+            )
+
+            result_proxy = self.session.execute(stmt)
+            results = result_proxy.all()
+
+            ids = [[] for _ in range(num_queries)]
+            distances = [[] for _ in range(num_queries)]
+            documents = [[] for _ in range(num_queries)]
+            metadatas = [[] for _ in range(num_queries)]
+
+            if not results:
+                return SearchResult(
+                    ids=ids,
+                    distances=distances,
+                    documents=documents,
+                    metadatas=metadatas,
+                )
+
+            for row in results:
+                qid = int(row.qid)
+                ids[qid].append(row.id)
+                distances[qid].append(row.distance)
+                documents[qid].append(row.text)
+                metadatas[qid].append(row.vmetadata)
+
+            return SearchResult(
+                ids=ids, distances=distances, documents=documents, metadatas=metadatas
+            )
+        except Exception as e:
+            print(f"Error during search: {e}")
+            return None
+
+    def query(
+        self, collection_name: str, filter: Dict[str, Any], limit: Optional[int] = None
+    ) -> Optional[GetResult]:
+        try:
+            query = self.session.query(DocumentChunk).filter(
+                DocumentChunk.collection_name == collection_name
+            )
+
+            for key, value in filter.items():
+                query = query.filter(DocumentChunk.vmetadata[key].astext == str(value))
+
+            if limit is not None:
+                query = query.limit(limit)
+
+            results = query.all()
+
+            if not results:
+                return None
+
+            ids = [[result.id for result in results]]
+            documents = [[result.text for result in results]]
+            metadatas = [[result.vmetadata for result in results]]
+
+            return GetResult(
+                ids=ids,
+                documents=documents,
+                metadatas=metadatas,
+            )
+        except Exception as e:
+            print(f"Error during query: {e}")
+            return None
+
+    def get(
+        self, collection_name: str, limit: Optional[int] = None
+    ) -> Optional[GetResult]:
+        try:
+            query = self.session.query(DocumentChunk).filter(
+                DocumentChunk.collection_name == collection_name
+            )
+            if limit is not None:
+                query = query.limit(limit)
+
+            results = query.all()
+
+            if not results:
+                return None
+
+            ids = [[result.id for result in results]]
+            documents = [[result.text for result in results]]
+            metadatas = [[result.vmetadata for result in results]]
+
+            return GetResult(ids=ids, documents=documents, metadatas=metadatas)
+        except Exception as e:
+            print(f"Error during get: {e}")
+            return None
+
+    def delete(
+        self,
+        collection_name: str,
+        ids: Optional[List[str]] = None,
+        filter: Optional[Dict[str, Any]] = None,
+    ) -> None:
+        try:
+            query = self.session.query(DocumentChunk).filter(
+                DocumentChunk.collection_name == collection_name
+            )
+            if ids:
+                query = query.filter(DocumentChunk.id.in_(ids))
+            if filter:
+                for key, value in filter.items():
+                    query = query.filter(
+                        DocumentChunk.vmetadata[key].astext == str(value)
+                    )
+            deleted = query.delete(synchronize_session=False)
+            self.session.commit()
+            print(f"Deleted {deleted} items from collection '{collection_name}'.")
+        except Exception as e:
+            self.session.rollback()
+            print(f"Error during delete: {e}")
+            raise
+
+    def reset(self) -> None:
+        try:
+            deleted = self.session.query(DocumentChunk).delete()
+            self.session.commit()
+            print(
+                f"Reset complete. Deleted {deleted} items from 'document_chunk' table."
+            )
+        except Exception as e:
+            self.session.rollback()
+            print(f"Error during reset: {e}")
+            raise
+
+    def close(self) -> None:
+        pass
+
+    def has_collection(self, collection_name: str) -> bool:
+        try:
+            exists = (
+                self.session.query(DocumentChunk)
+                .filter(DocumentChunk.collection_name == collection_name)
+                .first()
+                is not None
+            )
+            return exists
+        except Exception as e:
+            print(f"Error checking collection existence: {e}")
+            return False
+
+    def delete_collection(self, collection_name: str) -> None:
+        self.delete(collection_name)
+        print(f"Collection '{collection_name}' deleted.")

backend/open_webui/apps/retrieval/vector/dbs/qdrant.py

@@ -0,0 +1,184 @@
+from typing import Optional
+
+from qdrant_client import QdrantClient as Qclient
+from qdrant_client.http.models import PointStruct
+from qdrant_client.models import models
+
+from open_webui.apps.retrieval.vector.main import VectorItem, SearchResult, GetResult
+from open_webui.config import QDRANT_URI, QDRANT_API_KEY
+
+NO_LIMIT = 999999999
+
+
+class QdrantClient:
+    def __init__(self):
+        self.collection_prefix = "open-webui"
+        self.QDRANT_URI = QDRANT_URI
+        self.QDRANT_API_KEY = QDRANT_API_KEY
+        self.client = (
+            Qclient(url=self.QDRANT_URI, api_key=self.QDRANT_API_KEY)
+            if self.QDRANT_URI
+            else None
+        )
+
+    def _result_to_get_result(self, points) -> GetResult:
+        ids = []
+        documents = []
+        metadatas = []
+
+        for point in points:
+            payload = point.payload
+            ids.append(point.id)
+            documents.append(payload["text"])
+            metadatas.append(payload["metadata"])
+
+        return GetResult(
+            **{
+                "ids": [ids],
+                "documents": [documents],
+                "metadatas": [metadatas],
+            }
+        )
+
+    def _create_collection(self, collection_name: str, dimension: int):
+        collection_name_with_prefix = f"{self.collection_prefix}_{collection_name}"
+        self.client.create_collection(
+            collection_name=collection_name_with_prefix,
+            vectors_config=models.VectorParams(
+                size=dimension, distance=models.Distance.COSINE
+            ),
+        )
+
+        print(f"collection {collection_name_with_prefix} successfully created!")
+
+    def _create_collection_if_not_exists(self, collection_name, dimension):
+        if not self.has_collection(collection_name=collection_name):
+            self._create_collection(
+                collection_name=collection_name, dimension=dimension
+            )
+
+    def _create_points(self, items: list[VectorItem]):
+        return [
+            PointStruct(
+                id=item["id"],
+                vector=item["vector"],
+                payload={"text": item["text"], "metadata": item["metadata"]},
+            )
+            for item in items
+        ]
+
+    def has_collection(self, collection_name: str) -> bool:
+        return self.client.collection_exists(
+            f"{self.collection_prefix}_{collection_name}"
+        )
+
+    def delete_collection(self, collection_name: str):
+        return self.client.delete_collection(
+            collection_name=f"{self.collection_prefix}_{collection_name}"
+        )
+
+    def search(
+        self, collection_name: str, vectors: list[list[float | int]], limit: int
+    ) -> Optional[SearchResult]:
+        # Search for the nearest neighbor items based on the vectors and return 'limit' number of results.
+        if limit is None:
+            limit = NO_LIMIT  # otherwise qdrant would set limit to 10!
+
+        query_response = self.client.query_points(
+            collection_name=f"{self.collection_prefix}_{collection_name}",
+            query=vectors[0],
+            limit=limit,
+        )
+        get_result = self._result_to_get_result(query_response.points)
+        return SearchResult(
+            ids=get_result.ids,
+            documents=get_result.documents,
+            metadatas=get_result.metadatas,
+            distances=[[point.score for point in query_response.points]],
+        )
+
+    def query(self, collection_name: str, filter: dict, limit: Optional[int] = None):
+        # Construct the filter string for querying
+        if not self.has_collection(collection_name):
+            return None
+        try:
+            if limit is None:
+                limit = NO_LIMIT  # otherwise qdrant would set limit to 10!
+
+            field_conditions = []
+            for key, value in filter.items():
+                field_conditions.append(
+                    models.FieldCondition(
+                        key=f"metadata.{key}", match=models.MatchValue(value=value)
+                    )
+                )
+
+            points = self.client.query_points(
+                collection_name=f"{self.collection_prefix}_{collection_name}",
+                query_filter=models.Filter(should=field_conditions),
+                limit=limit,
+            )
+            return self._result_to_get_result(points.points)
+        except Exception as e:
+            print(e)
+            return None
+
+    def get(self, collection_name: str) -> Optional[GetResult]:
+        # Get all the items in the collection.
+        points = self.client.query_points(
+            collection_name=f"{self.collection_prefix}_{collection_name}",
+            limit=NO_LIMIT,  # otherwise qdrant would set limit to 10!
+        )
+        return self._result_to_get_result(points.points)
+
+    def insert(self, collection_name: str, items: list[VectorItem]):
+        # Insert the items into the collection, if the collection does not exist, it will be created.
+        self._create_collection_if_not_exists(collection_name, len(items[0]["vector"]))
+        points = self._create_points(items)
+        self.client.upload_points(f"{self.collection_prefix}_{collection_name}", points)
+
+    def upsert(self, collection_name: str, items: list[VectorItem]):
+        # Update the items in the collection, if the items are not present, insert them. If the collection does not exist, it will be created.
+        self._create_collection_if_not_exists(collection_name, len(items[0]["vector"]))
+        points = self._create_points(items)
+        return self.client.upsert(f"{self.collection_prefix}_{collection_name}", points)
+
+    def delete(
+        self,
+        collection_name: str,
+        ids: Optional[list[str]] = None,
+        filter: Optional[dict] = None,
+    ):
+        # Delete the items from the collection based on the ids.
+        field_conditions = []
+
+        if ids:
+            for id_value in ids:
+                field_conditions.append(
+                    models.FieldCondition(
+                        key="metadata.id",
+                        match=models.MatchValue(value=id_value),
+                    ),
+                ),
+        elif filter:
+            for key, value in filter.items():
+                field_conditions.append(
+                    models.FieldCondition(
+                        key=f"metadata.{key}",
+                        match=models.MatchValue(value=value),
+                    ),
+                ),
+
+        return self.client.delete(
+            collection_name=f"{self.collection_prefix}_{collection_name}",
+            points_selector=models.FilterSelector(
+                filter=models.Filter(must=field_conditions)
+            ),
+        )
+
+    def reset(self):
+        # Resets the database. This will delete all collections and item entries.
+        collection_names = self.client.get_collections().collections
+        for collection_name in collection_names:
+            if collection_name.name.startswith(self.collection_prefix):
+                self.client.delete_collection(collection_name=collection_name.name)

backend/open_webui/apps/retrieval/vector/main.py

@@ -0,0 +1,19 @@
+from pydantic import BaseModel
+from typing import Optional, List, Any
+
+
+class VectorItem(BaseModel):
+    id: str
+    text: str
+    vector: List[float | int]
+    metadata: Any
+
+
+class GetResult(BaseModel):
+    ids: Optional[List[List[str]]]
+    documents: Optional[List[List[str]]]
+    metadatas: Optional[List[List[Any]]]
+
+
+class SearchResult(GetResult):
+    distances: Optional[List[List[float | int]]]

backend/open_webui/apps/retrieval/web/bing.py

@@ -0,0 +1,73 @@
+import logging
+import os
+from pprint import pprint
+from typing import Optional
+import requests
+from open_webui.apps.retrieval.web.main import SearchResult, get_filtered_results
+from open_webui.env import SRC_LOG_LEVELS
+import argparse
+
+log = logging.getLogger(__name__)
+log.setLevel(SRC_LOG_LEVELS["RAG"])
+"""
+Documentation: https://docs.microsoft.com/en-us/bing/search-apis/bing-web-search/overview
+"""
+
+
+def search_bing(
+    subscription_key: str,
+    endpoint: str,
+    locale: str,
+    query: str,
+    count: int,
+    filter_list: Optional[list[str]] = None,
+) -> list[SearchResult]:
+    mkt = locale
+    params = {"q": query, "mkt": mkt, "answerCount": count}
+    headers = {"Ocp-Apim-Subscription-Key": subscription_key}
+
+    try:
+        response = requests.get(endpoint, headers=headers, params=params)
+        response.raise_for_status()
+        json_response = response.json()
+        results = json_response.get("webPages", {}).get("value", [])
+        if filter_list:
+            results = get_filtered_results(results, filter_list)
+        return [
+            SearchResult(
+                link=result["url"],
+                title=result.get("name"),
+                snippet=result.get("snippet"),
+            )
+            for result in results
+        ]
+    except Exception as ex:
+        log.error(f"Error: {ex}")
+        raise ex
+
+
+def main():
+    parser = argparse.ArgumentParser(description="Search Bing from the command line.")
+    parser.add_argument(
+        "query",
+        type=str,
+        default="Top 10 international news today",
+        help="The search query.",
+    )
+    parser.add_argument(
+        "--count", type=int, default=10, help="Number of search results to return."
+    )
+    parser.add_argument(
+        "--filter", nargs="*", help="List of filters to apply to the search results."
+    )
+    parser.add_argument(
+        "--locale",
+        type=str,
+        default="en-US",
+        help="The locale to use for the search, maps to market in api",
+    )
+
+    args = parser.parse_args()
+
+    results = search_bing(args.locale, args.query, args.count, args.filter)
+    pprint(results)

backend/open_webui/apps/retrieval/web/brave.py

@@ -1,16 +1,16 @@
 import logging
-from typing import List, Optional
-import requests
+from typing import Optional
 
-from apps.rag.search.main import SearchResult, get_filtered_results
-from config import SRC_LOG_LEVELS
+import requests
+from open_webui.apps.retrieval.web.main import SearchResult, get_filtered_results
+from open_webui.env import SRC_LOG_LEVELS
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["RAG"])
 
 
 def search_brave(
-    api_key: str, query: str, count: int, filter_list: Optional[List[str]] = None
+    api_key: str, query: str, count: int, filter_list: Optional[list[str]] = None
 ) -> list[SearchResult]:
     """Search using Brave's Search API and return the results as a list of SearchResult objects.
 

backend/open_webui/apps/retrieval/web/duckduckgo.py

@@ -1,15 +1,16 @@
 import logging
-from typing import List, Optional
-from apps.rag.search.main import SearchResult, get_filtered_results
+from typing import Optional
+
+from open_webui.apps.retrieval.web.main import SearchResult, get_filtered_results
 from duckduckgo_search import DDGS
-from config import SRC_LOG_LEVELS
+from open_webui.env import SRC_LOG_LEVELS
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["RAG"])
 
 
 def search_duckduckgo(
-    query: str, count: int, filter_list: Optional[List[str]] = None
+    query: str, count: int, filter_list: Optional[list[str]] = None
 ) -> list[SearchResult]:
     """
     Search using DuckDuckGo's Search API and return the results as a list of SearchResult objects.
@@ -18,7 +19,7 @@ def search_duckduckgo(
         count (int): The number of results to return
 
     Returns:
-        List[SearchResult]: A list of search results
+        list[SearchResult]: A list of search results
     """
     # Use the DDGS context manager to create a DDGS object
     with DDGS() as ddgs:

backend/open_webui/apps/retrieval/web/google_pse.py

@@ -1,10 +1,9 @@
-import json
 import logging
-from typing import List, Optional
-import requests
+from typing import Optional
 
-from apps.rag.search.main import SearchResult, get_filtered_results
-from config import SRC_LOG_LEVELS
+import requests
+from open_webui.apps.retrieval.web.main import SearchResult, get_filtered_results
+from open_webui.env import SRC_LOG_LEVELS
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["RAG"])
@@ -15,7 +14,7 @@ def search_google_pse(
     search_engine_id: str,
     query: str,
     count: int,
-    filter_list: Optional[List[str]] = None,
+    filter_list: Optional[list[str]] = None,
 ) -> list[SearchResult]:
     """Search using Google's Programmable Search Engine API and return the results as a list of SearchResult objects.
 

backend/open_webui/apps/retrieval/web/jina_search.py

@@ -1,15 +1,15 @@
 import logging
-import requests
-from yarl import URL
 
-from apps.rag.search.main import SearchResult
-from config import SRC_LOG_LEVELS
+import requests
+from open_webui.apps.retrieval.web.main import SearchResult
+from open_webui.env import SRC_LOG_LEVELS
+from yarl import URL
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["RAG"])
 
 
-def search_jina(query: str, count: int) -> list[SearchResult]:
+def search_jina(api_key: str, query: str, count: int) -> list[SearchResult]:
     """
     Search using Jina's Search API and return the results as a list of SearchResult objects.
     Args:
@@ -17,12 +17,10 @@ def search_jina(query: str, count: int) -> list[SearchResult]:
         count (int): The number of results to return
 
     Returns:
-        List[SearchResult]: A list of search results
+        list[SearchResult]: A list of search results
     """
     jina_search_endpoint = "https://s.jina.ai/"
-    headers = {
-        "Accept": "application/json",
-    }
+    headers = {"Accept": "application/json", "Authorization": f"Bearer {api_key}"}
     url = str(URL(jina_search_endpoint + query))
     response = requests.get(url, headers=headers)
     response.raise_for_status()

backend/open_webui/apps/retrieval/web/main.py

@@ -1,5 +1,6 @@
 from typing import Optional
 from urllib.parse import urlparse
+
 from pydantic import BaseModel
 
 
@@ -8,7 +9,8 @@ def get_filtered_results(results, filter_list):
         return results
     filtered_results = []
     for result in results:
-        domain = urlparse(result["url"]).netloc
+        url = result.get("url") or result.get("link", "")
+        domain = urlparse(url).netloc
         if any(domain.endswith(filtered_domain) for filtered_domain in filter_list):
             filtered_results.append(result)
     return filtered_results

backend/open_webui/apps/retrieval/web/mojeek.py

@@ -0,0 +1,40 @@
+import logging
+from typing import Optional
+
+import requests
+from open_webui.apps.retrieval.web.main import SearchResult, get_filtered_results
+from open_webui.env import SRC_LOG_LEVELS
+
+log = logging.getLogger(__name__)
+log.setLevel(SRC_LOG_LEVELS["RAG"])
+
+
+def search_mojeek(
+    api_key: str, query: str, count: int, filter_list: Optional[list[str]] = None
+) -> list[SearchResult]:
+    """Search using Mojeek's Search API and return the results as a list of SearchResult objects.
+
+    Args:
+        api_key (str): A Mojeek Search API key
+        query (str): The query to search for
+    """
+    url = "https://api.mojeek.com/search"
+    headers = {
+        "Accept": "application/json",
+    }
+    params = {"q": query, "api_key": api_key, "fmt": "json", "t": count}
+
+    response = requests.get(url, headers=headers, params=params)
+    response.raise_for_status()
+    json_response = response.json()
+    results = json_response.get("response", {}).get("results", [])
+    print(results)
+    if filter_list:
+        results = get_filtered_results(results, filter_list)
+
+    return [
+        SearchResult(
+            link=result["url"], title=result.get("title"), snippet=result.get("desc")
+        )
+        for result in results
+    ]

backend/open_webui/apps/retrieval/web/searchapi.py

@@ -0,0 +1,48 @@
+import logging
+from typing import Optional
+from urllib.parse import urlencode
+
+import requests
+from open_webui.apps.retrieval.web.main import SearchResult, get_filtered_results
+from open_webui.env import SRC_LOG_LEVELS
+
+log = logging.getLogger(__name__)
+log.setLevel(SRC_LOG_LEVELS["RAG"])
+
+
+def search_searchapi(
+    api_key: str,
+    engine: str,
+    query: str,
+    count: int,
+    filter_list: Optional[list[str]] = None,
+) -> list[SearchResult]:
+    """Search using searchapi.io's API and return the results as a list of SearchResult objects.
+
+    Args:
+      api_key (str): A searchapi.io API key
+      query (str): The query to search for
+    """
+    url = "https://www.searchapi.io/api/v1/search"
+
+    engine = engine or "google"
+
+    payload = {"engine": engine, "q": query, "api_key": api_key}
+
+    url = f"{url}?{urlencode(payload)}"
+    response = requests.request("GET", url)
+
+    json_response = response.json()
+    log.info(f"results from searchapi search: {json_response}")
+
+    results = sorted(
+        json_response.get("organic_results", []), key=lambda x: x.get("position", 0)
+    )
+    if filter_list:
+        results = get_filtered_results(results, filter_list)
+    return [
+        SearchResult(
+            link=result["link"], title=result["title"], snippet=result["snippet"]
+        )
+        for result in results[:count]
+    ]

backend/open_webui/apps/retrieval/web/searxng.py

@@ -1,10 +1,9 @@
 import logging
+from typing import Optional
+
 import requests
-
-from typing import List, Optional
-
-from apps.rag.search.main import SearchResult, get_filtered_results
-from config import SRC_LOG_LEVELS
+from open_webui.apps.retrieval.web.main import SearchResult, get_filtered_results
+from open_webui.env import SRC_LOG_LEVELS
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["RAG"])
@@ -14,9 +13,9 @@ def search_searxng(
     query_url: str,
     query: str,
     count: int,
-    filter_list: Optional[List[str]] = None,
+    filter_list: Optional[list[str]] = None,
     **kwargs,
-) -> List[SearchResult]:
+) -> list[SearchResult]:
     """
     Search a SearXNG instance for a given query and return the results as a list of SearchResult objects.
 
@@ -31,10 +30,10 @@ def search_searxng(
         language (str): Language filter for the search results; e.g., "en-US". Defaults to an empty string.
         safesearch (int): Safe search filter for safer web results; 0 = off, 1 = moderate, 2 = strict. Defaults to 1 (moderate).
         time_range (str): Time range for filtering results by date; e.g., "2023-04-05..today" or "all-time". Defaults to ''.
-        categories: (Optional[List[str]]): Specific categories within which the search should be performed, defaulting to an empty string if not provided.
+        categories: (Optional[list[str]]): Specific categories within which the search should be performed, defaulting to an empty string if not provided.
 
     Returns:
-        List[SearchResult]: A list of SearchResults sorted by relevance score in descending order.
+        list[SearchResult]: A list of SearchResults sorted by relevance score in descending order.
 
     Raise:
         requests.exceptions.RequestException: If a request error occurs during the search process.

backend/open_webui/apps/retrieval/web/serper.py

@@ -1,17 +1,17 @@
 import json
 import logging
-from typing import List, Optional
-import requests
+from typing import Optional
 
-from apps.rag.search.main import SearchResult, get_filtered_results
-from config import SRC_LOG_LEVELS
+import requests
+from open_webui.apps.retrieval.web.main import SearchResult, get_filtered_results
+from open_webui.env import SRC_LOG_LEVELS
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["RAG"])
 
 
 def search_serper(
-    api_key: str, query: str, count: int, filter_list: Optional[List[str]] = None
+    api_key: str, query: str, count: int, filter_list: Optional[list[str]] = None
 ) -> list[SearchResult]:
     """Search using serper.dev's API and return the results as a list of SearchResult objects.
 

backend/open_webui/apps/retrieval/web/serply.py

@@ -1,11 +1,10 @@
-import json
 import logging
-from typing import List, Optional
-import requests
+from typing import Optional
 from urllib.parse import urlencode
 
-from apps.rag.search.main import SearchResult, get_filtered_results
-from config import SRC_LOG_LEVELS
+import requests
+from open_webui.apps.retrieval.web.main import SearchResult, get_filtered_results
+from open_webui.env import SRC_LOG_LEVELS
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["RAG"])
@@ -19,7 +18,7 @@ def search_serply(
     limit: int = 10,
     device_type: str = "desktop",
     proxy_location: str = "US",
-    filter_list: Optional[List[str]] = None,
+    filter_list: Optional[list[str]] = None,
 ) -> list[SearchResult]:
     """Search using serper.dev's API and return the results as a list of SearchResult objects.
 

backend/open_webui/apps/retrieval/web/serpstack.py

@@ -1,10 +1,9 @@
-import json
 import logging
-from typing import List, Optional
-import requests
+from typing import Optional
 
-from apps.rag.search.main import SearchResult, get_filtered_results
-from config import SRC_LOG_LEVELS
+import requests
+from open_webui.apps.retrieval.web.main import SearchResult, get_filtered_results
+from open_webui.env import SRC_LOG_LEVELS
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["RAG"])
@@ -14,7 +13,7 @@ def search_serpstack(
     api_key: str,
     query: str,
     count: int,
-    filter_list: Optional[List[str]] = None,
+    filter_list: Optional[list[str]] = None,
     https_enabled: bool = True,
 ) -> list[SearchResult]:
     """Search using serpstack.com's and return the results as a list of SearchResult objects.

backend/open_webui/apps/retrieval/web/tavily.py

@@ -1,9 +1,8 @@
 import logging
 
 import requests
-
-from apps.rag.search.main import SearchResult
-from config import SRC_LOG_LEVELS
+from open_webui.apps.retrieval.web.main import SearchResult
+from open_webui.env import SRC_LOG_LEVELS
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["RAG"])
@@ -17,7 +16,7 @@ def search_tavily(api_key: str, query: str, count: int) -> list[SearchResult]:
         query (str): The query to search for
 
     Returns:
-        List[SearchResult]: A list of search results
+        list[SearchResult]: A list of search results
     """
     url = "https://api.tavily.com/search"
     data = {"query": query, "api_key": api_key}

backend/open_webui/apps/retrieval/web/testdata/bing.json

@@ -0,0 +1,58 @@
+{
+	"_type": "SearchResponse",
+	"queryContext": {
+		"originalQuery": "Top 10 international results"
+	},
+	"webPages": {
+		"webSearchUrl": "https://www.bing.com/search?q=Top+10+international+results",
+		"totalEstimatedMatches": 687,
+		"value": [
+			{
+				"id": "https://api.bing.microsoft.com/api/v7/#WebPages.0",
+				"name": "2024 Mexican Grand Prix - F1 results and latest standings ... - PlanetF1",
+				"url": "https://www.planetf1.com/news/f1-results-2024-mexican-grand-prix-race-standings",
+				"datePublished": "2024-10-27T00:00:00.0000000",
+				"datePublishedFreshnessText": "1 day ago",
+				"isFamilyFriendly": true,
+				"displayUrl": "https://www.planetf1.com/news/f1-results-2024-mexican-grand-prix-race-standings",
+				"snippet": "Nico Hulkenberg and Pierre Gasly completed the top 10. A full report of the Mexican Grand Prix is available at the bottom of this article. F1 results – 2024 Mexican Grand Prix",
+				"dateLastCrawled": "2024-10-28T07:15:00.0000000Z",
+				"cachedPageUrl": "https://cc.bingj.com/cache.aspx?q=Top+10+international+results&d=916492551782&mkt=en-US&setlang=en-US&w=zBsfaAPyF2tUrHFHr_vFFdUm8sng4g34",
+				"language": "en",
+				"isNavigational": false,
+				"noCache": false
+			},
+			{
+				"id": "https://api.bing.microsoft.com/api/v7/#WebPages.1",
+				"name": "F1 Results Today: HUGE Verstappen penalties cause major title change",
+				"url": "https://www.gpfans.com/en/f1-news/1033512/f1-results-today-mexican-grand-prix-huge-max-verstappen-penalties-cause-major-title-change/",
+				"datePublished": "2024-10-27T00:00:00.0000000",
+				"datePublishedFreshnessText": "1 day ago",
+				"isFamilyFriendly": true,
+				"displayUrl": "https://www.gpfans.com/en/f1-news/1033512/f1-results-today-mexican-grand-prix-huge-max...",
+				"snippet": "Elsewhere, Mercedes duo Lewis Hamilton and George Russell came home in P4 and P5 respectively. Meanwhile, the surprise package of the day were Haas, with both Kevin Magnussen and Nico Hulkenberg finishing inside the points.. READ MORE: RB star issues apology after red flag CRASH at Mexican GP Mexican Grand Prix 2024 results. 1. Carlos Sainz [Ferrari] 2. Lando Norris [McLaren] - +4.705",
+				"dateLastCrawled": "2024-10-28T06:06:00.0000000Z",
+				"cachedPageUrl": "https://cc.bingj.com/cache.aspx?q=Top+10+international+results&d=2840656522642&mkt=en-US&setlang=en-US&w=-Tbkwxnq52jZCvG7l3CtgcwT1vwAjIUD",
+				"language": "en",
+				"isNavigational": false,
+				"noCache": false
+			},
+			{
+				"id": "https://api.bing.microsoft.com/api/v7/#WebPages.2",
+				"name": "International Power Rankings: England flying, Kangaroos cruising, Fiji rise",
+				"url": "https://www.loverugbyleague.com/post/international-power-rankings-england-flying-kangaroos-cruising-fiji-rise",
+				"datePublished": "2024-10-28T00:00:00.0000000",
+				"datePublishedFreshnessText": "7 hours ago",
+				"isFamilyFriendly": true,
+				"displayUrl": "https://www.loverugbyleague.com/post/international-power-rankings-england-flying...",
+				"snippet": "LRL RECOMMENDS: England player ratings from first Test against Samoa as omnificent George Williams scores perfect 10. 2. Australia (Men) – SAME. The Kangaroos remain 2nd in our Power Rankings after their 22-10 win against New Zealand in Christchurch on Sunday. As was the case in their win against Tonga last week, Mal Meninga’s side weren ...",
+				"dateLastCrawled": "2024-10-28T07:09:00.0000000Z",
+				"cachedPageUrl": "https://cc.bingj.com/cache.aspx?q=Top+10+international+results&d=1535008462672&mkt=en-US&setlang=en-US&w=82ujhH4Kp0iuhCS7wh1xLUFYUeetaVVm",
+				"language": "en",
+				"isNavigational": false,
+				"noCache": false
+			}
+		],
+		"someResultsRemoved": true
+	}
+}

backend/open_webui/apps/retrieval/web/utils.py

@@ -0,0 +1,97 @@
+import socket
+import urllib.parse
+import validators
+from typing import Union, Sequence, Iterator
+
+from langchain_community.document_loaders import (
+    WebBaseLoader,
+)
+from langchain_core.documents import Document
+
+
+from open_webui.constants import ERROR_MESSAGES
+from open_webui.config import ENABLE_RAG_LOCAL_WEB_FETCH
+from open_webui.env import SRC_LOG_LEVELS
+
+import logging
+
+log = logging.getLogger(__name__)
+log.setLevel(SRC_LOG_LEVELS["RAG"])
+
+
+def validate_url(url: Union[str, Sequence[str]]):
+    if isinstance(url, str):
+        if isinstance(validators.url(url), validators.ValidationError):
+            raise ValueError(ERROR_MESSAGES.INVALID_URL)
+        if not ENABLE_RAG_LOCAL_WEB_FETCH:
+            # Local web fetch is disabled, filter out any URLs that resolve to private IP addresses
+            parsed_url = urllib.parse.urlparse(url)
+            # Get IPv4 and IPv6 addresses
+            ipv4_addresses, ipv6_addresses = resolve_hostname(parsed_url.hostname)
+            # Check if any of the resolved addresses are private
+            # This is technically still vulnerable to DNS rebinding attacks, as we don't control WebBaseLoader
+            for ip in ipv4_addresses:
+                if validators.ipv4(ip, private=True):
+                    raise ValueError(ERROR_MESSAGES.INVALID_URL)
+            for ip in ipv6_addresses:
+                if validators.ipv6(ip, private=True):
+                    raise ValueError(ERROR_MESSAGES.INVALID_URL)
+        return True
+    elif isinstance(url, Sequence):
+        return all(validate_url(u) for u in url)
+    else:
+        return False
+
+
+def resolve_hostname(hostname):
+    # Get address information
+    addr_info = socket.getaddrinfo(hostname, None)
+
+    # Extract IP addresses from address information
+    ipv4_addresses = [info[4][0] for info in addr_info if info[0] == socket.AF_INET]
+    ipv6_addresses = [info[4][0] for info in addr_info if info[0] == socket.AF_INET6]
+
+    return ipv4_addresses, ipv6_addresses
+
+
+class SafeWebBaseLoader(WebBaseLoader):
+    """WebBaseLoader with enhanced error handling for URLs."""
+
+    def lazy_load(self) -> Iterator[Document]:
+        """Lazy load text from the url(s) in web_path with error handling."""
+        for path in self.web_paths:
+            try:
+                soup = self._scrape(path, bs_kwargs=self.bs_kwargs)
+                text = soup.get_text(**self.bs_get_text_kwargs)
+
+                # Build metadata
+                metadata = {"source": path}
+                if title := soup.find("title"):
+                    metadata["title"] = title.get_text()
+                if description := soup.find("meta", attrs={"name": "description"}):
+                    metadata["description"] = description.get(
+                        "content", "No description found."
+                    )
+                if html := soup.find("html"):
+                    metadata["language"] = html.get("lang", "No language found.")
+
+                yield Document(page_content=text, metadata=metadata)
+            except Exception as e:
+                # Log the error and continue with the next URL
+                log.error(f"Error loading {path}: {e}")
+
+
+def get_web_loader(
+    url: Union[str, Sequence[str]],
+    verify_ssl: bool = True,
+    requests_per_second: int = 2,
+):
+    # Check if the URL is valid
+    if not validate_url(url):
+        raise ValueError(ERROR_MESSAGES.INVALID_URL)
+    return SafeWebBaseLoader(
+        url,
+        verify_ssl=verify_ssl,
+        requests_per_second=requests_per_second,
+        continue_on_failure=True,
+    )

backend/open_webui/apps/socket/main.py

@@ -0,0 +1,221 @@
+# TODO: move socket to webui app
+
+import asyncio
+import socketio
+import logging
+import sys
+import time
+
+from open_webui.apps.webui.models.users import Users
+from open_webui.env import (
+    ENABLE_WEBSOCKET_SUPPORT,
+    WEBSOCKET_MANAGER,
+    WEBSOCKET_REDIS_URL,
+)
+from open_webui.utils.utils import decode_token
+from open_webui.apps.socket.utils import RedisDict
+
+from open_webui.env import (
+    GLOBAL_LOG_LEVEL,
+    SRC_LOG_LEVELS,
+)
+
+
+logging.basicConfig(stream=sys.stdout, level=GLOBAL_LOG_LEVEL)
+log = logging.getLogger(__name__)
+log.setLevel(SRC_LOG_LEVELS["SOCKET"])
+
+
+if WEBSOCKET_MANAGER == "redis":
+    mgr = socketio.AsyncRedisManager(WEBSOCKET_REDIS_URL)
+    sio = socketio.AsyncServer(
+        cors_allowed_origins=[],
+        async_mode="asgi",
+        transports=(
+            ["polling", "websocket"] if ENABLE_WEBSOCKET_SUPPORT else ["polling"]
+        ),
+        allow_upgrades=ENABLE_WEBSOCKET_SUPPORT,
+        always_connect=True,
+        client_manager=mgr,
+    )
+else:
+    sio = socketio.AsyncServer(
+        cors_allowed_origins=[],
+        async_mode="asgi",
+        transports=(
+            ["polling", "websocket"] if ENABLE_WEBSOCKET_SUPPORT else ["polling"]
+        ),
+        allow_upgrades=ENABLE_WEBSOCKET_SUPPORT,
+        always_connect=True,
+    )
+
+
+# Dictionary to maintain the user pool
+
+if WEBSOCKET_MANAGER == "redis":
+    SESSION_POOL = RedisDict("open-webui:session_pool", redis_url=WEBSOCKET_REDIS_URL)
+    USER_POOL = RedisDict("open-webui:user_pool", redis_url=WEBSOCKET_REDIS_URL)
+    USAGE_POOL = RedisDict("open-webui:usage_pool", redis_url=WEBSOCKET_REDIS_URL)
+else:
+    SESSION_POOL = {}
+    USER_POOL = {}
+    USAGE_POOL = {}
+
+
+# Timeout duration in seconds
+TIMEOUT_DURATION = 3
+
+
+async def periodic_usage_pool_cleanup():
+    while True:
+        now = int(time.time())
+        for model_id, connections in list(USAGE_POOL.items()):
+            # Creating a list of sids to remove if they have timed out
+            expired_sids = [
+                sid
+                for sid, details in connections.items()
+                if now - details["updated_at"] > TIMEOUT_DURATION
+            ]
+
+            for sid in expired_sids:
+                del connections[sid]
+
+            if not connections:
+                log.debug(f"Cleaning up model {model_id} from usage pool")
+                del USAGE_POOL[model_id]
+            else:
+                USAGE_POOL[model_id] = connections
+
+            # Emit updated usage information after cleaning
+            await sio.emit("usage", {"models": get_models_in_use()})
+
+        await asyncio.sleep(TIMEOUT_DURATION)
+
+
+app = socketio.ASGIApp(
+    sio,
+    socketio_path="/ws/socket.io",
+)
+
+
+def get_models_in_use():
+    # List models that are currently in use
+    models_in_use = list(USAGE_POOL.keys())
+    return models_in_use
+
+
+@sio.on("usage")
+async def usage(sid, data):
+    model_id = data["model"]
+    # Record the timestamp for the last update
+    current_time = int(time.time())
+
+    # Store the new usage data and task
+    USAGE_POOL[model_id] = {
+        **(USAGE_POOL[model_id] if model_id in USAGE_POOL else {}),
+        sid: {"updated_at": current_time},
+    }
+
+    # Broadcast the usage data to all clients
+    await sio.emit("usage", {"models": get_models_in_use()})
+
+
+@sio.event
+async def connect(sid, environ, auth):
+    user = None
+    if auth and "token" in auth:
+        data = decode_token(auth["token"])
+
+        if data is not None and "id" in data:
+            user = Users.get_user_by_id(data["id"])
+
+        if user:
+            SESSION_POOL[sid] = user.id
+            if user.id in USER_POOL:
+                USER_POOL[user.id].append(sid)
+            else:
+                USER_POOL[user.id] = [sid]
+
+            # print(f"user {user.name}({user.id}) connected with session ID {sid}")
+            await sio.emit("user-count", {"count": len(USER_POOL.items())})
+            await sio.emit("usage", {"models": get_models_in_use()})
+
+
+@sio.on("user-join")
+async def user_join(sid, data):
+    # print("user-join", sid, data)
+
+    auth = data["auth"] if "auth" in data else None
+    if not auth or "token" not in auth:
+        return
+
+    data = decode_token(auth["token"])
+    if data is None or "id" not in data:
+        return
+
+    user = Users.get_user_by_id(data["id"])
+    if not user:
+        return
+
+    SESSION_POOL[sid] = user.id
+    if user.id in USER_POOL:
+        USER_POOL[user.id].append(sid)
+    else:
+        USER_POOL[user.id] = [sid]
+
+    # print(f"user {user.name}({user.id}) connected with session ID {sid}")
+
+    await sio.emit("user-count", {"count": len(USER_POOL.items())})
+
+
+@sio.on("user-count")
+async def user_count(sid):
+    await sio.emit("user-count", {"count": len(USER_POOL.items())})
+
+
+@sio.event
+async def disconnect(sid):
+    if sid in SESSION_POOL:
+        user_id = SESSION_POOL[sid]
+        del SESSION_POOL[sid]
+
+        USER_POOL[user_id] = [_sid for _sid in USER_POOL[user_id] if _sid != sid]
+
+        if len(USER_POOL[user_id]) == 0:
+            del USER_POOL[user_id]
+
+        await sio.emit("user-count", {"count": len(USER_POOL)})
+    else:
+        pass
+        # print(f"Unknown session ID {sid} disconnected")
+
+
+def get_event_emitter(request_info):
+    async def __event_emitter__(event_data):
+        await sio.emit(
+            "chat-events",
+            {
+                "chat_id": request_info["chat_id"],
+                "message_id": request_info["message_id"],
+                "data": event_data,
+            },
+            to=request_info["session_id"],
+        )
+
+    return __event_emitter__
+
+
+def get_event_call(request_info):
+    async def __event_call__(event_data):
+        response = await sio.call(
+            "chat-events",
+            {
+                "chat_id": request_info["chat_id"],
+                "message_id": request_info["message_id"],
+                "data": event_data,
+            },
+            to=request_info["session_id"],
+        )
+        return response
+
+    return __event_call__

backend/open_webui/apps/socket/utils.py

@@ -0,0 +1,59 @@
+import json
+import redis
+
+
+class RedisDict:
+    def __init__(self, name, redis_url):
+        self.name = name
+        self.redis = redis.Redis.from_url(redis_url, decode_responses=True)
+
+    def __setitem__(self, key, value):
+        serialized_value = json.dumps(value)
+        self.redis.hset(self.name, key, serialized_value)
+
+    def __getitem__(self, key):
+        value = self.redis.hget(self.name, key)
+        if value is None:
+            raise KeyError(key)
+        return json.loads(value)
+
+    def __delitem__(self, key):
+        result = self.redis.hdel(self.name, key)
+        if result == 0:
+            raise KeyError(key)
+
+    def __contains__(self, key):
+        return self.redis.hexists(self.name, key)
+
+    def __len__(self):
+        return self.redis.hlen(self.name)
+
+    def keys(self):
+        return self.redis.hkeys(self.name)
+
+    def values(self):
+        return [json.loads(v) for v in self.redis.hvals(self.name)]
+
+    def items(self):
+        return [(k, json.loads(v)) for k, v in self.redis.hgetall(self.name).items()]
+
+    def get(self, key, default=None):
+        try:
+            return self[key]
+        except KeyError:
+            return default
+
+    def clear(self):
+        self.redis.delete(self.name)
+
+    def update(self, other=None, **kwargs):
+        if other is not None:
+            for k, v in other.items() if hasattr(other, "items") else other:
+                self[k] = v
+        for k, v in kwargs.items():
+            self[k] = v
+
+    def setdefault(self, key, default=None):
+        if key not in self:
+            self[key] = default
+        return self[key]

backend/open_webui/apps/webui/internal/db.py

@@ -1,20 +1,25 @@
-import os
-import logging
 import json
+import logging
 from contextlib import contextmanager
+from typing import Any, Optional
 
+from open_webui.apps.webui.internal.wrappers import register_connection
+from open_webui.env import (
+    OPEN_WEBUI_DIR,
+    DATABASE_URL,
+    SRC_LOG_LEVELS,
+    DATABASE_POOL_MAX_OVERFLOW,
+    DATABASE_POOL_RECYCLE,
+    DATABASE_POOL_SIZE,
+    DATABASE_POOL_TIMEOUT,
+)
 from peewee_migrate import Router
-from apps.webui.internal.wrappers import register_connection
-
-from typing import Optional, Any
-from typing_extensions import Self
-
-from sqlalchemy import create_engine, types, Dialect
+from sqlalchemy import Dialect, create_engine, types
 from sqlalchemy.ext.declarative import declarative_base
-from sqlalchemy.orm import sessionmaker, scoped_session
+from sqlalchemy.orm import scoped_session, sessionmaker
+from sqlalchemy.pool import QueuePool, NullPool
 from sqlalchemy.sql.type_api import _T
-
-from config import SRC_LOG_LEVELS, DATA_DIR, DATABASE_URL, BACKEND_DIR
+from typing_extensions import Self
 
 log = logging.getLogger(__name__)
 log.setLevel(SRC_LOG_LEVELS["DB"])
@@ -42,34 +47,21 @@ class JSONField(types.TypeDecorator):
             return json.loads(value)
 
 
-# Check if the file exists
-if os.path.exists(f"{DATA_DIR}/ollama.db"):
-    # Rename the file
-    os.rename(f"{DATA_DIR}/ollama.db", f"{DATA_DIR}/webui.db")
-    log.info("Database migrated from Ollama-WebUI successfully.")
-else:
-    pass
-
-
 # Workaround to handle the peewee migration
 # This is required to ensure the peewee migration is handled before the alembic migration
 def handle_peewee_migration(DATABASE_URL):
+    # db = None
     try:
-        # Replace the postgresql:// with postgres:// and %40 with @ in the DATABASE_URL
-        db = register_connection(
-            DATABASE_URL.replace("postgresql://", "postgres://").replace("%40", "@")
-        )
-        migrate_dir = BACKEND_DIR / "apps" / "webui" / "internal" / "migrations"
+        # Replace the postgresql:// with postgres:// to handle the peewee migration
+        db = register_connection(DATABASE_URL.replace("postgresql://", "postgres://"))
+        migrate_dir = OPEN_WEBUI_DIR / "apps" / "webui" / "internal" / "migrations"
         router = Router(db, logger=log, migrate_dir=migrate_dir)
         router.run()
         db.close()
 
-        # check if db connection has been closed
-
     except Exception as e:
         log.error(f"Failed to initialize the database connection: {e}")
         raise
-
     finally:
         # Properly closing the database connection
         if db and not db.is_closed():
@@ -88,7 +80,20 @@ if "sqlite" in SQLALCHEMY_DATABASE_URL:
         SQLALCHEMY_DATABASE_URL, connect_args={"check_same_thread": False}
     )
 else:
-    engine = create_engine(SQLALCHEMY_DATABASE_URL, pool_pre_ping=True)
+    if DATABASE_POOL_SIZE > 0:
+        engine = create_engine(
+            SQLALCHEMY_DATABASE_URL,
+            pool_size=DATABASE_POOL_SIZE,
+            max_overflow=DATABASE_POOL_MAX_OVERFLOW,
+            pool_timeout=DATABASE_POOL_TIMEOUT,
+            pool_recycle=DATABASE_POOL_RECYCLE,
+            pool_pre_ping=True,
+            poolclass=QueuePool,
+        )
+    else:
+        engine = create_engine(
+            SQLALCHEMY_DATABASE_URL, pool_pre_ping=True, poolclass=NullPool
+        )
 
 
 SessionLocal = sessionmaker(
@@ -98,7 +103,6 @@ Base = declarative_base()
 Session = scoped_session(SessionLocal)
 
 
-# Dependency
 def get_session():
     db = SessionLocal()
     try:

backend/open_webui/apps/webui/internal/migrations/010_migrate_modelfiles_to_models.py

@@ -30,7 +30,7 @@ import peewee as pw
 from peewee_migrate import Migrator
 import json
 
-from utils.misc import parse_ollama_modelfile
+from open_webui.utils.misc import parse_ollama_modelfile
 
 with suppress(ImportError):
     import playhouse.postgres_ext as pw_pext