mirror of
https://github.com/harvard-edge/cs249r_book.git
synced 2026-04-29 00:59:07 -05:00
10a2dc4303
Quarto's Lua filter calls rsvg-convert to convert SVG figures to PDF during PDF builds. librsvg2-dev was present (C headers/lib) but the binary package librsvg2-bin was missing, causing a FATAL build error: 'Could not convert a SVG to a PDF. Please ensure rsvg-convert is on path' Also adds rsvg-convert to the Phase 2 verification checks so missing tools are caught at image build time, not at render time.
761 lines
28 KiB
Docker
761 lines
28 KiB
Docker
# MLSysBook Quarto Build Container
|
|
# Based on Ubuntu 22.04 with all dependencies pre-installed
|
|
# This container eliminates the 30-45 minute setup time for Linux builds
|
|
|
|
FROM ubuntu:22.04
|
|
|
|
# === PATH CONFIGURATION ===
|
|
# Paths are hardcoded since they're stable and ARG scope causes issues in multi-layer builds
|
|
|
|
# Set environment variables
|
|
ENV DEBIAN_FRONTEND=noninteractive
|
|
ENV R_LIBS_USER=/usr/local/lib/R/library
|
|
ENV QUARTO_LOG_LEVEL=INFO
|
|
ENV PYTHONIOENCODING=utf-8
|
|
ENV LANG=en_US.UTF-8
|
|
ENV LC_ALL=en_US.UTF-8
|
|
ENV PATH=/usr/local/texlive/bin/x86_64-linux:$PATH
|
|
|
|
# === PHASE 0: COPY DEPENDENCY FILES EARLY (for better cache efficiency) ===
|
|
# Copy dependency files early for better cache efficiency
|
|
COPY book/tools/dependencies/requirements.txt /tmp/requirements.txt
|
|
COPY book/tools/dependencies/install_packages.R /tmp/install_packages.R
|
|
COPY book/tools/dependencies/tl_packages /tmp/tl_packages
|
|
COPY book/docker/linux/verify_r_packages.R /tmp/verify_r_packages.R
|
|
|
|
# Note: Python packages are installed later in Phase 10
|
|
|
|
# Early cleanup
|
|
RUN rm -rf /var/lib/apt/lists/* && \
|
|
apt-get clean
|
|
|
|
# === PHASE 1: LOCALE CONFIGURATION ===
|
|
RUN echo "🚀 === STARTING LOCALE CONFIGURATION ===" && \
|
|
echo "🔍 Checking system readiness..." && \
|
|
if [ -f /etc/os-release ]; then \
|
|
echo "✅ OS release file found"; \
|
|
cat /etc/os-release | grep PRETTY_NAME; \
|
|
else \
|
|
echo "❌ OS release file not found"; \
|
|
exit 1; \
|
|
fi && \
|
|
echo "📦 Installing locales package..." && \
|
|
apt-get update && apt-get install -y locales && \
|
|
echo "📦 Locales package installed" && \
|
|
echo "🔧 Generating en_US.UTF-8 locale..." && \
|
|
locale-gen en_US.UTF-8 && \
|
|
echo "📄 Locale generated" && \
|
|
echo "🔧 Updating system locale..." && \
|
|
update-locale LANG=en_US.UTF-8 && \
|
|
echo "🔧 System locale updated" && \
|
|
echo "🧹 Cleaning package cache..." && \
|
|
rm -rf /var/lib/apt/lists/* && \
|
|
echo "✅ Locale configuration complete"
|
|
|
|
# === PHASE 2: SYSTEM DEPENDENCIES ===
|
|
RUN echo "🚀 === STARTING SYSTEM DEPENDENCIES INSTALLATION ===" && \
|
|
echo "⏰ Estimated time: 2-3 minutes" && \
|
|
echo "📊 Free disk space: $(df -h / | tail -1 | awk '{print $4}')" && \
|
|
start_time=$(date +%s) && \
|
|
\
|
|
echo "🔄 Updating package lists..." && \
|
|
apt-get update && \
|
|
\
|
|
echo "📦 Installing core system packages (25 packages)..." && \
|
|
echo "📋 Package list:" && \
|
|
echo " - fonts-dejavu" && \
|
|
echo " - fonts-freefont-ttf" && \
|
|
echo " - gdk-pixbuf2.0-bin" && \
|
|
echo " - libcairo2" && \
|
|
echo " - libfontconfig1" && \
|
|
echo " - libfontconfig1-dev" && \
|
|
echo " - libfreetype6" && \
|
|
echo " - libfreetype6-dev" && \
|
|
echo " - libpango-1.0-0" && \
|
|
echo " - libpangocairo-1.0-0" && \
|
|
echo " - libpangoft2-1.0-0" && \
|
|
echo " - libxml2-dev" && \
|
|
echo " - libcurl4-openssl-dev" && \
|
|
echo " - libjpeg-dev" && \
|
|
echo " - libtiff5-dev" && \
|
|
echo " - libpng-dev" && \
|
|
echo " - libharfbuzz-dev" && \
|
|
echo " - libfribidi-dev" && \
|
|
echo " - librsvg2-dev" && \
|
|
echo " - librsvg2-bin" && \
|
|
echo " - libgdal-dev" && \
|
|
echo " - libudunits2-dev" && \
|
|
echo " - wget" && \
|
|
echo " - curl" && \
|
|
echo " - git" && \
|
|
apt-get install -y \
|
|
fonts-dejavu \
|
|
fonts-freefont-ttf \
|
|
gdk-pixbuf2.0-bin \
|
|
libcairo2 \
|
|
libfontconfig1 \
|
|
libfontconfig1-dev \
|
|
libfreetype6 \
|
|
libfreetype6-dev \
|
|
libpango-1.0-0 \
|
|
libpangocairo-1.0-0 \
|
|
libpangoft2-1.0-0 \
|
|
libxml2-dev \
|
|
libcurl4-openssl-dev \
|
|
libjpeg-dev \
|
|
libtiff5-dev \
|
|
libpng-dev \
|
|
libharfbuzz-dev \
|
|
libfribidi-dev \
|
|
librsvg2-dev \
|
|
librsvg2-bin \
|
|
libgdal-dev \
|
|
libudunits2-dev \
|
|
wget \
|
|
curl \
|
|
git && \
|
|
echo "📦 All system packages installed successfully" && \
|
|
\
|
|
echo "🔍 Verifying critical packages..." && \
|
|
if command -v wget >/dev/null 2>&1; then \
|
|
echo "📦 wget available"; \
|
|
else \
|
|
echo "❌ wget not found"; \
|
|
exit 1; \
|
|
fi && \
|
|
if command -v curl >/dev/null 2>&1; then \
|
|
echo "📦 curl available"; \
|
|
else \
|
|
echo "❌ curl not found"; \
|
|
exit 1; \
|
|
fi && \
|
|
if command -v git >/dev/null 2>&1; then \
|
|
echo "📦 git available"; \
|
|
else \
|
|
echo "❌ git not found"; \
|
|
exit 1; \
|
|
fi && \
|
|
if command -v rsvg-convert >/dev/null 2>&1; then \
|
|
echo "📦 rsvg-convert available"; \
|
|
else \
|
|
echo "❌ rsvg-convert not found"; \
|
|
exit 1; \
|
|
fi && \
|
|
\
|
|
echo "🧹 Cleaning package cache..." && \
|
|
rm -rf /var/lib/apt/lists/* && \
|
|
\
|
|
end_time=$(date +%s) && \
|
|
duration=$((end_time - start_time)) && \
|
|
echo "✅ === SYSTEM DEPENDENCIES COMPLETE === (${duration}s)" && \
|
|
echo "📊 Free disk space: $(df -h / | tail -1 | awk '{print $4}')"
|
|
|
|
# === PHASE 3: INKSCAPE INSTALLATION ===
|
|
RUN echo "🚀 === STARTING INKSCAPE INSTALLATION ===" && \
|
|
echo "⏰ Estimated time: 1-2 minutes" && \
|
|
start_time=$(date +%s) && \
|
|
\
|
|
echo "🔄 Adding Inkscape PPA repository..." && \
|
|
apt-get update && \
|
|
echo "📦 Installing software-properties-common..." && \
|
|
apt-get install -y software-properties-common && \
|
|
echo "📦 software-properties-common installed" && \
|
|
echo "🔧 Adding Inkscape PPA..." && \
|
|
add-apt-repository ppa:inkscape.dev/stable -y && \
|
|
echo "📦 Inkscape PPA added" && \
|
|
\
|
|
echo "📦 Installing Inkscape..." && \
|
|
apt-get update && \
|
|
apt-get install -y inkscape && \
|
|
echo "📦 Inkscape installed" && \
|
|
\
|
|
echo "🧹 Cleaning package cache..." && \
|
|
rm -rf /var/lib/apt/lists/* && \
|
|
\
|
|
end_time=$(date +%s) && \
|
|
duration=$((end_time - start_time)) && \
|
|
echo "✅ === INKSCAPE INSTALLATION COMPLETE === (${duration}s)"
|
|
|
|
# Install font dependencies (note: fonts-freefont-ttf already installed above)
|
|
RUN echo "🚀 === STARTING FONT INSTALLATION ===" && \
|
|
echo "📦 Installing additional fonts..." && \
|
|
apt-get update && apt-get install -y \
|
|
fonts-liberation \
|
|
fontconfig && \
|
|
echo "📦 Fonts installed" && \
|
|
echo "🔧 Updating font cache..." && \
|
|
fc-cache -fv && \
|
|
echo "📄 Font cache updated" && \
|
|
echo "🧹 Cleaning package cache..." && \
|
|
rm -rf /var/lib/apt/lists/* && \
|
|
echo "✅ Font installation complete"
|
|
|
|
# Test Inkscape SVG to PDF conversion (same as your workflow)
|
|
RUN echo "🚀 === STARTING INKSCAPE TEST ===" && \
|
|
echo "📋 Creating test SVG file..." && \
|
|
echo '<svg xmlns="http://www.w3.org/2000/svg" width="100" height="100"><circle cx="50" cy="50" r="40" fill="red"/></svg>' > test.svg && \
|
|
echo "📄 Test SVG created" && \
|
|
echo "🔄 Converting SVG to PDF..." && \
|
|
inkscape --export-type=pdf --export-filename=test.pdf test.svg && \
|
|
echo "📦 Conversion completed" && \
|
|
if [ -f test.pdf ]; then \
|
|
echo "✅ Inkscape SVG to PDF conversion successful!"; \
|
|
echo "📊 PDF file details:"; \
|
|
ls -lh test.pdf; \
|
|
else \
|
|
echo "❌ Inkscape SVG to PDF conversion failed."; \
|
|
exit 1; \
|
|
fi && \
|
|
echo "🧹 Cleaning up test files..." && \
|
|
rm -f test.svg test.pdf && \
|
|
echo "✅ Inkscape test complete"
|
|
|
|
# === PHASE 4: GHOSTSCRIPT INSTALLATION ===
|
|
RUN echo "🚀 === STARTING GHOSTSCRIPT INSTALLATION ===" && \
|
|
echo "⏰ Estimated time: 30 seconds" && \
|
|
start_time=$(date +%s) && \
|
|
\
|
|
echo "📦 Installing Ghostscript..." && \
|
|
apt-get update && apt-get install -y ghostscript && \
|
|
echo "📦 Ghostscript installed" && \
|
|
echo "📊 Ghostscript version:" && \
|
|
gs --version && \
|
|
\
|
|
echo "🧹 Cleaning package cache..." && \
|
|
rm -rf /var/lib/apt/lists/* && \
|
|
\
|
|
end_time=$(date +%s) && \
|
|
duration=$((end_time - start_time)) && \
|
|
echo "✅ === GHOSTSCRIPT INSTALLATION COMPLETE === (${duration}s)"
|
|
|
|
# === PHASE 5: TEX LIVE INSTALLATION ===
|
|
# Step 5.1: Install TeX Live prerequisites
|
|
RUN echo "🚀 === STARTING TEX LIVE INSTALLATION ===" && \
|
|
echo "⏰ Estimated time: 8-12 minutes (largest phase)" && \
|
|
echo "📊 Free disk space before: $(df -h / | tail -1 | awk '{print $4}')" && \
|
|
\
|
|
echo "📦 Installing TeX Live prerequisites..." && \
|
|
echo "📋 Prerequisites:" && \
|
|
echo " - perl" && \
|
|
echo " - wget" && \
|
|
echo " - xzdec" && \
|
|
apt-get update && apt-get install -y \
|
|
perl \
|
|
wget \
|
|
xzdec && \
|
|
echo "📦 Prerequisites installed" && \
|
|
\
|
|
echo "🔍 Verifying prerequisites..." && \
|
|
if command -v perl >/dev/null 2>&1; then \
|
|
echo "📦 perl available"; \
|
|
else \
|
|
echo "❌ perl not found"; \
|
|
exit 1; \
|
|
fi && \
|
|
if command -v wget >/dev/null 2>&1; then \
|
|
echo "📦 wget available"; \
|
|
else \
|
|
echo "❌ wget not found"; \
|
|
exit 1; \
|
|
fi && \
|
|
if command -v xzdec >/dev/null 2>&1; then \
|
|
echo "📦 xzdec available"; \
|
|
else \
|
|
echo "❌ xzdec not found"; \
|
|
exit 1; \
|
|
fi && \
|
|
\
|
|
rm -rf /var/lib/apt/lists/* && \
|
|
echo "✅ TeX Live prerequisites installed"
|
|
|
|
# Step 5.2: Download and extract TeX Live installer with mirror fallbacks
|
|
RUN echo "🔄 Downloading TeX Live installer (~4MB) with mirror fallbacks..." && \
|
|
SUCCESS=false && \
|
|
for mirror in \
|
|
"https://mirror.ctan.org/systems/texlive/tlnet/install-tl-unx.tar.gz" \
|
|
"https://mirrors.mit.edu/CTAN/systems/texlive/tlnet/install-tl-unx.tar.gz" \
|
|
"https://ctan.math.washington.edu/tex-archive/systems/texlive/tlnet/install-tl-unx.tar.gz" \
|
|
"https://mirrors.rit.edu/CTAN/systems/texlive/tlnet/install-tl-unx.tar.gz" \
|
|
"https://mirror.las.iastate.edu/tex-archive/systems/texlive/tlnet/install-tl-unx.tar.gz"; do \
|
|
echo "🔄 Trying mirror: $mirror" && \
|
|
if wget --timeout=30 --tries=2 -O /tmp/install-tl-unx.tar.gz "$mirror"; then \
|
|
echo "✅ Download successful from: $mirror" && \
|
|
SUCCESS=true && \
|
|
break; \
|
|
else \
|
|
echo "❌ Failed to download from: $mirror" && \
|
|
rm -f /tmp/install-tl-unx.tar.gz; \
|
|
fi; \
|
|
done && \
|
|
if [ "$SUCCESS" = false ]; then \
|
|
echo "❌ All mirrors failed, cannot proceed" && \
|
|
exit 1; \
|
|
fi && \
|
|
echo "📥 Download completed" && \
|
|
echo "📊 Downloaded file size:" && \
|
|
ls -lh /tmp/install-tl-unx.tar.gz && \
|
|
\
|
|
echo "📦 Extracting TeX Live installer..." && \
|
|
cd /tmp && tar -xzf install-tl-unx.tar.gz && \
|
|
echo "📦 Extraction completed" && \
|
|
echo "📊 Extracted files:" && \
|
|
ls -la /tmp/install-tl-* && \
|
|
echo "✅ TeX Live installer ready"
|
|
|
|
# Step 5.3: Create TeX Live profile and install base system with retry logic
|
|
RUN echo "🔧 Creating TeX Live installation profile..." && \
|
|
echo "selected_scheme scheme-medium" > /tmp/texlive.profile && \
|
|
echo "tlpdbopt_install_docfiles 0" >> /tmp/texlive.profile && \
|
|
echo "tlpdbopt_install_srcfiles 0" >> /tmp/texlive.profile && \
|
|
echo "TEXDIR /usr/local/texlive" >> /tmp/texlive.profile && \
|
|
echo "TEXMFCONFIG /usr/local/texlive/texmf-config" >> /tmp/texlive.profile && \
|
|
echo "TEXMFHOME /usr/local/texlive/texmf-home" >> /tmp/texlive.profile && \
|
|
echo "TEXMFLOCAL /usr/local/texlive/texmf-local" >> /tmp/texlive.profile && \
|
|
echo "TEXMFSYSCONFIG /usr/local/texlive/texmf-config" >> /tmp/texlive.profile && \
|
|
echo "TEXMFSYSVAR /usr/local/texlive/texmf-var" >> /tmp/texlive.profile && \
|
|
echo "TEXMFVAR /usr/local/texlive/texmf-var" >> /tmp/texlive.profile && \
|
|
echo "📄 Profile created" && \
|
|
echo "📊 Profile contents:" && \
|
|
cat /tmp/texlive.profile && \
|
|
\
|
|
echo "🔄 Installing TeX Live base system with retry logic..." && \
|
|
INSTALL_SUCCESS=false && \
|
|
for attempt in 1 2 3; do \
|
|
echo "🔄 Installation attempt $attempt/3..." && \
|
|
if /tmp/install-tl-*/install-tl --profile=/tmp/texlive.profile; then \
|
|
echo "✅ TeX Live installation successful on attempt $attempt" && \
|
|
INSTALL_SUCCESS=true && \
|
|
break; \
|
|
else \
|
|
echo "❌ Installation attempt $attempt failed" && \
|
|
if [ $attempt -lt 3 ]; then \
|
|
echo "⏳ Waiting 10 seconds before retry..." && \
|
|
sleep 10; \
|
|
fi; \
|
|
fi; \
|
|
done && \
|
|
if [ "$INSTALL_SUCCESS" = false ]; then \
|
|
echo "❌ All installation attempts failed" && \
|
|
echo "📋 Checking for partial installation..." && \
|
|
ls -la /usr/local/texlive/ || echo "No TeX Live directory found" && \
|
|
exit 1; \
|
|
fi && \
|
|
echo "📦 TeX Live base system installed"
|
|
|
|
# Step 5.4: Configure TeX Live PATH
|
|
RUN echo "🔧 Setting up TeX Live PATH..." && \
|
|
echo 'export PATH=/usr/local/texlive/bin/x86_64-linux:$PATH' >> /etc/bash.bashrc && \
|
|
export PATH=/usr/local/texlive/bin/x86_64-linux:$PATH && \
|
|
echo "🔧 TeX Live PATH configured" && \
|
|
\
|
|
echo "📍 Checking if tlmgr is available..." && \
|
|
if command -v tlmgr >/dev/null 2>&1; then \
|
|
echo "📦 tlmgr available"; \
|
|
tlmgr --version | head -1; \
|
|
else \
|
|
echo "❌ tlmgr not found or not working"; \
|
|
exit 1; \
|
|
fi && \
|
|
echo "✅ TeX Live PATH and tlmgr verified"
|
|
|
|
# Step 5.5: Install TeX Live packages with improved error handling
|
|
RUN echo "📊 Analyzing tl_packages file..." && \
|
|
collection_count=$(grep -c '^collection-' /tmp/tl_packages) && \
|
|
echo "📦 Found $collection_count TeX Live collections to install" && \
|
|
\
|
|
echo "🔄 Installing TeX Live collections with retry logic..." && \
|
|
export PATH=/usr/local/texlive/bin/x86_64-linux:$PATH && \
|
|
i=1 && \
|
|
failed_packages="" && \
|
|
while IFS= read -r collection; do \
|
|
case "$collection" in \
|
|
collection-*) \
|
|
echo "📦 [$i/$collection_count] Installing $collection..."; \
|
|
if command -v tlmgr >/dev/null 2>&1; then \
|
|
success=false; \
|
|
for retry in 1 2; do \
|
|
if tlmgr install "$collection"; then \
|
|
echo "✅ Successfully installed $collection"; \
|
|
success=true; \
|
|
break; \
|
|
else \
|
|
echo "❌ Attempt $retry failed for $collection"; \
|
|
if [ $retry -lt 2 ]; then \
|
|
echo "⏳ Retrying in 5 seconds..."; \
|
|
sleep 5; \
|
|
fi; \
|
|
fi; \
|
|
done; \
|
|
if [ "$success" = false ]; then \
|
|
echo "⚠️ Failed to install $collection after retries, continuing..."; \
|
|
failed_packages="$failed_packages $collection"; \
|
|
fi; \
|
|
else \
|
|
echo "⚠️ tlmgr not available, skipping $collection"; \
|
|
fi; \
|
|
i=$((i+1)); \
|
|
;; \
|
|
esac; \
|
|
done < /tmp/tl_packages && \
|
|
if [ -n "$failed_packages" ]; then \
|
|
echo "⚠️ Some packages failed to install:$failed_packages"; \
|
|
echo "📋 This may not be critical for basic functionality"; \
|
|
fi && \
|
|
echo "✅ TeX Live packages installation completed"
|
|
|
|
# Step 5.6: TeX Live cleanup and completion
|
|
RUN echo "🧹 Cleaning up TeX Live installer..." && \
|
|
rm -rf /tmp/install-tl-* /tmp/texlive.profile /tmp/install-tl-unx.tar.gz && \
|
|
echo "✅ === TEX LIVE INSTALLATION COMPLETE ===" && \
|
|
echo "📊 Free disk space after: $(df -h / | tail -1 | awk '{print $4}')" && \
|
|
echo "📊 TeX Live disk usage: $(du -sh /usr/local/texlive 2>/dev/null || echo 'N/A')"
|
|
|
|
# Verify TeX Live installation (with error handling)
|
|
RUN echo "🔄 Verifying TeX Live installation..." && \
|
|
export PATH=/usr/local/texlive/bin/x86_64-linux:$PATH && \
|
|
echo "📍 PATH: $PATH" && \
|
|
echo "📍 Checking TeX Live directory:" && \
|
|
ls -la /usr/local/texlive/ || echo "❌ TeX Live directory not found" && \
|
|
ls -la /usr/local/texlive/bin/ || echo "❌ TeX Live bin directory not found" && \
|
|
if [ -f /usr/local/texlive/bin/x86_64-linux/lualatex ]; then \
|
|
echo "✅ lualatex found"; \
|
|
/usr/local/texlive/bin/x86_64-linux/lualatex --version | head -1; \
|
|
else \
|
|
echo "❌ lualatex not found, checking for alternative locations"; \
|
|
find /usr/local/texlive -name "lualatex" -type f 2>/dev/null || echo "No lualatex found anywhere"; \
|
|
fi && \
|
|
echo "✅ TeX Live verification complete (allowing partial failures)"
|
|
|
|
# === PHASE 6: R INSTALLATION (R 4.5 from CRAN) ===
|
|
RUN echo "🚀 === STARTING R INSTALLATION ===" && \
|
|
echo "⏰ Estimated time: 1-2 minutes" && \
|
|
start_time=$(date +%s) && \
|
|
\
|
|
echo "📦 Adding CRAN repository for R 4.5..." && \
|
|
apt-get update && apt-get install -y --no-install-recommends software-properties-common dirmngr && \
|
|
wget -qO- https://cloud.r-project.org/bin/linux/ubuntu/marutter_pubkey.asc | tee -a /etc/apt/trusted.gpg.d/cran_ubuntu_key.asc > /dev/null && \
|
|
add-apt-repository "deb https://cloud.r-project.org/bin/linux/ubuntu jammy-cran40/" && \
|
|
apt-get update && \
|
|
\
|
|
echo "📦 Installing R and development packages..." && \
|
|
echo "📋 R packages:" && \
|
|
echo " - r-base" && \
|
|
echo " - r-base-dev" && \
|
|
echo " - r-recommended" && \
|
|
apt-get install -y --no-install-recommends \
|
|
r-base \
|
|
r-base-dev \
|
|
r-recommended && \
|
|
if [ $? -ne 0 ]; then \
|
|
echo "❌ Failed to install R packages"; \
|
|
exit 1; \
|
|
fi && \
|
|
echo "📦 R packages installed" && \
|
|
\
|
|
echo "🧹 Cleaning package cache..." && \
|
|
rm -rf /var/lib/apt/lists/* && \
|
|
\
|
|
echo "📊 R version: $(R --version | head -1)" && \
|
|
if [ $? -ne 0 ]; then \
|
|
echo "❌ Failed to verify R installation"; \
|
|
exit 1; \
|
|
fi && \
|
|
\
|
|
end_time=$(date +%s) && \
|
|
duration=$((end_time - start_time)) && \
|
|
echo "✅ === R INSTALLATION COMPLETE === (${duration}s)"
|
|
|
|
# === PHASE 7: R PACKAGE INSTALLATION ===
|
|
RUN echo "🚀 === STARTING R PACKAGE INSTALLATION ===" && \
|
|
echo "⏰ Estimated time: 2-3 minutes" && \
|
|
start_time=$(date +%s) && \
|
|
\
|
|
echo "Running R setup script with the following files:" && \
|
|
echo " - /tmp/install_packages.R (from book/tools/dependencies/install_packages.R)" && \
|
|
echo " - /tmp/verify_r_packages.R (from book/docker/linux/verify_r_packages.R)" && \
|
|
\
|
|
echo "📦 Installing R packages..." && \
|
|
Rscript /tmp/install_packages.R && \
|
|
if [ $? -ne 0 ]; then \
|
|
echo "❌ Failed to install R packages"; \
|
|
exit 1; \
|
|
fi && \
|
|
echo "✅ R packages installed successfully" && \
|
|
\
|
|
echo "🔍 Verifying R packages..." && \
|
|
Rscript /tmp/verify_r_packages.R && \
|
|
if [ $? -ne 0 ]; then \
|
|
echo "❌ Failed to verify R packages"; \
|
|
exit 1; \
|
|
fi && \
|
|
echo "✅ R packages verified successfully" && \
|
|
\
|
|
echo "🧹 Cleaning up R installation files..." && \
|
|
rm -f /tmp/install_packages.R /tmp/verify_r_packages.R && \
|
|
\
|
|
end_time=$(date +%s) && \
|
|
duration=$((end_time - start_time)) && \
|
|
echo "✅ === R PACKAGE INSTALLATION COMPLETE === (${duration}s)"
|
|
|
|
# === PHASE 8: PYTHON INSTALLATION ===
|
|
RUN echo "🚀 === STARTING PYTHON INSTALLATION ===" && \
|
|
echo "⏰ Estimated time: 1 minute" && \
|
|
start_time=$(date +%s) && \
|
|
\
|
|
echo "📦 Installing Python 3 and development packages..." && \
|
|
echo "📋 Python packages:" && \
|
|
echo " - python3" && \
|
|
echo " - python3-pip" && \
|
|
echo " - python3-dev" && \
|
|
apt-get update && apt-get install -y \
|
|
python3 \
|
|
python3-pip \
|
|
python3-dev && \
|
|
if [ $? -ne 0 ]; then \
|
|
echo "❌ Failed to install Python packages"; \
|
|
exit 1; \
|
|
fi && \
|
|
echo "📦 Python packages installed" && \
|
|
\
|
|
echo "🧹 Cleaning package cache..." && \
|
|
rm -rf /var/lib/apt/lists/* && \
|
|
\
|
|
echo "📊 Python version: $(python3 --version)" && \
|
|
if [ $? -ne 0 ]; then \
|
|
echo "❌ Failed to verify Python installation"; \
|
|
exit 1; \
|
|
fi && \
|
|
echo "📊 Pip version: $(pip3 --version)" && \
|
|
if [ $? -ne 0 ]; then \
|
|
echo "❌ Failed to verify pip installation"; \
|
|
exit 1; \
|
|
fi && \
|
|
\
|
|
end_time=$(date +%s) && \
|
|
duration=$((end_time - start_time)) && \
|
|
echo "✅ === PYTHON INSTALLATION COMPLETE === (${duration}s)"
|
|
|
|
# === PHASE 9: QUARTO INSTALLATION ===
|
|
RUN echo "🚀 === STARTING QUARTO INSTALLATION ===" && \
|
|
echo "⏰ Estimated time: 1 minute" && \
|
|
start_time=$(date +%s) && \
|
|
\
|
|
echo "📦 Downloading Quarto 1.9.27..." && \
|
|
wget -q https://github.com/quarto-dev/quarto-cli/releases/download/v1.9.27/quarto-1.9.27-linux-amd64.deb && \
|
|
if [ $? -ne 0 ]; then \
|
|
echo "❌ Failed to download Quarto"; \
|
|
exit 1; \
|
|
fi && \
|
|
echo "📥 Download completed" && \
|
|
echo "📊 Downloaded file size:" && \
|
|
ls -lh quarto-1.9.27-linux-amd64.deb && \
|
|
\
|
|
echo "📦 Installing Quarto..." && \
|
|
dpkg -i quarto-1.9.27-linux-amd64.deb && \
|
|
if [ $? -ne 0 ]; then \
|
|
echo "❌ Failed to install Quarto"; \
|
|
exit 1; \
|
|
fi && \
|
|
echo "📦 Quarto installed" && \
|
|
\
|
|
echo "🧹 Cleaning up installer..." && \
|
|
rm quarto-1.9.27-linux-amd64.deb && \
|
|
echo "🧹 Installer cleaned up" && \
|
|
\
|
|
end_time=$(date +%s) && \
|
|
duration=$((end_time - start_time)) && \
|
|
echo "✅ === QUARTO INSTALLATION COMPLETE === (${duration}s)"
|
|
|
|
# Create R library directory
|
|
RUN echo "🚀 === STARTING R LIBRARY SETUP ===" && \
|
|
echo "📁 Creating R library directory..." && \
|
|
mkdir -p $R_LIBS_USER && \
|
|
echo "✅ R library directory created: $R_LIBS_USER" && \
|
|
echo "✅ R library setup complete"
|
|
|
|
|
|
|
|
# === PHASE 10: PYTHON PACKAGES ===
|
|
RUN echo "🚀 === STARTING PYTHON PACKAGE INSTALLATION ===" && \
|
|
echo "⏰ Estimated time: 1-2 minutes" && \
|
|
start_time=$(date +%s) && \
|
|
\
|
|
echo "🔄 Upgrading pip..." && \
|
|
pip3 install --upgrade pip && \
|
|
\
|
|
echo "📊 Analyzing requirements.txt..." && \
|
|
package_count=$(grep -v '^#' /tmp/requirements.txt | grep -v '^$' | wc -l) && \
|
|
echo "📦 Found $package_count Python packages to install" && \
|
|
\
|
|
echo "🔄 Installing Python packages with space optimization..." && \
|
|
pip3 install --no-cache-dir -r /tmp/requirements.txt && \
|
|
if [ $? -ne 0 ]; then \
|
|
echo "❌ Failed to install Python packages"; \
|
|
exit 1; \
|
|
fi && \
|
|
\
|
|
echo "🧹 Cleaning Python installation caches..." && \
|
|
pip3 cache purge && \
|
|
find /usr -name "*.pyc" -delete && \
|
|
find /usr -name "__pycache__" -type d -exec rm -rf {} + || true && \
|
|
\
|
|
echo "📊 Installed Python packages:" && \
|
|
pip3 list | head -10 && \
|
|
echo "📊 Total packages: $(pip3 list | wc -l)" && \
|
|
\
|
|
end_time=$(date +%s) && \
|
|
duration=$((end_time - start_time)) && \
|
|
echo "✅ === PYTHON PACKAGES COMPLETE === (${duration}s)"
|
|
|
|
# === PHASE 9: R PACKAGES ===
|
|
# Step 9.1: Set up R environment and install remotes
|
|
RUN echo "🚀 === STEP 9.1: SETTING UP R ENVIRONMENT ===" && \
|
|
R --slave -e " \
|
|
options(repos = c(CRAN = 'https://cran.rstudio.com')); \
|
|
cat('🔄 Setting up R environment...\n'); \
|
|
cat(paste('R library path:', Sys.getenv('R_LIBS_USER'), '\n')); \
|
|
lib_path <- Sys.getenv('R_LIBS_USER'); \
|
|
dir.create(lib_path, showWarnings = FALSE, recursive = TRUE); \
|
|
.libPaths(lib_path); \
|
|
cat('📦 Installing remotes package...\n'); \
|
|
install.packages('remotes'); \
|
|
cat('✅ R environment setup complete\n'); \
|
|
" && \
|
|
if [ $? -ne 0 ]; then \
|
|
echo "❌ Failed to set up R environment"; \
|
|
exit 1; \
|
|
fi && \
|
|
echo "✅ R environment setup successful"
|
|
|
|
# Step 9.2: Install R packages from install_packages.R or fallback
|
|
RUN echo "🚀 === STEP 9.2: INSTALLING R PACKAGES ===" && \
|
|
R --slave -e " \
|
|
options(repos = c(CRAN = 'https://cran.rstudio.com')); \
|
|
lib_path <- Sys.getenv('R_LIBS_USER'); \
|
|
.libPaths(lib_path); \
|
|
if (file.exists('/tmp/install_packages.R')) { \
|
|
cat('📦 Installing packages from tools/dependencies/install_packages.R...\n'); \
|
|
source('/tmp/install_packages.R'); \
|
|
} else { \
|
|
cat('⚠️ No tools/dependencies/install_packages.R found, installing common packages\n'); \
|
|
pkgs <- c('rmarkdown', 'knitr', 'tidyverse', 'ggplot2', 'bookdown'); \
|
|
cat(paste('📦 Installing packages:', paste(pkgs, collapse=', '), '\n')); \
|
|
install.packages(pkgs); \
|
|
}; \
|
|
cat('✅ R package installation complete\n'); \
|
|
" && \
|
|
if [ $? -ne 0 ]; then \
|
|
echo "❌ Failed to install R packages"; \
|
|
exit 1; \
|
|
fi && \
|
|
echo "✅ R packages installed successfully"
|
|
|
|
# Step 9.3: Verify R package installation
|
|
RUN echo "🚀 === STEP 9.3: VERIFYING R PACKAGES ===" && \
|
|
R --slave -e " \
|
|
lib_path <- Sys.getenv('R_LIBS_USER'); \
|
|
.libPaths(lib_path); \
|
|
cat('📊 Installed packages:\n'); \
|
|
ip <- installed.packages()[, 'Package']; \
|
|
print(head(ip, 10)); \
|
|
cat(paste('Total packages installed:', nrow(ip), '\n')); \
|
|
" && \
|
|
if [ $? -ne 0 ]; then \
|
|
echo "❌ Failed to verify R packages"; \
|
|
exit 1; \
|
|
fi && \
|
|
echo "✅ R package verification successful"
|
|
|
|
# === PHASE 11: R PACKAGE VERIFICATION ===
|
|
# Note: R package verification is now done immediately after R package installation (Phase 6.5)
|
|
|
|
# === PHASE 12: COMPREHENSIVE CLEANUP ===
|
|
RUN echo "🚀 === STARTING COMPREHENSIVE CLEANUP ===" && \
|
|
echo "📊 Disk space before cleanup: $(df -h / | tail -1 | awk '{print $4}')" && \
|
|
start_time=$(date +%s) && \
|
|
\
|
|
echo "🧹 Removing temporary files..." && \
|
|
rm -rf /tmp/* && \
|
|
rm -rf /var/tmp/* && \
|
|
\
|
|
echo "🧹 Cleaning package caches..." && \
|
|
apt-get clean && \
|
|
rm -rf /var/lib/apt/lists/* && \
|
|
rm -rf /var/cache/apt/* && \
|
|
\
|
|
echo "🧹 Cleaning Python caches..." && \
|
|
find /usr -name "*.pyc" -delete && \
|
|
find /usr -name "__pycache__" -type d -exec rm -rf {} + || true && \
|
|
pip3 cache purge || true && \
|
|
\
|
|
echo "🧹 Cleaning R temporary files..." && \
|
|
rm -rf /tmp/Rtmp* || true && \
|
|
rm -rf /var/lib/R/site-library/*/help || true && \
|
|
\
|
|
echo "🧹 Cleaning TeX Live caches and docs..." && \
|
|
rm -rf /usr/local/texlive/*/texmf-var/luatex-cache/* || true && \
|
|
rm -rf /usr/local/texlive/*/texmf-var/web2c/* || true && \
|
|
\
|
|
echo "🧹 Removing unnecessary system files..." && \
|
|
rm -rf /usr/share/doc/* && \
|
|
rm -rf /usr/share/man/* && \
|
|
rm -rf /usr/share/info/* && \
|
|
rm -rf /var/log/* && \
|
|
\
|
|
end_time=$(date +%s) && \
|
|
duration=$((end_time - start_time)) && \
|
|
echo "✅ === COMPREHENSIVE CLEANUP COMPLETE === (${duration}s)" && \
|
|
echo "📊 Final disk space: $(df -h / | tail -1 | awk '{print $4}')"
|
|
|
|
# Set working directory
|
|
WORKDIR /workspace
|
|
|
|
# === PHASE 13: FINAL VERIFICATION ===
|
|
RUN echo "🚀 === STARTING FINAL VERIFICATION ===" && \
|
|
export PATH=/usr/local/texlive/bin/x86_64-linux:$PATH && \
|
|
echo "📊 Checking Quarto..." && \
|
|
if command -v quarto >/dev/null 2>&1; then \
|
|
quarto --version && \
|
|
echo "✅ Quarto verified"; \
|
|
else \
|
|
echo "❌ Quarto not found in PATH"; \
|
|
echo "📍 Checking for quarto in common locations:"; \
|
|
find /usr -name "quarto" -type f 2>/dev/null || echo "No quarto found"; \
|
|
exit 1; \
|
|
fi && \
|
|
echo "📊 Checking Python..." && \
|
|
if command -v python3 >/dev/null 2>&1; then \
|
|
python3 --version && \
|
|
echo "✅ Python verified"; \
|
|
else \
|
|
echo "❌ Python3 not found"; \
|
|
exit 1; \
|
|
fi && \
|
|
echo "📊 Checking R..." && \
|
|
if command -v R >/dev/null 2>&1; then \
|
|
R --version && \
|
|
echo "✅ R verified"; \
|
|
else \
|
|
echo "❌ R not found"; \
|
|
exit 1; \
|
|
fi && \
|
|
echo "📊 Checking LaTeX..." && \
|
|
if command -v lualatex >/dev/null 2>&1; then \
|
|
lualatex --version && \
|
|
echo "✅ LaTeX verified"; \
|
|
else \
|
|
echo "❌ lualatex not found"; \
|
|
echo "📍 Checking for lualatex in TeX Live:"; \
|
|
find /usr/local/texlive -name "lualatex" -type f 2>/dev/null || echo "No lualatex found"; \
|
|
exit 1; \
|
|
fi && \
|
|
echo "✅ Final verification complete"
|
|
|
|
# Health check
|
|
RUN export PATH=/usr/local/texlive/bin/x86_64-linux:$PATH && \
|
|
echo "✅ Container build completed successfully" && \
|
|
echo "📊 Quarto version: $(quarto --version)" && \
|
|
echo "📊 Python version: $(python3 --version)" && \
|
|
echo "📊 R version: $(R --version | head -1)" && \
|
|
echo "📊 TeX Live: $(lualatex --version | head -1)"
|