Add comprehensive cross-referencing system for ML Systems book

Features:
- Semantic similarity-based cross-reference generation using sentence transformers
- 3-stage workflow: train embeddings, generate suggestions, optional LLM enhancement
- Simplified title-only approach (no LLM) for zero-cost operation
- Optional GPT-4 enhancement with pedagogical analysis and smart content extraction
- Quarto Lua filter for automatic cross-reference injection
- Comprehensive testing with real book content
- Multiple workflow options from simple (/bin/zsh) to enhanced (~/bin/zsh.28)

Core Components:
- scripts/cross_referencing/cross_referencing.py: Main tool with 3-stage workflow
- lua/inject_xrefs.lua: Quarto filter for cross-reference injection
- Extensive documentation and testing utilities
- Cost-optimized approach with 64% token reduction strategies

Ready for production use with flexible enhancement options.

.gitignore

@@ -48,3 +48,4 @@ swap_quarto.bat
 
 Machine-Learning-Systems.log
 
+__pycache__/