feat(postgis): add PostGIS documentation support

[cbc3929]

Summary

This PR adds comprehensive PostGIS documentation support to pg-aiguide, enabling AI coding assistants to provide better guidance for spatial database operations.

New Features

• PostGIS Documentation Scraper (ingest/postgis_docs.py)
  • Dedicated scraper for PostGIS manual (DocBook HTML format)
  • Supports both file and database storage modes
  • Header-based markdown chunking with token counting
• Search APIs
  • semantic_search_postgis_docs - Vector similarity search for PostGIS documentation
  • keyword_search_postgis_docs - BM25 keyword search for PostGIS documentation
• Database Migration
  • postgis_pages and postgis_chunks tables
  • HNSW index for fast vector similarity search

Enhanced Embedding Configuration

Added support for custom embedding providers (beyond OpenAI):

• OPENAI_BASE_URL - Custom OpenAI-compatible API endpoint (e.g., Ollama, SiliconFlow)
• EMBEDDING_MODEL - Custom embedding model name
• EMBEDDING_DIMENSIONS - Configurable vector dimensions This allows users to use alternative embedding services while maintaining compatibility with the existing database schema.

Testing

• ✅ TypeScript build passes
• ✅ Python syntax validation passes
• ✅ Database migration tested
• ✅ Scraper tested with file mode (5 pages)
• ✅ Scraper tested with database mode (3 pages, 43 chunks)
• ✅ Semantic search verified with vector similarity queries

Usage

# Scrape PostGIS documentation
cd ingest
uv run python postgis_docs.py --version 3.5 --storage-type database
# With custom embedding provider
export OPENAI_BASE_URL=https://api.siliconflow.cn/v1
export EMBEDDING_MODEL=Qwen/Qwen3-Embedding-8B
uv run python postgis_docs.py --version 3.5 --storage-type database
Checklist
- Code follows project conventions
- All comments in English
- Documentation updated (README.md, .env.sample)
- Database migration included
- Tests performed locally