feat: Phase 4 KB Integration - pdf_kb_ingest tool

[krisoye]

Summary

Implements Phase 4 of Epic #21 - KB Integration for the Document Analysis MCP Server.

Features

• pdf_kb_ingest tool - One-shot PDF ingestion for knowledge bank integration
• Smart text chunking respecting paragraph/sentence boundaries
• Page number tracking for source references
• Document classification (LLM with heuristic fallback)
• Caching support with parameter-specific keys
• Comprehensive input validation and error handling

Changes

• New src/document_analysis_mcp/tools/kb_ingest.py (1,200+ lines)
• Updated server.py with pdf_kb_ingest tool registration
• Updated tools/init.py exports
• 43 new tests in tests/test_kb_ingest.py

Test Results

• 266 tests passing (43 new)
• All linting checks pass
• Code review approved with zero P0/P1 issues

Test plan

• Unit tests for chunking logic
• Unit tests for classification (LLM and heuristic)
• Integration tests for end-to-end flows
• Edge case tests (empty PDFs, large docs, invalid input)
• CI pipeline validation

Resolves Epic #21 Phase 4
Refs: krisoye/project-tracker#96

🤖 Generated with Claude Code