|
|
3327137536
|
🚀 v2.0.5: Fix page range parsing across all PDF tools
Major architectural improvements and bug fixes in the v2.0.x series:
## v2.0.5 - Page Range Parsing (Current Release)
- Fix page range parsing bug affecting 6 mixins (e.g., "93-95" or "11-30")
- Create shared parse_pages_parameter() utility function
- Support mixed formats: "1,3-5,7,10-15"
- Update: pdf_utilities, content_analysis, image_processing, misc_tools, table_extraction, text_extraction
## v2.0.4 - Chunk Hint Fix
- Fix next_chunk_hint to show correct page ranges
- Dynamic calculation based on actual pages being extracted
- Example: "30-50" now correctly shows "40-49" for next chunk
## v2.0.3 - Initial Range Support
- Add page range support to text extraction ("11-30")
- Fix _parse_pages_parameter to handle ranges with Python's range()
- Convert 1-based user input to 0-based internal indexing
## v2.0.2 - Lazy Import Fix
- Fix ModuleNotFoundError for reportlab on startup
- Implement lazy imports for optional dependencies
- Graceful degradation with helpful error messages
## v2.0.1 - Dependency Restructuring
- Move reportlab to optional [forms] extra
- Document installation: uvx --with mcp-pdf[forms] mcp-pdf
## v2.0.0 - Official FastMCP Pattern Migration
- Migrate to official fastmcp.contrib.mcp_mixin pattern
- Create 12 specialized mixins with 42 tools total
- Architecture: mixins_official/ using MCPMixin base class
- Backwards compatibility: server_legacy.py preserved
Technical Improvements:
- Centralized utility functions (DRY principle)
- Consistent behavior across all PDF tools
- Better error messages with actionable instructions
- Library-specific adapters for table extraction
Files Changed:
- New: src/mcp_pdf/mixins_official/utils.py (shared utilities)
- Updated: 6 mixins with improved page parsing
- Version: pyproject.toml, server.py → 2.0.5
PyPI: https://pypi.org/project/mcp-pdf/2.0.5/
|
2025-11-03 17:12:37 -07:00 |
|
