Skip to content

v1.0.0

@teaguesterling teaguesterling tagged this 16 Jul 05:44
Add HTML type system with full libxml2 HTML parser integration:
- HTML type with proper DuckDB type registration and casting
- parse_html() scalar function for reading/parsing HTML files
- read_html_objects() table function for batch HTML processing

Implement HTML extraction functions following markdown extension patterns:
- html_extract_text() with XPath support for text extraction
- html_extract_links() returning structured link data (href, text, title)
- html_extract_images() returning image metadata (src, alt, title, dimensions)
- html_extract_tables() table function for structured table extraction

Key features:
- Full XML/HTML function compatibility (xml_extract_* work on HTML)
- Robust HTML parsing with malformed HTML recovery
- Smart empty file handling (returns <html></html>)
- Proper RAII memory management with XMLCharPtr smart pointers
- Clean API design avoiding function naming conflicts

All tests passing: 22/22 test cases, 417 assertions

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <noreply@anthropic.com>
Assets 2
Loading