Spaces:
Sleeping
Sleeping
Commit History
-
feat: intelligent search-based navigation for trending/popular content
- Add _infer_navigation_paths() for generic path inference based on intent - Add _build_search_navigation_url() for search-based fallback - Prefer search URLs for top/trending/popular queries (work without auth) - Rewrite _extract_markdown_link_rows() with smarter content detection: - Skip image links (![Image]) in markdown - Extract metrics (views, likes, comments) from nearby lines - Filter boilerplate labels and URL tokens aggressively - Rank results by keyword relevance and engagement metrics - Expand boilerplate filtering in _fallback_extraction_code() - Fix alias matching in registry.py (boundary-aware matching) - Add regression tests for alias matching This enables generic extraction for JS-heavy sites like YouTube by using search URLs that return content without requiring authentication.0735175
-
test: add full site template matrix coverage
- add registry tests that validate every template domain maps correctly - add API tests for /api/sites list/get/match endpoints - add site template matrix report under docs/test - validate 56 template catalog with comprehensive assertions Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>7f50e50