Files
stack/tests/rex/comments/test_combine.py
kert b2ad400764 feat(comments): DOCX/text extract + combined.md writer (refs #253)
- extract_attachment now dispatches PDF/DOCX/TXT/HTML; DOCX via
  python-docx, text-like via plain read with cheap HTML strip.
- combine.py writes one combined.md per comment dir with YAML
  frontmatter (per-attachment status + chars) and per-section body.
- extract_comment is idempotent (existing combined.md is a no-op
  unless force=True).
- parse_combined() splits frontmatter ↔ body for downstream consumers.

Walker + DuckDB view land in the next batch.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-23 15:42:32 -04:00

2.9 KiB