v1.0.4
Patch release. Two follow-ups deferred from v1.0.3 (#17):
- Pipeline (T7):
dedupe_toc_tablenow normalizes leader-dot padding (Purpose..........) from TOC cells before matching against body H2/H3 titles. Docling-rendered TOC tables now get properly stripped under aggressive/maximum profiles. - Audit script:
scripts/audit-outputs.pyleading-toc-tablecheck no longer false-flags documents whose only GFM table sits at the end with no H2 in the body.
See CHANGELOG.md for full notes.
Install:
pip install --upgrade 'any2md[high-fidelity]==1.0.4'