Changelog
View Source0.9.2
- Bump required
doc_redlinesdependency to~> 0.6.0. - Refresh package metadata after the
doc_redlines 0.6.0release.
0.9.0
- Bumped version for the public release with legacy
.docextraction support viadoc_redlines. - Added a synthetic
.docfixture for local validation (non-proprietary).
0.8.0
- Added optional legacy
.docextraction support viadoc_redlines(Redlines.extract/2now supports:docand.docinference). - Added
Redlines.DOCadapter to normalize.docredline output intoRedlines.Change. - Added a synthetic
.docfixture for local validation.
0.7.1
- Note:
0.7.0was not published; these changes shipped as0.7.1. - DOCX cleaning now also accepts moved text (
<w:moveFrom>,<w:moveTo>) and drops revision history/markers (e.g.*PrChange,*RangeStart/*RangeEnd). - Added warning-returning helpers:
Redlines.clean_docx_with_warnings/2,Redlines.clean_docx_binary_with_warnings/2. - Added DOCX cleaning helpers to accept track changes (remove
<w:del>and unwrap<w:ins>) and return a cleaned DOCX:Redlines.clean_docx/2,Redlines.clean_docx_binary/2.
0.6.1
- Bump
pdf_redlinesto~> 0.7.1(fixes precompiled NIF checksum mismatch)
0.6.0
- Breaking: Removed
has_redlines?/1. Callers should useextract/2and checkresult.changes == []instead. The old heuristic pre-check doubled the PDF rendering cost and risked false negatives on non-standard redline colors. - Bumped
pdf_redlinesto~> 0.7.0. - Added performance section to README (under 700 ms on 35 MB+ scanned documents).
0.5.1
- Bump
pdf_redlinesdependency to~> 0.6.3
0.5.0
pdf_redlinesis now a required dependency (precompiled NIF, no toolchain needed).- Removed runtime detection of
PDFRedlinesmodule availability. - Expanded README with badges, full API docs, and usage examples.
0.1.0
- Initial release.
- DOCX track-changes extraction from
word/document.xml. - Optional PDF extraction via the
pdf_redlinespackage. - Unified
Redlines.Changeshape andRedlines.format_for_llm/1.