BREAKING CHANGE
- parsing_metadata keys pptx_pictures_found and
pptx_pictures_captioned (added in 0.193.0) are renamed to
pictures_found and pictures_captioned.
Feat
- processors: read .xlsx natively via openpyxl (ADR-038)
- processors: read .docx natively via python-docx (ADR-038)
Fix
- xlsx: order xl/media naturally so the caption cap takes the first pictures
- docx: treat direct numbering as a list item; pin same-paragraph picture order
- config: warn when OFFICE_CAPTION_* and PPTX_CAPTION_* are both set
Refactor
- ooxml: share pptx picture captioning with the other OOXML readers