🌿 fiche-extract v1.1.0 · MIT

Extract text and embedded media from an XLSX technical sheet — one pass, source file never modified.

What it does

Text + images together: openpyxl and soffice --convert-to csv only give you cell text — real-world spec sheets often carry critical info in embedded photos/diagrams. fiche-extract gets both.
One CSV per sheet via LibreOffice headless; every xl/media/* entry extracted and inventoried (bytes, type).
Safe by construction: the source file is opened read-only; everything lands in a separate output folder.
Explicit JSON report: sheets, line counts, media inventory, warnings (macros detected, unexpected formats, missing soffice…).
Honest exit codes: 0 content extracted · 1 failure/no content · 2 bad input (missing, not .xlsx, zero bytes).

pip install git+https://mandrilly.com/git/fiche-extract.git

Installs a fiche-extract command (zero Python dependencies — uses your system soffice):

fiche-extract spec.xlsx            # → ./fiche_extract_spec/{csv/, media/, report.json}
fiche-extract spec.xlsx -o out_dir

python3 fiche_extract.py spec.xlsx            # → /tmp/fiche_extract_spec/{csv/, media/, report.json}
python3 fiche_extract.py spec.xlsx -o out_dir

git clone https://mandrilly.com/git/fiche-extract.git

Requires Python 3.8+ and LibreOffice (soffice). Test fixtures included.