Hi,

I ran a shell script to verify or falsify the existence of the listed packages 
in Debian.

for pkg in "${packages[@]}"; do
    
    if apt-cache show "$pkg" "python3-$pkg" "lib$pkg" >/dev/null 2>&1; then
        echo "$pkg: FOUND"
    elif [ $(apt-cache search "$pkg" | wc -l) -gt 0 ]; then
        echo "$pkg: PARTIAL MATCH (Check manually)"
    else
        echo "$pkg: NOT FOUND"
    fi
done

html2text: FOUND
libportal: PARTIAL MATCH (Check manually)
libspelling: PARTIAL MATCH (Check manually)
markitdown: NOT FOUND
matplotlib: FOUND
odfpy: PARTIAL MATCH (Check manually)
openai: FOUND
opencv: FOUND
pydbus: FOUND
pyicu: NOT FOUND
pillow: FOUND
requests: FOUND
vte: PARTIAL MATCH (Check manually)
youtube-transcript-api: NOT FOUND

I rechecked the partial matches manually and came to following conclusion:

libportal FOUND
libspelling FOUND
pyodf FOUND (python-odf)
vte FOUND

False negatives (packages marked as NOT FOUND but are FOUND):
pyicu FOUND

This results in following missing packages:

1. markitdown 
2. youtube-transcript-api

Both packages seem to be DFSG compliant at first look.

Youtube transcript API depends exclusively on packages already existing in 
Debian. [1]

Markitdown includes according to Microsoft trademarked material [2] which has 
to be removed as needed.

Following dependencies are required by markitdown but are missing in Debian 
[3]:
* markdownify
* magika~=0.6.1



References:
[1] https://github.com/jdepoix/youtube-transcript-api/blob/master/
pyproject.toml#L51-L54
[2] https://github.com/microsoft/markitdown/tree/main/packages/
markitdown#trademarks
[3] https://github.com/microsoft/markitdown/blob/main/packages/markitdown/
pyproject.toml#L26-L34


Regards,
-- 
Arian Ott
[email protected]

Attachment: signature.asc
Description: This is a digitally signed message part.

Reply via email to