On Thu, 2024-06-27 at 10:02 +0200, Richard wrote:
> You could try if Googles ML model "magika" can do a better job
> (available via pypi). Otherwise, what exactly does "file" or better
> "file -i" say? Worst case, you could open the files in a hex editor
> and google the first few bits. Chances are the format uses "magic
> bits", so the first few bits in hex are identical for all files of
> that format.
>
> Am Do., 27. Juni 2024 um 06:33 Uhr schrieb Van Snyder
> <van.sny...@sbcglobal.net>:
> > I downloaded everything with the same base name as I sent -- a file
> > and a directory. LibreOffice can't read any of it. Calibre can't
> > read any of it, either in the download or in the mounted Kindle.
> > "file" has no idea what any of the files are.
> >
> >
> >
Thanks to Richard for the suggestion to try magika:
# find ./ -type f -exec ~/.local/bin/magika {} \;
Whence-Energy-2_YI5JSRVWOVJ5MO6EWC57GJKTBZLXXI23.sdr/Whence-Energy-
2_YI5JSRVWOVJ5MO6EWC57GJKTBZLXXI23bbd37f35b4b251f45c1bf42879a62d70.yjf:
Unknown binary data (unknown) [Low-confidence model best-guess: JPEG
image data (image), score=28]
Whence-Energy-
2_YI5JSRVWOVJ5MO6EWC57GJKTBZLXXI23.sdr/data/.pagination.cache/887fcb7c:
Unknown binary data (unknown) [Low-confidence model best-guess: PE
executable (executable), score=71]
Whence-Energy-
2_YI5JSRVWOVJ5MO6EWC57GJKTBZLXXI23.sdr/YI5JSRVWOVJ5MO6EWC57GJKTBZLXXI23
.mf: Generic text document (text) [Low-confidence model best-guess:
JSON document (code), score=59]
Whence-Energy-2_YI5JSRVWOVJ5MO6EWC57GJKTBZLXXI23.sdr/Whence-Energy-
2_YI5JSRVWOVJ5MO6EWC57GJKTBZLXXI23bbd37f35b4b251f45c1bf42879a62d70.yjr:
Unknown binary data (unknown) [Low-confidence model best-guess: Intel
80386 COFF (executable), score=76]
Whence-Energy-
2_YI5JSRVWOVJ5MO6EWC57GJKTBZLXXI23.sdr/AssetDownloadMetadata.meta: JSON
document (code)
Whence-Energy-2_YI5JSRVWOVJ5MO6EWC57GJKTBZLXXI23.kfx: Unknown binary
data (unknown) [Low-confidence model best-guess: BMP image data
(image), score=89]
So except for one file, it's still a mystery what e-mailing a PDF to a
Kindle reader with "convert" (without quotes) in the subject actually
produces. The Kindle reader can read it, but so far nothing else I've
tried can. In the end, I want to edit it and re-publish it on Amazon.
okular thinks the .kfx file is a mobipocket, but can't open it. It
doesn't know what the .yjf or .yjr files are.
So, back to trying to find a competent PDF -to- ePub or PDF -to- mobi
converter (I haven't yet tried texmate to create a mobi or ePub from
the LaTeX).
