[PDF → ePUB] Alternative to Calibre?

[Shohreh]

Is there a way to check which version of poppler Calibre uses?

At this point, the latest is 0.87 from March 28, 2020 : https://poppler.freedesktop.org/

It says that it uses "PDF Input (1.0.0) by Kovid Goyal and John Schember" to convert PDF to HTML.

[kovidgoyal]

look at sources.json in the calibre source code

[Shohreh]

Thank you.

    {
        "name": "poppler",
        "unix": {
            "filename": "poppler-0.76.1.tar.xz",
            "hash": "sha256:e62cd4053f851eecc4a6aa2b81d739fbab1d67b3c60b31ebac7049483cc617b1",
            "urls": ["https://poppler.freedesktop.org/{filename}"]
        }
    },

    {
        "name": "podofo",
        "unix": {
            "filename": "podofo-0.9.6.tar.gz",
            "hash": "md5:46336fc4c4ce4be814bb5fbb4d918334",
            "urls": ["https://downloads.sourceforge.net/podofo/{filename}"]
        }
    },

https://github.com/kovidgoyal/calibr...y/sources.json

[Shohreh]

I had an idea: Considering PDF should display plain text just fine on an e-reader, and only some parts/pages won't because of more complex layout… an easier alternative to trying to use Caliber to convert a PDF into EPUB through poppler, is to use LibreOffice to edit just the pages/parts that don't work, and rebuild a PDF.

1. Save each page as a single PDF:

qpdf --progress --split-pages infile.pdf %d.pdf

2. In LibreOffice Draw, open and edit each problematic page to replace all nasty parts (remove/rewrite insets and multi-column text, replace tables with just screenshots that fit your e-reader, eg. 768x1024 pixels). While you're at it, use a regex to remove the headers/footers and page numbers

3. Merge all the pages back into a single PDF:

qpdf --empty --pages *.pdf -- out.pdf

[Shohreh]

For others' benefit: The easiest solution I found is to not bother trying to convert a PDF to EPUB, even with Calibre's courageous attempt to clean up the HTML generated by Poppler's pdftohtml.

Since probably all e-readers sold these days can read PDF somehow, first try copying the original, and see if it looks good enough, with or without playing with settings on the reader.

If it doesn't look good enough, an easy thing to try, is to run it through k2pdfopt, which will massage the PDF specifically for use on an e-reader.

Using the following command and disabling "Reflow text" on my reader makes for a pretty good output, even though it'll never be as good as a native EPUB:

#Fit width, portrait, 0.2 inch margin
k2pdfopt -mode fw -ls- -om 0.2 input.pdf