KindleUnpack (MobiUnpack): Extracts text, images and metadata from Kindle/Mobi files

[pdurrant]

Originally Posted by TK423

Beginner question here but what is the actual difference between using KindleUnpack versus simply using Calibre to convert a AZW3 file to EPUB?

KindleUnpack extracts what is already in the AZW3 with minimal changes. Doing a convert in calibre extracts the contents and reformats it according to the settings in calibre.

[DiapDealer]

Originally Posted by TK423

Beginner question here but what is the actual difference between using KindleUnpack versus simply using Calibre to convert a AZW3 file to EPUB?

It depends on your goals, too, I guess. If you're simply looking for a conversion from a Kindlebook to an EPUB, then there's probably very little difference for you.

But KindleUnpack (originally MobiUnpack) began life as an inspection/extraction tool for Kindle books. Markup was extracted with as much attention to detail as possible. There are no assumptions/changes made when unpacking a Kindle book with KindleUnpack.

Massaging that unpacked Kindle book markup into an EPUB was a relatively new feature for KindleUnpack. And to be quite frank: the resulting EPUB can be quite the non-compliant mess unless the Kindle book in question was actually created from an epub in the first place (granted: most of them are made from EPUBS any more).

[DNSB]

Originally Posted by TK423

Beginner question here but what is the actual difference between using KindleUnpack versus simply using Calibre to convert a AZW3 file to EPUB?

For me, the main difference is that the calibre conversion messes with the css files and calibre can split a single xhtml file into multiple files where some nudnick has used multiple header tags in a single file. When the input file to Amazon was an ePub, KindleUnpack is a lot closer to the original structure.

If you are not planning on actually mucking with the ePub in an editor (Sigil, calibre ebook-editor, whatever), there is not much real difference. If you are planning on editing the ePub, KindleUnpack is the way to go.

I won't get into the philosophical differences I have with Kovid Goyal over attempting to create output files that are compliant to the ePub specs. There is quite a bit of cruft in the ePub specs but many publishers will insist that your ePub pass epubcheck before it will be accepted for publishing.

[Never_Sm1le]

Is there a way to batch unpacking mobi instead of manually selecting each one?

[DiapDealer]

Originally Posted by Never_Sm1le

Is there a way to batch unpacking mobi instead of manually selecting each one?

Not really. Not easily anyway. The KindleUnpack.pyw file is merely a gui wrapper around the lib/kindleunpack.py file. The latter can be scripted from the command line easily enough, but you'd need to automatically provide a different output directory for each kindlebook, or else the next book would overwrite (more like "be lumped in with") the contents of the previous. If you need to batch unpack to epub, the calibre plugin version of Kindle unpack might better suit your needs.

[JSWolf]

Originally Posted by DNSB

For me, the main difference is that the calibre conversion messes with the css files and calibre can split a single xhtml file into multiple files where some nudnick has used multiple header tags in a single file. When the input file to Amazon was an ePub, KindleUnpack is a lot closer to the original structure.

If you are not planning on actually mucking with the ePub in an editor (Sigil, calibre ebook-editor, whatever), there is not much real difference. If you are planning on editing the ePub, KindleUnpack is the way to go.

I won't get into the philosophical differences I have with Kovid Goyal over attempting to create output files that are compliant to the ePub specs. There is quite a bit of cruft in the ePub specs but many publishers will insist that your ePub pass epubcheck before it will be accepted for publishing.

There are cases where it's better to do a calibre conversion. I've seen a number of Kindle eBooks where the entire book is one file. I've also seen cases where the eBook has no CSS and this was KF8. Basically, if the KindleUnpacked ePub is too much of a mess, it's better to convert then the shift.

[DiapDealer]

As already mentioned: when the source for the Kindle book was NOT an epub, the KindleUnpack generated epub could end up being a non-compliant mess. Both DNSB and myself specified "when the kindlebook was/was not created from an epub" in our replies.

KindleUnpack's epub is generally only a "mess" when the Kindlebook is not a KF8 created from an epub. When the KindleUnpack plugin for Sigil encounters an old style mobi (non-kf8) kindlebook, the epub generated will nearly always be a slam bam conversion from mobi markup to xhtml/css that will need a lot of work.