[GUI Plugin] KindleUnpack - The Plugin

[DiapDealer]

Originally Posted by Nitro

It's happening with some polish books, nothing special about them.

You're trying to split this mobi? Or do you get that error during a normal unpack/extract?

If you're trying to split, what version of Kindlegen/KindlePreview were the dual-format mobis created with?

[Nitro]

I'm trying to split original azw3 file from bookshop, to get separate mobi and kf8 files.
Not sure how to check Kindlegen/KindlePreview version.

[pdurrant]

Originally Posted by Nitro

I'm trying to split original azw3 file from bookshop, to get separate mobi and kf8 files.
Not sure how to check Kindlegen/KindlePreview version.

An azw3 downloaded from Amazon doesn't have a mobi portion. It's already a separate KF8 file.

[Nitro]

Originally Posted by pdurrant

An azw3 downloaded from Amazon doesn't have a mobi portion. It's already a separate KF8 file.

It's not from amazon. It's from local polish shop.
One thing. We don't have protection on files here, just a watermark.

[Nitro]

Hmm... Ok, my bad.
@pdurrant gave me a good hint.
I re-checked my files, the original ones were in mobi format and they did contain azw3 and another mobi file. I added them to calibre and a months later I was trying to split azw3 file this time and it gave me an error. I just saw possibility to do it so I tried
Sorry for confusion and help

[DiapDealer]

No problem. Although ideally, the plugin's menu shouldn't even be giving you the option to split a file that isn't "splittable". Before each menu is built, each kindlebook format's headers are briefly examined to determine which actions are appropriate for that particular type of Kindlebook.

If you can provide a brief, non-copyright-violating example of a kindlebook that the plugin mistakenly indicates can be split, I'll take a look.

[Nitro]

PM sent. Again, thanks for your time

[DiapDealer]

Have you been able to split azw3 files from this particular bookstore with KindleUnpack before, or is this something you just recently decided to try?

[Nitro]

azw3 file was created by calibre/kindleunpack when I splitted mobi file from this bookstore.
They do not provide azw3, just joint mobi with mobi and azw3 inside.

[DiapDealer]

Originally Posted by Nitro

azw3 file was created by calibre/kindleunpack when I splitted mobi file from this bookstore.
They do not provide azw3, just joint mobi with mobi and azw3 inside.

So you're saying the azw3 was already split from a dual-format mobi (and kindle unpack STILL thinks it's a splittable file)?

That's interesting. There must be something different about the way those particular dual-format books are being generated. Normally all dual-header info and boundary markers are removed from an azw3 that's created when KindleUnpack splits a dual-format mobi. The one I have still indicates (for all intents and purposes) that it's still a dual format mobi.

I'll have to do some more investigating.

[Nitro]

Yes, azw3 file was split from dual-format mobi.

I have files from at least 2 different book stores, which behaves the same.

[DiapDealer]

There's definitely something different about the dual-format mobis from those stores ... but I'm not sure what. The resulting azw3 files from splitting those dual formats are structurally different than ones I've previously split from kindlegen created ones. Could simply be a newer version of kindlegen/kindle previewer (I admit I haven't tried to keep up) .

Either way, I think I can work around the issue by slightly beefing up my initial "quick" dual-format test that creates the menu actions.

NOTE: this issue is plugin-related only. No errors result when checking the "Split" box on the standalone KindleUnpack tool (there is simply no separate azw3/mobi created).

[DiapDealer]

Released a new version (0.81.4) that I hope fixes AZW3s from being falsely reported as being "splittable" in the plugin's menus.

[DiapDealer]

Released a new version (0.81.5) that utilizes the latest upstream changes to the core KindleUnpack files (addresses urls to SVG images in CSS not unpacking properly).

[democrite]

Hello,

I've been using this for a long time without issues. Got a azw3 recently that errors:

Processing ncx / toc
Traceback (most recent call last):
  File "kindleunpack.py", line 1016, in <module>
    sys.exit(main())
  File "kindleunpack.py", line 1004, in main
    unpackBook(infile, outdir, apnxfile, epubver, use_hd)
  File "kindleunpack.py", line 919, in unpackBook
    process_all_mobi_headers(files, apnxfile, sect, mhlst, K8Boundary, False, epubver, use_hd)
  File "kindleunpack.py", line 836, in process_all_mobi_headers
    processMobi8(mh, metadata, sect, files, rscnames, pagemapproc, k8resc, obfuscate_data, apnxfile, epubver)
  File "kindleunpack.py", line 522, in processMobi8
    ncx_data = ncx.parseNCX()
  File "./mobi_ncx.py", line 45, in parseNCX
    outtbl, ctoc_text = self.mi.getIndexData(self.ncxidx, "NCX")
  File "./mobi_index.py", line 38, in getIndexData
    ctocdict = self.readCTOC(cdata)
  File "./mobi_index.py", line 131, in readCTOC
    pos, ilen = getVariableWidthValue(txtdata, offset)
  File "./mobi_index.py", line 157, in getVariableWidthValue
    if ord(v) & 0x80:
TypeError: ord() expected a character, but string of length 0 found

I didn't really try to figure out the code though adding if len(v) is 0, finished = True at least made it complete though I'm not sure if there's anything missing from the output.