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

[pdurrant]

Originally Posted by odamizu

Thank you! This is exciting

Was it necessary for me to compile and run the AppleScript after copying over the files? Or would copying the new files into the AppleScript have been sufficient?

And thank you, pdurrant, as always, for your wonderful AppleScripts. I love them so. I have tried to learn to write AppleScripts, but I don't get far before hitting a wall in my inadequate brain

If the Applescript application already runs on your machine, just replacing the scripts should be sufficient.

[odamizu]

Originally Posted by pdurrant

If the Applescript application already runs on your machine, just replacing the scripts should be sufficient.

Thank you! I'm ridiculously happy that I succeeded in updating this myself

[kakaka]

Hello, noob here. Not sure if this is the right place to ask.
But in mobi, image embeds has an attribute recindex which allows you to easily get the href of the image file from content.opf.
How is that done in the azw format?
When unpacking those files, image embeds look something like this (epub converted to azw3 using calibre).

<p class="c5" aid="N"><a href="kindle:pos:fid:002A:off:0000000000" title="linked image" aid="O"><img alt="Decorative graphic" title="Decorative graphic" src="kindle:embed:003V?mime=image/jpeg" class="calibre4"/></a></p>

I don't understand any of this.

[DiapDealer]

Not sure what the question is. Is KindleUnpack failing to unpack a book?

[kakaka]

Originally Posted by DiapDealer

Not sure what the question is. Is KindleUnpack failing to unpack a book?

No sorry, this probably isn't the correct place to ask this question. This is more about the kindle format in general. But oh well. My question is that I don't understand how image embeds work in azw. Basically the example above is an image embed in the rawml of an unpacked azw file. I want to know which image, in the image OEBPS directory of the unpacked azw file, that the image embed refers to using that example.

[DiapDealer]

Not really sure to be honest. We've focused primarily on Kindlebooks created with the official tools from Amazon. We try not to break anything when working with calibre-generated kindle books, but there may be something going on there we haven't accounted for.

But there should be no need to examine the rawml to find which embeds go with which images. KindleUnpack uses an algorithm to convert the binary offsets to the image data into urls. So the unpacked html (or the generated epub) should have the links to the correct image in the OEBPS/images directory. The rawml binary offsets will naturally be rather meaningless to the human eye/brain.

[kakaka]

Ok cool, Could you point me to that algorithm?

[KevinH]

It differs if embedded image, svg image, image url in css files, and of course by mobi6 vs mobi8 (KF8).

Check out the comments in the appropriate routines in mobi_html.py for some info.

For KF8 you will need to understand how aids are base 32 numbers to offsets in the rawml file, what a skeleton vs fragment tables, are etc.

You may just be better off looking at the generated xhtml files in the epub it creates to find the mapping.

Originally Posted by kakaka

Ok cool, Could you point me to that algorithm?

[WEricBr]

In looking at the ReadMe for KindleUnpack, it shows that Python 2.6 or 2.7 is required. Can the 3.x version be used, or is 2.6 or 2.7 the version that has to be used?

Thanks,

Eric

[JSWolf]

Originally Posted by WEricBr

In looking at the ReadMe for KindleUnpack, it shows that Python 2.6 or 2.7 is required. Can the 3.x version be used, or is 2.6 or 2.7 the version that has to be used?

Thanks,

Eric

If you use Calibre, you can use the KindleUnpack plugin.

[WEricBr]

I am wanting to get just the headers using the "KindleUnpack -r" option. I tried unpack in Calibre, but did not see a way to dump the headers.

I have used the Calibre-Debug to dump the headers, but wanted to see if there were an differences in the 2 since the header definitions differ in the source code.

Thanks,

Eric

[JSWolf]

Can someone please download the 0.83 version and replace the older version in the first post? Thanks.

https://github.com/kevinhendricks/KindleUnpack

[JSWolf]

Originally Posted by WEricBr

I am wanting to get just the headers using the "KindleUnpack -r" option. I tried unpack in Calibre, but did not see a way to dump the headers.

I have used the Calibre-Debug to dump the headers, but wanted to see if there were an differences in the 2 since the header definitions differ in the source code.

Thanks,

Eric

Problem solved. The version in the first post of the KindleUnpack thread is an older version and the one at Github is a newer version.

The KindleUnpack program requires Python 2.7.X or Python 3.4 or later to function properly.

https://github.com/kevinhendricks/KindleUnpack

[WEricBr]

Thanks, I had a back version that stated 2.6 or 2.7 and I wanted to make sure that I installed the correct version since I have read that 2.x and 3.x are somewhat incompatible.

Thanks again,

Eric

[JSWolf]

Originally Posted by WEricBr

Thanks, I had a back version that stated 2.6 or 2.7 and I wanted to make sure that I installed the correct version since I have read that 2.x and 3.x are somewhat incompatible.

Thanks again,

Eric

Python code can sometimes be made compatible with 2.x and 3.x as is the case here.