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

[pdurrant]

Originally Posted by mergen3107

@pdurrant, thank you!
Can you please tell me:
1) Can KindleUnpack extract only TOC? For the purpose of my project I don't need all raw files;
2) How can I get location of TOC entries in a mobi8 file? So far, I can see that it only provides ordered TOC entry reference numbers. As a temporary solution to find percentage of a chapter is to calculated respective cumulative file sizes, but that's very rough.

The python code of KindleUnpack should tell you how to obtain all the information you need, if it's actually in the file.

[mergen3107]

OK, thank you!

[xeshxzh]

Hi pdurrant,

Thank you for maintaining the KindleUnpack library. Really appreciate it.

I am wondering if we can convert a .mobi dictionary file into a txt file. Each line is one word and its definition separated by a tab.

The reason I'm asking is:

I use KindleMate to convert words in Vocabulary Builder to Anki cards. And I want to use my own dictionary (Longman) when creating the Anki cards. Currently KindleMate only takes dictionary in txt format.

That's why I need to convert my mobi dictionary file into a txt.

Any help would be greatly appreciated.

Thank you,
Hannah

[DiapDealer]

Kindleunpack won't help you convert it to text. It will help you to convert your mobi to html. From there, you'd need to use something else to get it into a text-only format.

Just note that there isn't 100% support for kindle dictionaries. Some things work. Some things don't.

[xeshxzh]

Thank you DiapDealer!

In my case, it doesn't seem to work. I can't even convert the mobi dictionary to html. No useful error message either.

Do you know if there is any other library that can parse mobi files? It doesn't have to be python.

Thank you again,
Hannah

Originally Posted by DiapDealer

Kindleunpack won't help you convert it to text. It will help you to convert your mobi to html. From there, you'd need to use something else to get it into a text-only format.

Just note that there isn't 100% support for kindle dictionaries. Some things work. Some things don't.

[DiapDealer]

It's not the mobi format in general that's the problem. It's the extra bits that dictionaries use. There's always been issues with infections and such. I'm not the expert there, so I can't really say what the trouble is with mobi-dictionaries.

You might try calibre. I don't know if it will do any better on dictionaries.

But more than likely, anything that claims to convert/unpack mobi files/dictionaries is probably based on calibre or KindleUnpack code (and thus the same limitations). And a lot of the mobi stuff in KindleUnpack was ported from the older Perl-based Mobiperl project. Somebody blasts though every now and again with high hopes of creating a C/C++ or C# version, but I'm not sure if anyone has a fully-functional alternative yet.

[xeshxzh]

Yea, I can understand why. I tried Calibre and it gave me "Memory Error".

But still, thank you for taking the time to explain to me. Appreciate it.

With gratitude,
Hannah

Originally Posted by DiapDealer

It's not the mobi format in general that's the problem. It's the extra bits that dictionaries use. There's always been issues with infections and such. I'm not the expert there, so I can't really say what the trouble is with mobi-dictionaries.

You might try calibre. I don't know if it will do any better on dictionaries.

But more than likely, anything that claims to convert/unpack mobi files/dictionaries is probably based on calibre or KindleUnpack code (and thus the same limitations). And a lot of the mobi stuff in KindleUnpack was ported from the older Perl-based Mobiperl project. Somebody blasts though every now and again with high hopes of creating a C/C++ or C# version, but I'm not sure if anyone has a fully-functional alternative yet.

[jhowell]

Originally Posted by xeshxzh

Yea, I can understand why. I tried Calibre and it gave me "Memory Error".

That error is common if you try to convert a large complex book, such as a dictionary, using the 32-bit version of calibre. Switching to the 64-bit version usually allows conversion to complete, however it can take hours or even days to do so.

[patrik]

Is there anything I can do to get kindle samples to work for extracting epub?

[pdurrant]

Originally Posted by patrik

Is there anything I can do to get kindle samples to work for extracting epub?

Kindle samples are not complete files, so I expect that the unpack fall over when it gets to the bit of the file that's missing.

[KevinH]

If jhowell is okay with it, I would like to pull his kfx unpack code and epub-like conversion code in kindleunpack so users have one stop shopping. There is/was also an official ION github code library someplace that I once saw. Perhaps pulling that lib in (as long as the licenses are compatible), would make sense. We could just add interface routines to his code so we can easily keep up with his changes.

Doug, Paul, and jhowell, what do you think?

[NiLuJe]

I see Amazon has a few ION-related stuff on GH, f.g., this in Python: https://github.com/amzn/ion-python

Was that what you were thinking of?

[KevinH]

Yes that looks like it. It has an Apache license file. If jhowell think it might be useful, we could pull it in, but if his code would not benefit from it, that is fine too.

[DiapDealer]

Originally Posted by KevinH

Doug, Paul, and jhowell, what do you think?

I'm OK with the idea, but to what extent? Is there any part of the current KindleUnpack codebase that will be useful in the venture? I admit that I'm woefully out of the loop RE KFx, but I was under the impression that it was a completely different beast altogether. Much like Topaz was (which we didn't incorporate). Is there even a mobi header/exth to contend with?

What would the unpacked structure look like? I guess I'm just trying to determine if we will be doing more converting, than actual unpacking/inspection.

[jhowell]

Originally Posted by KevinH

If jhowell is okay with it, I would like to pull his kfx unpack code and epub-like conversion code in kindleunpack so users have one stop shopping...

I am fine with anyone reusing my code. You didn’t need to ask, but thanks for doing so.

What I have to share at this point is contained in the latest release of the KFX Input plugin for calibre. It is not in GitHub at the moment and probably won’t be any time soon.

The calibre plugin and CLI are at the top level, with more generic code for handling KFX contained in the kfxlib module.

Originally Posted by KevinH

If jhowell think it might be useful, we could pull it in, but if his code would not benefit from it, that is fine too.

My code contains my own implementation of Amazon Ion. It would probably be more trouble than it is worth to integrate Amazon’s.