[Conversion Plugin] KFX Input

[yoshi]

Right-click on "KFX Input.zip" in the Calibre installation folder, select "Restore previous versions," and if you're lucky, you can revert to an older version.

[jhowell]

Originally Posted by Overkatefinated

I had zero issue using the previous version but now my books will no longer convert, they are stuck in the "KFX, MD" format.

If you see a format of MD in calibre after importing a book downloaded using Kindle for PC that indicates that you selected the wrong file for import. The correct main KFX file has an extension of ".azw". A supplemental metadata file found in the same folder has as extension of ".azw.md". Importing that will not work since calibre sees ".md" as an indication that the book is in markdown format instead of KFX.

This can be confusing because the Windows File Explorer does not show file extensions by default so a file with the extension ".azw.md" will appear there as ".azw". There is a setting in File Explorer that can be checked to show extensions. I recommend that be done to avoid this sort of problem in the future.

[jhowell]

Originally Posted by mallomar

Hi, I just imported a half-dozen books from my Kindle. Five of the six are fine, but when I go to convert the sixth to EPUB I get this error:

calibre, version 7.26.0
ERROR: KFX conversion failed: <p><b>Cannot convert Mrs. Caliban: A Novel</b><br><br>Exception('Book is incomplete. All of the KFX container files that make up the book must be combined into a KFX-ZIP file for successful conversion. (Missing containers CR!NMZERT4J5137Z6CAZW343TT7HKHW)')

I thought it might be an issue of the book not being fully downloaded but it shows as fully downloaded and displays without issue on my Kindle.

That error indicates there is in fact a file missing despite the book opening on the Kindle. KFX format is split into multiple files and Kindles will allow a book to be opened for reading before all of the files have downloaded in order to speed up the process.

Occasionally the supplemental files do not all download correctly. When that happens you need to remove the download from the device and re-download the book again.

[jhowell]

Originally Posted by Overkatefinated

Well I have the same issue as PikeUK but with ALL books from the US. Is there a way to rollback to 2.19??

That error was fixed in version 2.20 of KFX Input so it very unlikely to be the same problem. Please post a copy of the conversion error log and I will review it.

[mallomar]

Originally Posted by jhowell

That error indicates there is in fact a file missing despite the book opening on the Kindle. KFX format is split into multiple files and Kindles will allow a book to be opened for reading before all of the files have downloaded in order to speed up the process.

Occasionally the supplemental files do not all download correctly. When that happens you need to remove the download from the device and re-download the book again.

Thanks, I removed the download and redownloaded the book and it worked. Bizarre that it showed as fully downloaded when it wasn't.

There are some books in which the book description on Amazon shows they reference pages from the printed book (what Amazon calls real pages), but there's no page map or page list I can see in the EPUB I converted from the KFX. Is it likely these also didn't download incorrectly, despite no errors, or is it something else?

[jhowell]

Originally Posted by mallomar

There are some books in which the book description on Amazon shows they reference pages from the printed book (what Amazon calls real pages), but there's no page map or page list I can see in the EPUB I converted from the KFX. Is it likely these also didn't download incorrectly, despite no errors, or is it something else?

If you did not receive an error then all of the files that make up the book were found.

Page numbers will be retained if present in KFX format. They might appear in different places in the EPUB based on how you did the conversion and the target EPUB version. Check for a pageList in toc.ncx if EPUB 2 or nav.xhtml if EPUB 3. If you cannot find it let me know the ASINs of some books that are missing page numbers and I will try to take a look. Posting a conversion log would also help since that will show the number of page entries found if any.

[akirby83]

I've written a powershell script to convert 500+ kfx books in my Calibre library using the calibre-debug command line... the problem is it doesn't add the new files into my library, unlike the From KFX toolbar command in the GUI which does. Is there a command line argument I'm missing that will accomplish this?

[jhowell]

Originally Posted by akirby83

I've written a powershell script to convert 500+ kfx books in my Calibre library using the calibre-debug command line... the problem is it doesn't add the new files into my library, unlike the From KFX toolbar command in the GUI which does. Is there a command line argument I'm missing that will accomplish this?

After converting you can use the calibredb add command to add the result to your calibre library.

[valyent]

Hey @jhowell, thanks for making KFX Input!

I ran into a bug that preventing me from converting a few books. I was able to get around it by adding a `try/except` block around `name = self.strip_font_name(name)` in yj_to_epub_properties.py on line 2039.

Thanks again.

[jhowell]

Originally Posted by valyent

Hey @jhowell, thanks for making KFX Input!

I ran into a bug that preventing me from converting a few books. I was able to get around it by adding a `try/except` block around `name = self.strip_font_name(name)` in yj_to_epub_properties.py on line 2039.

Thanks again.

Thanks for the bug report! Would you mind sharing the ASIN of some books that exhibit this problem? I would like to determine the root cause of the failure and see if there might be a better way to handle it.

[akirby83]

Originally Posted by jhowell

After converting you can use the calibredb add command to add the result to your calibre library.

I eventually got it to work but I had to experiment. I have modified a lot of the titles in Calibre so Automerge was creating new entries instead of rolling them into the modified KFX version. The GUI version of your plugin seems to do this merge automatically but I had to experiment here. My solution was to import the book folder instead of the book file, treating everything inside as 1 book and ignoring kfx files. Not ignoring the KFX does seem to work while throwing errors that imply that it hasn't worked. This command worked for me, in case anyone is in my same situation:

calibredb add --automerge ignore -1 -- ignore *.kfx <folder that the book is in>

[Overkatefinated]

Originally Posted by jhowell

If you see a format of MD in calibre after importing a book downloaded using Kindle for PC that indicates that you selected the wrong file for import. The correct main KFX file has an extension of ".azw". A supplemental metadata file found in the same folder has as extension of ".azw.md". Importing that will not work since calibre sees ".md" as an indication that the book is in markdown format instead of KFX.

This can be confusing because the Windows File Explorer does not show file extensions by default so a file with the extension ".azw.md" will appear there as ".azw". There is a setting in File Explorer that can be checked to show extensions. I recommend that be done to avoid this sort of problem in the future.

So using the "Add Books" option in Calibre is the wrong way cause it has ALWAYS given me the KFX, MD after import. Someone on Reddit suggested a different way to import so I am gonna try that and see if it helps.

[DNSB]

Originally Posted by Overkatefinated

So using the "Add Books" option in Calibre is the wrong way cause it has ALWAYS given me the KFX, MD after import. Someone on Reddit suggested a different way to import so I am gonna try that and see if it helps.

If you enable see file extensions in File Explorer for Windows, you will see the real extensions. Alternatively you could import the file with the ASIN as the first part file name (B0D6RSYDSSN_EBOK.azw is a lot shorter than the ones that start CR!<whatever.azw.md)