[Conversion Plugin] KFX Input

[theducks]

metadata.opf is not a book.
That is the metadata backup file for Calibre.
If there is no book format, that is an 'empty book' (no formats) entry in calibre

[jhowell]

Originally Posted by GreenEyedLady

ERROR: Loading book failed: Failed to open the book at C:\Users\dlare\Calibre Library\Steven L. Peck\A Short Stay in Hell (8946)\metadata.opf. Click "Show details" for more info.

As theducks pointed out it appears that you are trying to import the wrong file, one that is already inside of your calibre library.

See the section "Importing KFX books into calibre for conversion" in the first post in this thread for information on how to locate books downloaded using Kindle for PC.

[GreenEyedLady]

Thanks, I was trying to find something that might be helpful.

For some reason I can see the book cover and name, but file is still unconverted KFX. Must have done something wrong. I've tried a few different books.

Deb

[GreenEyedLady]

Duh
I figurged it out. After I get KFX I needed to convert to EPUb. I can read now.
Thanks I do appreciate your responses.

[zenqoo]

I cant convert kindle book to EPUB,
when i tried to convert, the error appears ( i have imported the book to calibre)

calibre, version 7.11.0
ERROR: KFX conversion failed: <p><b>Cannot convert D-genesis 01</b><br><br>Exception("ruby_content b4GZ 3 has extra data: ($146: [($155: 6004, $159: $271, $584: '', $601: $283, $157: s4HF, $175: e4HD)])\nexternal_resource has extra data: (e4HD: ($161: $285, $162: 'image/jpg', $165: 'resource/rsrc6ZD', $422: 128, $175: e4HD, $423: 128))\nraw_media has extra data: (resource/rsrc6ZD: *** 3852 byte BLOB bd77ec268c95ee850a1e9f9d92f6fd38dc7753a8 ***)\nkfx styles has extra data: (s4HF: ($56: ($307: 1.0, $306: $308), $57: ($307: 1.0, $306: $308), $546: $377, $44: $350, $173: s4HF))")</p>

--

calibre, version 7.11.0
ERROR: KFX conversion failed: <p><b>Cannot convert D-genesis 05</b><br><br>Exception("ruby_content b558 1 has extra data: ($146: [($155: 6314, $584: '', $601: $283, $157: s55D, $159: $271, $175: e55B)])\nexternal_resource has extra data: (e55B: ($161: $285, $162: 'image/jpg', $165: 'resource/rsrc69W', $422: 128, $423: 128, $175: e55B))\nraw_media has extra data: (resource/rsrc69W: *** 4319 byte BLOB 4579fd02cf1679d8c6701776d31969ad309eef41 ***)\nkfx styles has extra data: (s55D: ($56: ($307: 1.0, $306: $308), $57: ($307: 1.0, $306: $308), $546: $377, $44: $350, $173: s55D))")</p>

[jhowell]

Originally Posted by zenqoo

I cant convert kindle book to EPUB,
when i tried to convert, the error appears ( i have imported the book to calibre)

Thank you for reporting this problem. There are variations on ruby content that are not handled properly by the current version of this plugin. This will be fixed in the next plugin release, expected within a few days.

[jhowell]

Version 2.12.0 - 28 May 2024

Improve handling of ruby content. Fixes ("ruby_content xxxx has extra data: ...")

Ignore extraneous BookManifest.kfx files produced by the new Kindle for Mac app.

[Jaws]

...that may well be related to the source files, but it's so unpredictable that I'm hesitant to assume anything. This appears in some KFX books (and, as far as I've been able to determine, only KFX books) converted to epub.

Sometimes word/phrase formatting — such as a word appearing in italics, "<i>Ibid.</i>" — is hardcoded:

<i>Ibid.</i>

and sometimes the conversion results in a style that is the format variation only, included in both the stylesheet and appearing in the text as:

<span style="class_s47">Ibid.</span>

This can result in some bizarre display glitches, but I'm more puzzled that they're unpredictable. The difference can appear in volume 2, but not volumes 1 or 3, of the same series from the same publisher released as e-books in the same month. I've seen this across multiple publishers. The only predictability is that they're consistent across the KFX (that is, it would be every appearance of italics in a given KFX source — or none of them).

Please confirm whether this is something internal to KFX's conversion routines, something that's slightly broader in Calibre's conversion routines, or something that's in the source files (such as different people doing the conversions at the publishers doing something different in InDesign or Word resulting in slightly different output code — which can/does happen). If you can, that is... and this is far from urgent. Just bizarre.

[tomsem]

As you might expect, the Scribe update to support multiple templates per notebook presents a new use case for KFX Input. It only converts pages using initial template. Example attached.

Software versions: KFX Input 2.12.0, calibre 7.12, macOS-14.5-x86_64-i386-64bit
KFX Input plugin help is available at https://www.mobileread.com/forums/showthread.php?t=291290

Processing bc106476-b4a1-9d8c-8806-e1102deac839
Processing container: bc106476-b4a1-9d8c-8806-e1102deac839/nbk
Removed 1 KDF SQLite file fingerprint(s)
Found local_action_fragments table
Found local_delta_fragments table
Features: db.delta_update-1, db.schema-1, max_id-825
Metadata: book_id=7cJe926iQ1Oq4QxclSjdVA0, reading_orders=2
Converting bc106476-b4a1-9d8c-8806-e1102deac839 to EPUB
Converting book to EPUB 3
Book symbol format is base64
ERROR: section ctTcMACReRu_dR0ElBZZ4ow1 has extra data: {nmdl.template_id: $349}
ERROR: section ctTcMACReRu_dR0ElBZZ4owE has extra data: {nmdl.template_id: ztTcMACReRu_dR0ElBZZ4owK}
ERROR: section ctTcMACReRu_dR0ElBZZ4ow13 has extra data: {nmdl.template_id: ztTcMACReRu_dR0ElBZZ4ow18}
ERROR: section ctTcMACReRu_dR0ElBZZ4ow1P has extra data: {nmdl.template_id: ztTcMACReRu_dR0ElBZZ4ow18}
ERROR: section ctTcMACReRu_dR0ElBZZ4ow1T has extra data: {nmdl.template_id: $349}
ERROR: section ctTcMACReRu_dR0ElBZZ4ow26 has extra data: {nmdl.template_id: $349}
Notebook template: medium_rule_1860x2480
Notebook template: dotted_1860x2480
Format is fixed layout notebook
Converted book saved to bc106476-b4a1-9d8c-8806-e1102deac839.epub

[jhowell]

Originally Posted by tomsem

As you might expect, the Scribe update to support multiple templates per notebook presents a new use case for KFX Input. It only converts pages using initial template. Example attached.

The next release of KFX Input will handle Scribe notebooks that use multiple templates. Thanks for the sample.

[jhowell]

Originally Posted by Jaws

This appears in some KFX books (and, as far as I've been able to determine, only KFX books) converted to epub.

Sometimes word/phrase formatting — such as a word appearing in italics, "<i>Ibid.</i>" — is hardcoded:

<i>Ibid.</i>

and sometimes the conversion results in a style that is the format variation only, included in both the stylesheet and appearing in the text as:

<span style="class_s47">Ibid.</span>

I do not understand how you could be seeing "<i>Ibid.</i>" in an EPUB generated by this plugin.

KFX format has no direct equivalent to the <i> HTML element. When books are converted to KFX any <i> is replaced by the KFX equivalent to a span with CSS styling of "font-style: italic". The KFX Input plugin retains this formatting when converting to EPUB. It never produces <i> elements.

Originally Posted by Jaws

This can result in some bizarre display glitches, but I'm more puzzled that they're unpredictable. The difference can appear in volume 2, but not volumes 1 or 3, of the same series from the same publisher released as e-books in the same month. I've seen this across multiple publishers. The only predictability is that they're consistent across the KFX (that is, it would be every appearance of italics in a given KFX source — or none of them).

What do you mean by bizarre display glitches?

Can you provide the titles (and ASINs) of books that demonstrate this?

Originally Posted by Jaws

Please confirm whether this is something internal to KFX's conversion routines, something that's slightly broader in Calibre's conversion routines, or something that's in the source files (such as different people doing the conversions at the publishers doing something different in InDesign or Word resulting in slightly different output code — which can/does happen). If you can, that is... and this is far from urgent. Just bizarre.

I do not understand what is happening here. It is possible that calibre's conversion process is making changes to the EPUB produced by KFX Input. For example there is a calibre conversion option to italicize common words and patterns. That could be coming into play if you have it enabled under heuristic processing.

One thing you could try to do is use the "From KFX" function built into this plugin to bypass the effects of calibre conversion. See the first post in this thread for more information.

[Michael Andersen]

Hi

New older member. Just purchased a Kobo. Very happy. Trying to convert Kindle eBooks. Downloaded Calibre Version 7.12, transferred Kindle books in, tried to convert, no luck, DRM protected. Have downloaded DeDRM, and tried to install in Calibre. Error message received is;

calibre, version 7.12.0
ERROR: Unhandled exception: <b>InvalidPlugin</b>:The plugin in 'E:\\Users\\Michael\\Downloads\\DeDRM_tools_10.0.9 .zip' is invalid. It does not contain a top-level __init__.py file

calibre 7.12 embedded-python: True
Windows-10-10.0.19045-SP0 Windows ('64bit', 'WindowsPE')
('Windows', '10', '10.0.19045')
Python 3.11.5
Windows: ('10', '10.0.19045', 'SP0', 'Multiprocessor Free')
Interface language: None
EXE path: C:\Program Files\Calibre2\calibre.exe
Successfully initialized third party plugins: Gather KFX-ZIP (from KFX Input) (2, 12, 0) && Package KFX (from KFX Input) (2, 12, 0) && Barnes & Noble (1, 5, 5) && Extract ISBN (1, 6, 3) && Fantastic Fiction (1, 7, 0) && Fantastic Fiction Adults (1, 3, 1) && FictionDB (1, 4, 2) && Generate Cover (2, 3, 4) && Goodreads (1, 8, 1) && ISFDB3 (1, 4, 0) && Job Spy (1, 0, 239) && K2pdfopt Plugin (0, 5, 4) && KFX metadata reader (from KFX Input) (2, 12, 0) && From KFX (2, 12, 0) && KFX Input (2, 12, 0) && Set KFX metadata (from KFX Output) (2, 8, 0) && KFX Output (2, 8, 0) && Kobo Books (1, 9, 2) && Kobo Metadata (1, 5, 0) && Kobo Utilities (2, 16, 13) && Perrypedia (1, 9, 0) && Search The Internet (1, 11, 2) && Skoob Books (1, 5, 5) && Standard Ebooks (1, 0, 0)
Traceback (most recent call last):
File "calibre\gui2\preferences\plugins.py", line 322, in add_plugin
File "calibre\customize\ui.py", line 525, in add_plugin
File "calibre\customize\ui.py", line 76, in load_plugin
File "calibre\customize\zipplugin.py", line 297, in load
File "calibre\customize\zipplugin.py", line 392, in _locate_code
calibre.customize.InvalidPlugin: The plugin in 'E:\\Users\\Michael\\Downloads\\DeDRM_tools_10.0.9 .zip' is invalid. It does not contain a top-level __init__.py file

Can you help please?

[theducks]

Last line: the DRM plugin is old.
Search MR for the link to NoDRM

[jhowell]

Originally Posted by Michael Andersen

ERROR: Unhandled exception: <b>InvalidPlugin</b>:The plugin in 'E:\\Users\\Michael\\Downloads\\DeDRM_tools_10.0.9 .zip' is invalid. It does not contain a top-level __init__.py file

That is a common error resulting from not following the instructions for installing the DRM removal plugin. Detailed assistance with DRM removal is not allowed here on MobileRead. Either Google the error message or ask for help at the site where you obtained the plugin.

[Jaws]

Just a couple responses for clarity, I think JHowell's response tells me "it's the publisher."

The bizarre display issues are line breaks around the span specifying italics, but only around that style definition (no other style definitions, including both bold-only and small-caps only, have that problem). This has popped up in several different readers for both Windows and phones. Replacing that style entirely with HTML italics fixes it.

My suspicion is that the hardcoded italics may arise from particular publisher workflows, and those workflows would tend to be volume/book by volume/book in converting back catalogs (and prior editions) to KFX. Particularly for "updated editions" and "updated texts" for which the publisher was working from either PDFs-intended-for-the-bindery or raw PostScript files -- which in my recollection are all of the places I've seen the inconsistency -- there's probably something weird in the conversion process that is embedding both a style and a hardcode. I've encountered this in publisher epub editions -- a style for font-style: italic wrapped around hardcoded italics. Proving yet again that data-format transformation is as much an art as translating human languages, especially when the speaker may not be all that articulate in the original language.

At least it's not imposing strong and em, so I suppose I should be happy.