[Conversion Output] KFX Output

[jhowell]

Originally Posted by HowlerX

So even if I specifically put "widows:0" and "orphans:0" in the CSS it will be ignored?

Yes. The 5.12.2 Kindle firmware prevents widows/orphans with no way to enable them.

I expect that the widows and orphans properties in CSS will be obeyed properly in the next Kindle firmware release. The Kindle for iOS, Android, and Fire apps all handle them as expected in their latest releases.

Also, as theducks wrote, the correct value should be "1" to allow widows/orphans. In my testing using "0" also works the same for Amazon/KFX creation, but I would not rely on that.

[user105]

I'm getting epub to kfx:
calibre, version 4.3.0
ERROR: KFX conversion failed: <p><b>Cannot convert Das dunkle Herz des Waldes</b><br><br><b>Conversion error:</b> Kindle Previewer error: Error:E00813: Unknown failure occurred during processing</p>

epub to mobi to kfx works


any tipps?

[jhowell]

Originally Posted by user105

I'm getting epub to kfx:
calibre, version 4.3.0
ERROR: KFX conversion failed: <p><b>Cannot convert Das dunkle Herz des Waldes</b><br><br><b>Conversion error:</b> Kindle Previewer error: Error:E00813: Unknown failure occurred during processing</p>

epub to mobi to kfx works


any tipps?

Since converting to MOBI first prevents the problem it is likely due to some formatting that the Kindle Previewer does not handle properly. Much formatting is removed in that process since MOBI is limited compared with EPUB.

Since the error message is so unhelpful I suggest either skipping KFX for that book or using the version that works via MOBI.

You might be able to track down the cause of the problem by trial and error. Selectively deleting parts of the book using the Calibre editor to try to narrow down the source of the error. It may not be worth the effort involved.

[user105]

Originally Posted by jhowell

You might be able to track down the cause of the problem by trial and error. Selectively deleting parts of the book using the Calibre editor to try to narrow down the source of the error. It may not be worth the effort involved.

would that be helpfull to you for improving the output-plug in, or is this a pure previewer problem?

thank you for the effort!

[jhowell]

Originally Posted by user105

would that be helpfull to you for improving the output-plug in, or is this a pure previewer problem?

The error is produced by the Kindle Previewer. Knowing what triggers it might be useful to someone else who encounters the same problem, but it is unlikely to lead to any changes in this plugin.

[VovoKrall]

I couldn't convert to KFX because it was missing a file - a gif, which doesn't make any sense for a book - but I could convert to azw3. Do you know why?

calibre, version 4.5.0
ERRO: KFX conversion failed: <p><b>Cannot convert How to Fail at Almost Everything and Still Win Big</b><br><br><b>Conversion error:</b> Kindle Previewer error: Error:E00001: Cannot find the file name C:\Users\adria\AppData\Local\Temp\calibre_nw0p2q\k c_hqh\eqlf2lcxklyjin3uni4w\0000\conv_resources\con v_tmp\temp\extractedEpub\OEBPS\styles\images\00004 .gif</p>

[jhowell]

Originally Posted by VovoKrall

I couldn't convert to KFX because it was missing a file - a gif, which doesn't make any sense for a book - but I could convert to azw3. Do you know why?

Conversion to KFX is designed by Amazon to fail if the source file contains incorrect or missing content. Conversion to KF8 (azw3) will succeed for just about any book regardless of errors. That's just the way it is.

[jhowell]

Version 1.35.0 - 07 Dec 2019

Support additional KFX features. (Fixes "Unexpected Ion symbols used: $46, $784")

Prevent incorrect date from being used if the publication date of a book to be converted is undefined.

[phossler]

I'm using calibredb to extract (copy?) the KFX books to a new single folder:

export --progress --all --dont-save-cover --dont-write-opf --format kfx --single-dir --template "{authors}-{series}{series_index:0>2s|[|]-}{title}" --to-dir "D:\eBooks\KFX"


It seems like there's a lot more happening than just the copy and it seems like that if it's embedding the metadata, it takes a lot longer than for the same for an ePub.

Is that just a 'side effect' of the plugin or something else?

[jhowell]

Originally Posted by phossler

It seems like there's a lot more happening than just the copy and it seems like that if it's embedding the metadata, it takes a lot longer than for the same for an ePub.

Is that just a 'side effect' of the plugin or something else?

Updating metadata in KFX files is fairly slow. Add the --dont-update-metadata option to speed things up. From the manual:

Normally, calibre will update the metadata in the saved files from what is in the calibre library. Makes saving to disk slower. Specifying this switch will turn this behavior off.

[phossler]

Thanks, but unfortunately I think I do need the metadata updated

I was just wondering since it seems that the KFX part of the calibredb job takes longer than the epub part

[Silverbacknet]

I can't find a repo or any way to make a pull request, so I'll just post here. This is a small fix for when Kindle Previewer is installed in a different folder.

--- calibre\plugins\KFX Output_old\kfxlib\generate_kpf.py
+++ calibre\plugins\KFX Output\kfxlib\generate_kpf.py
@@ -464,7 +464,13 @@
         }

     def __init__(self):
-        self.program_path = os.path.join(windows_user_dir(local_appdata=True), "Amazon", "Kindle Previewer 3")
+        from _winreg import *
+        default_path = os.path.join(windows_user_dir(local_appdata=True), "Amazon", "Kindle Previewer 3")
+        try:
+            reg_key = OpenKey(ConnectRegistry(None, HKEY_CURRENT_USER),r'Software\Amazon\Kindle Previewer 3')
+            self.program_path = QueryValueEx(reg_key,None)[0]
+        except (OSError, WindowsError) as e:
+            self.program_path = default_path
         KindlePreviewer.__init__(self)

     def prepare_java_app_args(self):

Moderator Notice
I suspect this is best placed in the KFX Output plugin thread. I only copied the post here in case I'm wrong.

[jhowell]

I suspect this is best placed in the KFX Output plugin thread. I only copied the post here in case I'm wrong.

Yes, this applies to the KFX Output plugin.

Originally Posted by Silverbacknet

I can't find a repo or any way to make a pull request, so I'll just post here. This is a small fix for when Kindle Previewer is installed in a different folder.

Having the plugin automatically detect alternate installation locations for the Kindle Previewer would be useful. I will take a look at this and add it or something similar to the plugin when I have some free time. Thanks for the code contribution.

[Silverbacknet]

For some reason I hadn't found this thread in search, oops. Thanks for taking care of it!

FWIW, the "OSError" isn't used in Py2, only WindowsError, it's to future-proof for Py3 (in which 2to3 would change _winreg to winreg automatically, the only other Py3 compatibility concern).

[gers1978]

Error using latest plugin and Calibre 4.6:

calibre, version 4.6.0
ERROR: KFX conversion failed: <p><b>Cannot convert <book name></b><br><br><b>Conversion error:</b> Kindle Previewer error: Error:E00015: Unsupported tag or style was found. Tag: (), HTML Attribute: (), CSS Style: (-webkit-box-shadow=rgb(170, 170, 170) 0px 0px 5px inset).</p>

Happens on multiple input epubs