[Conversion Output] KFX Output

[davidfor]

Originally Posted by jhowell

I was reluctant to implement it that way since it requires manual effort to set the desired number of pages for each book converted, thus not being usable for bulk conversion. But since I don't see a better alternative I will go ahead and do that in the next plugin release.

Just to check, but is this a conversion thing or a metadata thing? If it's metadata, then maybe the metadata writer should do it. That should have access to the custom columns.

[jhowell]

Originally Posted by davidfor

Just to check, but is this a conversion thing or a metadata thing? If it's metadata, then maybe the metadata writer should do it. That should have access to the custom columns.

Thanks for the suggestion!

Moving the page number creation function from the converter to the metadata writer may be the best solution. I will look into it.

[davidfor]

Originally Posted by jhowell

Thanks for the suggestion!

Moving the page number creation function from the converter to the metadata writer may be the best solution. I will look into it.

After writing my suggestion, I realised I didn't know how this was to be used. Is this to generate something like the page list in an ePub? Does it need to modify the actual text? To add an anchor or other reference? If it is, then the metadata writer is probably not a good place. If it is an independent file or structure, referencing existing points in the book, it should work except for the extra time.

[jhowell]

Originally Posted by davidfor

After writing my suggestion, I realised I didn't know how this was to be used. Is this to generate something like the page list in an ePub? Does it need to modify the actual text? To add an anchor or other reference? If it is, then the metadata writer is probably not a good place. If it is an independent file or structure, referencing existing points in the book, it should work except for the extra time.

In KFX the page number information is somewhere in between. It is a list of pointers into the content that is included within the KFX file. I would prefer to generate it as part of conversion, and will still do so as part of the KFX Output plugin, but I will also add the ability to update the page list as part of the KFX Metadata Writer since the content of custom columns is available there.

[davidfor]

Originally Posted by jhowell

In KFX the page number information is somewhere in between. It is a list of pointers into the content that is included within the KFX file. I would prefer to generate it as part of conversion, and will still do so as part of the KFX Output plugin, but I will also add the ability to update the page list as part of the KFX Metadata Writer since the content of custom columns is available there.

So, like the page list but no anchors in the book. That sounds like it could be a separate plugin, but the metadata writer is probably sensible. And both also make sense.

And what I find interesting is that it sounds like the page count doesn't have to have any real relationship to the size of the book. If it is a 100,000 word book and someone says it is 10 pages long then the page index would be built that way.

An alternative is to do the page count while doing the conversion. The Count Pages plugin does have a callback for other plugins to use it, but it runs the count in a job, so it probably isn't suitable here. But, it does have sample code for a few algorithms.

[jhowell]

Originally Posted by davidfor

So, like the page list but no anchors in the book. That sounds like it could be a separate plugin, but the metadata writer is probably sensible. And both also make sense.

And what I find interesting is that it sounds like the page count doesn't have to have any real relationship to the size of the book. If it is a 100,000 word book and someone says it is 10 pages long then the page index would be built that way.

An alternative is to do the page count while doing the conversion. The Count Pages plugin does have a callback for other plugins to use it, but it runs the count in a job, so it probably isn't suitable here. But, it does have sample code for a few algorithms.

Where this process is best implemented seems a gray area to me. The number of pages in the book could be considered to be a piece of metadata, but their exact placement seems more like content to me. At the moment I plan to allow page number creation during conversion and allow it to be updated when metadata is embedded.

I plan to allow those users who care about having a specific number of pages to determine it themselves using their preferred method. For everyone else there is already a simple method implemented within the plugin that yields a page count close to the actual number of pages as given by a pre-existing EPUB pagelist in a set of books I tested with.

The exact placement of those pages within the book is already handled using another simple algorithm and I do not plan on changing that.

I am still considering the best approach, but I don't want to spend much more time on this. There aren't that many users of this plugin.


Added: After checking again I was able to find the custom column information in a place accessible during conversion so I will add the ability to configure the lookup name for the custom column containing the page count.

[jhowell]

Version 1.14.0 - 05 Jan 2018

Allow the lookup name for a custom column holding the desired number of pages for each book to be specified when creating approximate page numbers.

Improve support for Kindle Previewer version 3.17. Handle change to the format of conversion guidance. Allow conversion of books containing MathML.

Remove support for Previewer versions older than 3.15. (Newer Previewer versions produce better results.)

Retry conversion with location map generation disabled if the first conversion fails with an accuracy threshold error. This allows more books to be successfully converted. (Previewer 3.17 or later is required.)

Additional fixes applied to the source EPUB before conversion.

Prevent plugin installation under Linux since Kindle Previewer 3 does not run under that OS.

[jhowell]

Version 1.14.1 - 05 Jan 2018

Fix error when updating to version 1.14.0: ImportError: cannot import name AUTO_PAGES


Note: If you received this error while updating to the previous version you may need to re-install the plugin.

[terendel]

Since I installed 1.13, although I am selecting the option to make the KFX files books instead of "Docs," every time I convert and send to my Oasis 2, they display as Docs on my Kindle. Well, until they don't. Sometimes a previously loaded book suddenly displays as a Doc and sometimes a previously displayed Doc suddenly shows as a book.

It seems kind of random and maybe it's just something weird with my Kindle, but it wasn't happening before 1.13.

[jhowell]

Originally Posted by terendel

Since I installed 1.13, although I am selecting the option to make the KFX files books instead of "Docs," every time I convert and send to my Oasis 2, they display as Docs on my Kindle. ...

I would like some clarification on what you are seeing.

Is it that when you select the drop down choice at the top of the library screen to show only "Books" or "Docs" those books only show under the Docs category?

If you long press the book entry on the library screen do you see the menu of choices appropriate for Books or Docs? (Books include "Book Description", "Search This Book", and also "View on Goodreads", "Add to Goodreads Shelf", and "Share" if you have Goodreads enabled. Docs include "Search This Document" and "Delete This Document".)

Is it that the cover thumbnail isn't shown? (It will only show if you have an ASIN set for the book in calibre prior to conversion.)


Added: I left out the most obvious indication. If you switch the library to "List" instead of "Grid" does the word "DOC" appear in the line under the title of the book?

[terendel]

Yes, DOC appears under the title when I am in list view

[jhowell]

Version 1.14.2 - 08 Jan 2018

Fix bug introduced in version 1.14.0. The ASIN value from book identifiers is not being used during conversion.

[jhowell]

Originally Posted by terendel

Sometimes a previously loaded book suddenly displays as a Doc and sometimes a previously displayed Doc suddenly shows as a book.

The type of a book changing sometime after it is loaded may possibly point to a problem in indexing books. This is a process that occurs in the background over time after books are loaded to the Kindle. Sometimes unexpected content causes the indexer to fail in strange ways and since the KFX is undocumented I cannot be sure that the files created by this plugin do not contain content that could trigger this problem.

It hasn't happened to me, at least not that I have noticed. If you discover anything else that could help me to reproduce the problem please post again here.

[thecfg]

Stupidly (I think!) I allowed Kindle Previewer to update to 3.17.1, and now my KFX conversion (on lots of books, ones that have converted no probs) with plugin 1.14.2 yields

calibre, version 3.15.0
ERROR: KFX conversion failed: Cannot convert <book> Conversion error: Kindle Previewer error: Exception in thread "main" com.b.a.n: Unknown option: -allowYJConversionForJP

perhaps someone could point me to a safe downgrade version of previewer? or another option?

am I dumb? LOVE what this plugin does btw

thanks!

[jhowell]

Originally Posted by thecfg

I allowed Kindle Previewer to update to 3.17.1, and now my KFX conversion (on lots of books, ones that have converted no probs) with plugin 1.14.2 yields...

Someone had to be the first to try it!

As you describe, Kindle Previewer 3.17.1 is not compatible with this plugin. There appear to be several changes that impact compatibility so I am not sure how long it will take to fix this. Is is even possible, but unlikely, that I will not be able to work around the changes.

Until then I suggest refusing the update or reverting back to 3.17.0 if you have already updated. I do not have a link to a previous version. Amazon reuses the same URL for new releases.