[Conversion Output] KFX Output

[jhowell]

Originally Posted by abogomolov

Wow, isn't it something eager to be optimized?

I never bothered to optimize it because it worked fast enough with my books. But there is one part of the code where the time used increases rapidly with the complexity of the book.

I have now optimized that and it will be part of the next plugin release. One large book I tested takes three hours to convert with the current version. It will take about three minutes with the new one. There will be only a slight increase in speed for shorter and less complex books.

[mathil]

That's great news, jhowell! In case anyone is interested, the book I was trying to convert as per my previous post was The Complete Malazan Book of the Fallen (omnibus edition containing 10 books of the series), 12.8 mb ca. in size. It was successfully converted to KFX format, with the conversion taking 6 hours in total.

[jhowell]

I plan to release the new plugin version in the next few days. I need to test it with the just released Kindle Previewer 3.13. (No longer called a "beta" by Amazon.)

Added: So far the plugin works properly with Previewer 3.13. As with most Previewer releases, this one converts more books than prior ones.

A new version of the plugin should be ready sometime next week. Until then the current version is OK to use with Previewer 3.13.

[AcidWeb]

Re-tested 3.13 comic output. No changes. Still no Panel View and e-books are unstable.

[jhowell]

Version 1.9.0 - 09 Aug 2017

Remove support for Previewer versions older than 3.13.

Allow English language MS Word DOC and DOCX files and MOBI files (created using kindlegen version 2.9 without the -dont_append_source option) to be converted to KFX using the Command Line Interface.

Allow Chinese language books to be converted.

Improve performance when converting large and complex books.

Improve reporting of conversion errors. Do not truncate (tail) Previewer conversion logs in the calibre conversion job log.

Handle more EPUB problems that would otherwise cause the Previewer to fail. (badly sized transparent GIFs, "onload" attribute)

Further internal changes based on the Amazon Ion specifications.

[bobb40]

Can the plugin work (or be made to work) with kindlegen? That's available to linux, where previewer isn't, but previewer is said to use kindlegen to convert epub to kindle.

[AcidWeb]

No. If KindleGen could generate KFX this entire plugin would not be needed.

[shoelesshunter]

I am using the KFX output plugin on Window 7. When I return home, I would like to continue using it on my linux machine.

The plugin looks for Kindle Previewer in it's default location on Windows.

Is there a trick to using Wine to make this happen on linux?

Thanks.

Moderator Notice
Post moved - please post questions about a plugin in the plugin's thread.

tks - BR

[Kaan88]

Is there an automated way to generate page numbers for epub files? NCX pageList, NAV page-list, page-map.xml all require manually adding them to epub file as far as I could find. Thanks.

[jhowell]

Originally Posted by bobb40

Can the plugin work (or be made to work) with kindlegen? That's available to linux, where previewer isn't, ...

Originally Posted by shoelesshunter

Is there a trick to using Wine to make this happen on linux?

The plugin launches a Windows application (Java) that is part of the Previewer in order to perform the bulk of the KFX conversion process. That in turn launches several other Windows executables to perform subtasks.

I do not use Linux or know how this would work under that platform. If anyone has been able to get the Previewer to run in that environment using Wine or knows why it cannot do so perhaps they will post here.

Originally Posted by Kaan88

Is there an automated way to generate page numbers for epub files? NCX pageList, NAV page-list, page-map.xml all require manually adding them to epub file as far as I could find. Thanks.

I don't know of any software that can do this.

It could be done using a calibre plugin. The problem would be finding someone interested enough to develop it. (Not me.)

[sdspieg]

Any ideas on this one?

ERROR: yj.eidhash_eid_section_map has extra: XW26tKp0SoqBSddYmlQhlw118/XW26tKp0SoqBSddYmlQhlw119

[jhowell]

Originally Posted by sdspieg

Any ideas on this one?
ERROR: yj.eidhash_eid_section_map has extra: ...

The plugin was written based on my reverse engineering of the output of Kindle Previewer. I put in some safety checks for conditions that I never expected to occur and that is one of them.

Since I do not have any books that cause that condition I cannot replicate the problem. I do not know what is causing it or how to prevent it.

[jhowell]

Version 1.10.0 - 30 Aug 2017

Adjustments to output format to better match KFX files currently produced by Amazon. (Reading these files with the latest Kindle firmware/app version is recommended for best results.)

Fix occasional error during conversion: "TypeError: info() takes exactly 2 arguments (3 given)"

Attempt to work around occasional error during conversion caused by unexpected data produced by the Previewer: "yj.eidhash_eid_section_map has extra: ..."

[HowlerX]

Don't know what you changed, but now code like:

<span epub:type="pagebreak" id="page39" title="39"></span>

or

<span id="page39" title="39"></span>

is adding a line of space where there shouldn't be one. These are usually used for page numbers in epub. Before version 1.10 of your plugin the line space wasn't there, but now it is. Explanation?

[JSWolf]

Originally Posted by HowlerX

Don't know what you changed, but now code like:

<span epub:type="pagebreak" id="page39" title="39"></span>

or

<span id="page39" title="39"></span>

is adding a line of space where there shouldn't be one. These are usually used for page numbers in epub. Before version 1.10 of your plugin the line space wasn't there, but now it is. Explanation?

They are not usually used for page numbers. They are usually used to mark where some version pBook page number is in the ePub. In most ePub, those two sets of spans can be deleted. Also, it's also inserted as <a id="page39"/>.

What I do is regex remove them and the do a check of the ePub to make sure they are not needed. If not too many are needed, I leave those and dump the rest.