[Device Plugin] Kobo Touch Extended Driver

[jgoguen]

New version uploaded that seems to fix it. Optimizing the filename callback went a little too far

Calibre should offer the update soon.

[Terisa de morgan]

Originally Posted by jgoguen

New version uploaded that seems to fix it. Optimizing the filename callback went a little too far

Calibre should offer the update soon.

That wonderful code embellishment Thank you very much.

[patrik]

Originally Posted by jgoguen

New version uploaded that seems to fix it. Optimizing the filename callback went a little too far

Calibre should offer the update soon.

Tested with new 2.5.1. version. All is well again!

Thanks

[anavin]

I just noticed a behaviour on the plugin (not sure if this is a new behaviour Update: no it's not).

After conversion, some of the punctuations are wrapped separately (specifically, ".?!").
I think it is made for "span each sentence instead of paragraph", which should be more precise on reading position or highlight.

However, in some of my books, numbers / punctuations are wrapped in a span like this:

<span class="tcy">12</span>
<span class="tcy">?!</span>

And the tags become like this after convertion (new line added by me):

<span class="tcy">
<span class="koboSpan" id="kobo.196.2">?</span>
<span class="koboSpan" id="kobo.196.3">!</span>
</span>

Which breaks the meaning that "?" and "!" should be wrapped together.

I know that "span each sentence" does have its needs, but is there any way to make it not happen on such situation? (eg. don't "span each sentence" when the book is in CJK, or make it an option)

[jgoguen]

Are the numbers handled correctly in this case?

[anavin]

Originally Posted by jgoguen

Are the numbers handled correctly in this case?

Yes, they are still wrapped together.

To be more precisely, the numbers one becomes:

<span class="tcy"><span class="koboSpan" id="xxx">12</span></span>

The Access render engine actually won't make the CSS set for class="tcy" (in this case, I use ".tcy: {-epub-text-combine: horizontal}" on it) take effect on the "12" when it is wrapped again in another span, so I did have trouble on this case too. But this can be solved by writing this in kobo_extra.css:

.tcy span { -epub-text-combine: horizontal; }

[anavin]

Just saw my backup plugin; I don't think its a "new" behaviour at all. I didn't encounter it simply because I hacked the container.py in all previous versions (return from __append_kobo_spans_from_text when encounter all span or a).

[jgoguen]

So, since you've already fixed the bug, perhaps you can contribute the fix to the Github repo?

[Lilarcor]

@jgoguen, thanks for the great plugin .I am using Kobo H2O and fresh install calibre with your plugin, other kobo plugin is disabled. When I send a book to device, it never convert book to kepub in device. Does this plugin support H2O.

False alarm, I disabled all other driver, now it works.

[Patricius]

There is a bug. You can use the old version before it gets fixed.
https://www.mobileread.com/forums/att...7&d=1433608839

[anavin]

Originally Posted by jgoguen

So, since you've already fixed the bug, perhaps you can contribute the fix to the Github repo?

Hmm... what I had done is simply "not kobospan all <span> and <a>*", won't it be too over...?

Anyway I'll do that later today after I leave office (and learn how to use github)

*<a> is for footnotes which looks like <a class="footnote">[12]</a>

[jgoguen]

Ah, my mistake, I was thinking you'd fixed that bug

[jgoguen]

Originally Posted by Patricius

There is a bug. You can use the old version before it gets fixed.
https://www.mobileread.com/forums/att...7&d=1433608839

Way ahead of you!

Originally Posted by jgoguen

New version uploaded that seems to fix it. Optimizing the filename callback went a little too far

Calibre should offer the update soon.

[norbusan]

Hi everyone,
I have a question concerning the different plugins and the conversion of Japanese books. I have the following plugins activated:
* Kobo Reader Device Interface 2.1.8 (Tomothy Legge and David Forrester)
* KoboTouch 2.1.8 (David Forrester)
* KoboTouchExtended 2.5.1 (Joel Goguen)
* Pandigital Kobo device interface 1.0.0 (Kovid Goyal)
* Kobo Reader 1.0.0 (Kovid Goyal)
* Kobo 1.0.1 (John Schember)

I have converted a epub (kf8) Japanese book that displays correctly on the Kindle in vertical settings right-to-left using Calibre "Convert books individually". I have selected the output Kepub, and removed the ticks from "cleanup epub" and "remove language code", but still the book arrives on the Kobo (GloHD) with the following properties:
* text runs left-to-right
* pagination is right-to-left
(the most inconvenient way ;-)

I faintly remember that in former times I used the Kobo Extended Plugin only (without the Kepub Input and Output plugins) and it did properly convert to Japanese epub3/kepub.

Thanks for any information and all the best

Norbert
* Kobo Reader

[anavin]

Originally Posted by jgoguen

Ah, my mistake, I was thinking you'd fixed that bug

So I guess I don't need to learn how to use github for now...

Anyway, here's what I had done to my container.py:

Spoiler:

...
    def __append_kobo_spans_from_text(self, node, text):
        if node.tag.endswith('span') or node.tag.endswith('a'):
            return False
        if text is not None:
...

...as you can see, I don't even know how to recognize the node tag properly

Originally Posted by norbusan

Hi everyone,
...

Hi Norbert,

I have the same plugin sets on my calibre too (except my KTE hack), but my books transferred to my kobo display as my expect: rtl + vertical-rl.

Just did a test on my Glo (3.2.0; a rather old version): copy an vertical layout book to my kobo, without rename it.
It seems that RMSDK engine won't recognize page-progression-direction="rtl" property on the opf file.
So, since the pagination goes from right-to-left, I think the book is, at least, renamed as .kepub.epub on device and is using Access engine.
On the other hand, since the text becomes left to right, there might be some modification on the CSS, making the text flow setting invalid.

I guess you can try:
1. copy-paste the book to the device, rename it as .kepub.epub, and see if the text flow is as the expect.
2. Then, check the book transferred on the device by calibre, and see what modification were done on it.