[GUI Plugin] Prince PDF

[Geremia]

I get this error:

calibre, version 6.14.0
ERROR: Unhandled exception: <b>AttributeError</b>:type object 'TextEditWithTooltip' has no attribute 'NoWrap'

[Jellby]

Good news. The new version 1.6.0 should be compatible with calibre 6 (probably not with older calibre versions though). There may still be some issues, please report any remaining problem.

Sorry for the long delay. I just needed some quality time to upgrade the OS in my computer, and I didn't get it until the Easter break.

[bravosx]

I've been waiting patiently and it's finally here.

@Jellby
Many thanks for your time and work.

[naisren]

Originally Posted by Jellby

Good news. The new version 1.6.0 should be compatible with calibre 6 (probably not with older calibre versions though). There may still be some issues, please report any remaining problem.

Sorry for the long delay. I just needed some quality time to upgrade the OS in my computer, and I didn't get it until the Easter break.

many thanks, wish you every day goes well.

[Comfy.n]

@Jellby

Hi! This is just a heads up, as I think I found something that interferes with this plugin's conversions: in Calibre 6.22 it fails converting some books with long titles like The Trained Memory (Newly Revised Edition With Commentary by Anthony Metivier) or Vegetable Diet: As Sanctioned by Medical Men, and by Experience in All Ages / Including a System of Vegetable Cookery

If I temporarily set the title to something shorter, it then succeeds.

Thanks for this awesome plugin!

[Jellby]

I'll try to have a look. Could it be that's not the length, but the presence of some special characters, like : or / ? (On which OS, by the way?)

[Comfy.n]

Originally Posted by Jellby

I'll try to have a look. Could it be that's not the length, but the presence of some special characters, like : or / ? (On which OS, by the way?)

yes, I forgot to mention, this is on Windows (10...ugh). I reckon this plugin is more tailored to Unixes. And I also had thought, at first, it could be related to some special characters, but that would need further investigation.

[Comfy.n]

Hi, I did some tests today and have attached a bunch of logs. The message is always the same: "error: can't open output file: No such file or directory"

The reason I said this issue could be related to long titles, instead of special characters like : and /, is because I had tried just removing those from the titles, but even then conversion would fail.

In doing the tests I found a simple workaround to make the conversion succeed. I'd just clear the title sort field in the metadata edit dialog, and then any of the troublesome ebooks would convert fine. Clearing the author sort field also seems to work. Then, after converting, I'd just revert that change.

I wasn't able to find a primary culprit, but I have some suspicions:

- sometimes, titles starting with "The" get that error
- titles with multiple authors sometimes wouldn't convert (I've used author:#=1 to filter those)
- maybe some titles with multiple commas, colons and parentheses can be troublesome

Anyway, using the workaround I'll be fine with Prince. BTW, it would be perfect if there was a way to run it with a single-click action, once the desired CSS is already selected! Is this possible within Calibre GUI?

[Jellby]

If I remember correctly, the output should be using the "Save to disk" template. What happens if you just save to disk from calibre? Does it work? Does it save it with the same path that failed with prince or is it somehow truncated/modified? Could you run with "calibre-debug --gui" (https://manual.calibre-ebook.com/gen...bre-debug.html) and see if there's more information?

[Comfy.n]

Originally Posted by Jellby

If I remember correctly, the output should be using the "Save to disk" template. What happens if you just save to disk from calibre? Does it work? Does it save it with the same path that failed with prince or is it somehow truncated/modified? Could you run with "calibre-debug --gui" (https://manual.calibre-ebook.com/gen...bre-debug.html) and see if there's more information?

Saving to disk seems to work fine with the same title mentioned in the attached log.

EPUB format was saved to: C:\tesssst\Burgos, Edgar Baro & Moreno, David\poder del ayuno_ Guia practica para bajar de aumentar tu bienestar (Spanish Edition), El\poder del ayuno_ Guia practica para bajar de peso, dition), El - Edgar Barrionuevo Burgos & David Moreno.epub

After clearing the Title Sort in MDE, and converting it with Prince, I tried the Save only PDF format to disk, and it was saved to: "C:\tesssst\Burgos, Edgar Barrionuevo & Moreno, David\Unknown\Unknown - Edgar Barrionuevo Burgos & David Moreno.pdf"

edit: the conversion recorded in the log file did not succeed, maybe I should clarify. The PDF saving test was done after doing the workaround metadata edit

[Comfy.n]

Prince is blazing fast, most times.

glad we have this alternative!

[Jellby]

I'm confused. The "save to disk" path is shortened to a different length than the output file for Prince (e.g. "Barrionuevo" becomes "Barrvo" or "Baro"). I confess I don't know how this is supposed to work, and it's been a while since I last looked at the code. I still need to find some time for proper debugging, but you could also try to change your save-to-disk template to use shorter paths and see if that helps.

[JSWolf]

Originally Posted by Comfy.n

yes, I forgot to mention, this is on Windows (10...ugh). I reckon this plugin is more tailored to Unixes. And I also had thought, at first, it could be related to some special characters, but that would need further investigation.

Nothing wrong with Windows 10. It's quote good.

[Comfy.n]

Oh I hadn't noticed that author middle name change... though that trimming of 'Edition' to ', dition)' did puzzle me.

I have this in save template: {author_sort}/{title}/{title} - {authors} , and just now I spotted this possibly relevant option: "Convert non-English characters to English equivalents" (it's checked).

Anyway, that "clearing" workaround seems enough for my current use. I loved that Prince is pretty fast and allows setting a transparent background for PDFs.

[Jellby]

Since this happens, apparently, in the Prince step, perhaps you could manually run Prince on command line with different output paths, and see if you can reproduce the error with some combination of long filename and "unusual" characters.

Nothing wrong with Windows 10. It's quote good.

Do you mean it's "good"