[GUI Plugin] Hyphenate This!

[MarkFalk]

I am using "Hyphenate-This!" with azw3-files for "old" kindle-reader and German texts. With nearly every (German) text the hyphenation with the plug-in works.
But some days ago I have converted a German text from pdf to epub (and after editing converting again to azw3, but his makes no difference) and tried several time adding soft-hyphens with the "Hyphenate This!"-plugin. After some time of investigations I realized that the plug-in only added soft-hyphens to text which is enclosed by / between <i class="calibre9"> and </i>, not into "normal" text w/o this tags. Checking this with ebook-edit and searching for "\xad" as "regex" soft-hyphens are only found in text in/between <i class="calibre9"></i> and only this text is displayed with hyphenation. After some research I have learned hat also ebook-polish does insert soft-hyphenations and after doing this with ebook-polish the soft-hyphens are in the complete text: They are found with "\xad" and also are working when displaying the text.
Seems as if the plug-in has a bug and ebook-polish does some things different than the plug-in.
After this I remembered vague that I have realized in the past, that the plug-in did not work with a few other ebooks. Unfortunately I don´t know enough about the structure of .epub, .azw3 etc. for further analyzes.

[Ancanar]

Hello.
I use hyphenate this on my ebooks but although it makes the words hyhen here and there it still does not make all spaces between words equally white as I expected it to do. You can see on the attached screenshot that although some words are hyphenated the space marked in red is much wider than the one marked in green (it's from Calibre preview but looks exactly the same on my Kindle). Is the way to set up the plugin to make spaces equally wide like in real paper books?

[DNSB]

@Ancanar: Posting duplicate messages is frowned upon. Please avoid doing so.

See this thread in the Amazon Kindle forum: Hyphenate This doesn't hyphen "as much as it should" .

[tudoapampa]

Originally Posted by Ruskie_it

Ivano I got that!
Quick and easy: you don't have to load the entire oxt file into hypenate this!
It normally should work, but for some reason it confuses the plugin.
Just open the .oxt file with winzip, then extract hyph_it_IT.dic file from "dictionaries" subdirectory and load that into Hypenate this!
This way the hypenation will work.

PS: Sorry for the repeated posting. :P

thank you! You saved me

[C4thie]

I would like to use KFX format, but it doesnt support hyphenation for my native language. Is it possible to add hyphens with this plugin to EPUB file first. Then convert it to KFX to keep hyphenation? Or converting to KFX is going to remove these changes?

[JSWolf]

Originally Posted by C4thie

I would like to use KFX format, but it doesnt support hyphenation for my native language. Is it possible to add hyphens with this plugin to EPUB file first. Then convert it to KFX to keep hyphenation? Or converting to KFX is going to remove these changes?

What is your native language? A hyphenation dictionary would have to be found for your language.

You would have to convert to KF8 and then add soft-hyphens. I just did a test and the soft-hyphens I added to the ePub were removed from the KFX. So for KFX, it's a no.

I'm thinking that reason soft-hyphens get removed is because KFX already has hyphenation.

[shamanNS]

Originally Posted by C4thie

I would like to use KFX format, but it doesnt support hyphenation for my native language. Is it possible to add hyphens with this plugin to EPUB file first. Then convert it to KFX to keep hyphenation? Or converting to KFX is going to remove these changes?

I tested this last autumn with sending soft-hyphanted epub via Send to Kindle web page for book in serbian language (serbian Latin script actually, didn't test the Cyrillic script) that also not official supported by KFX format and KFX file deliver to Kindle for Android app had hyphens.

So, couple more notes about this:
- converting books in serbian language using Kindle Previewer directly (so not using KFX output Calibre plugin that now does the language metadata temporary change as a workaround for unsupported languages) wouldn't produce a KFX/KPF file because the language is unsupported.

- if you send epub without soft-hyphens added using Send to Kindle you'll get a KFX file with no hyphens renderer when reading a book. Also no need to fake / change the language metadata when sending via Send to Kindle web page, unlike with Kindle Previewer.

[karakai]

The latest Hungarian dictionary for the plugin:
hyph_hu_HU_processed.zip

[Moonbase59]

Is @SauliusP still active here? The forum says last activity in 2020?

If not, who is maintaining "Hyphenate This!"? Is there a GitHub page?

I’d like the original author (or current maintainer) to check with my German hyphenation file, since it seems to do nothing on the file I’ve tested with.

Other users using German are also invited to test, of course. I attach my current hyph_de_DE.dic file here. In the plugin, I’ve set min word length to 6, and both left and right hyphen min to 3. The dictionary displays as "deu.dic - German" in the plugin, for some reason.

[BeckyEbook]

@Moonbase59: I do not know the reason (perhaps an incomplete implementation), your version of the dictionary does not work, but I know how to fix it.
Simply delete the last 229 lines from the dictionary file (the entire sections on U+00A0 and U+202F).

[Moonbase59]

Awww… that sucks. So I have to make yet another "special version" of a standard hyphenation dic?
Thanks for the feedback anyway, @BeckyEbook.

Looks like "Hyphenate This!" must be using some old or incomplete implementation of the dic reader then. I’ve tested the hyph_de_DE.dic with lots of software, from e-readers to LibreOffice and many others (it is also my standard Linux hyphen dic). The extra (weight 8) additions were introduced because OpenOffice at that time couldn’t handle hyphenation at word boundaries (adjacent to one of these characters) correctly, and some readers couldn’t as well (Tolino Vision 5 for example).

This could result in real bad things like

Bla bla. »Hallo
!«.

Would be real great to know who now maintains this plugin, so he/she could correct the bug…

[BetterRed]

Originally Posted by Moonbase59

. . .
Would be real great to know who now maintains this plugin, so he/she could correct the bug…

SaliusP. hasn't logged on for almost 5 years.

So, no one is actively maintaining it, and AFAIK no one has shown any interest in doing so.

BR

[Moonbase59]

So I did a quick check. "Hyphenate This!" uses the Python hyphenator which is from Dec 2019.

This apparently doesn’t work well with the NOHYPEN replacement stuff I added for OpenOffice and RMSDK a long time ago, while, as @BeckyEbook pointed out, it works fine with that "special part" left off:

>>> from hyphenator import Hyphenator
>>> h = Hyphenator("/home/matthias/Projekte/wortliste/build-patterns-2-2/dehyph\
n-x/hyph_de_DE.dic")
>>> h.left = 3
>>> h.right = 3
>>> h.inserted('Donaudampfschifffahrtsgesellschaft')
'Donaudampfschifffahrtsgesellschaft'
>>> 
>>> h = Hyphenator("/home/matthias/Projekte/wortliste/build-patterns-2-2/dehyph\
n-x/hyph_de_DE_nohyphen.dic")
>>> h.left = 3 # recommended for German
>>> h.right = 3 # recommended for German
>>> h.inserted('Donaudampfschifffahrtsgesellschaft')
'Donau-dampf-schiff-fahrts-ge-sell-schaft'
>>> h.left = 2 # not recommended for German
>>> h.right = 2 # not recommended for German
>>> h.inserted('Donaudampfschifffahrtsgesellschaft')
'Do-nau-dampf-schiff-fahrts-ge-sell-schaft'
>>>

So, for the time being, I include my current German hyph_de_DE.dic (for all normal uses) as well as hyph_de_DE_nohyphen.dic (for everything that uses the Python hyphenator, like the Calibre "Hyphenate This!" plugin).

Many thanks to @BeckyEbook for finding the issue!

Happy hyphenating!

N.B.: I still don’t recommend adding soft hyphens to files—hyphenation should be done by the reader software (as KOReader shows nicely, or even RMSDK using my hyph_de.dic). But I do know there are some readers where this is almost impossible (or at least not easy), so now we at least have the choice.