Better Hyphenation

[JSWolf]

Originally Posted by serser

All right!

I just found one last issue today: when a word starts with an exclamation or question mark, the Spanish dictionary sometimes hyphenates words like this: ¿S- / abías, when it should be at least ¿Sa- / bías.

Could you address this issue in the same way you resolved the one for the Catalan dictionary?

Thank you SOO MUCH!

I've edited the Spanish hyphenation dictionary. Please let me know if it works better. Thanks.

[serser]

Originally Posted by JSWolf

I've edited the Spanish hyphenation dictionary. Please let me know if it works better. Thanks.

The previous issue seems to be solved, but I’ve found two new problematic cases (with the Spanish dictionary):

1) The word pesadilla (followed by a comma) is hyphenated as pes- / adilla, but in Spanish, the syllables of the word are: pe.sa.di.lla. Therefore, the hyphenation should be one of the following: pe- / sadilla, pesa- / dilla, or pesadi- / lla.

2) The word radar (followed by a period) is hyphenated as rada- / r., but it should be ra- / dar.

Are these issues caused by a faulty dictionary? If they are, here you can find another Spanish spellcheck dictionary: https://extensions.libreoffice.org/e...h-dictionaries

[serser]

The previous issue seems to be solved, but I’ve found two new problematic cases (with the Spanish dictionary):

1) The word pesadilla (followed by a comma) is hyphenated as pes- / adilla,, but in Spanish, the syllables of the word are: pe.sa.di.lla. Therefore, the hyphenation should be one of the following: pe- / sadilla, pesa- / dilla, or pesadi- / lla.

2) The word radar (followed by a period) is hyphenated as rada- / r., but it should be ra- / dar.

Are these issues caused by a faulty dictionary? If they are, here you can find another Spanish spellcheck dictionary: https://extensions.libreoffice.org/e...h-dictionaries

Thank you!

[JSWolf]

Originally Posted by serser

The previous issue seems to be solved, but I’ve found two new problematic cases (with the Spanish dictionary):

1) The word pesadilla (followed by a comma) is hyphenated as pes- / adilla,, but in Spanish, the syllables of the word are: pe.sa.di.lla. Therefore, the hyphenation should be one of the following: pe- / sadilla, pesa- / dilla, or pesadi- / lla.

2) The word radar (followed by a period) is hyphenated as rada- / r., but it should be ra- / dar.

Are these issues caused by a faulty dictionary? If they are, here you can find another Spanish spellcheck dictionary: https://extensions.libreoffice.org/e...h-dictionaries

Thank you!

The problem is that I can either have the hyphens allowed after two characters or three. If I go back to two, then the previous problem will happen again. The dictionary using is a Libre Office dictionary. So you can either have it as it was or as it is. Your choice.

[serser]

Alright. But if you allow hyphenation after either 2 or 3 characters, shouldn’t the dictionary only permit correct word hyphenation in that language given that condition? For example, if you allow hyphenation after 2 characters, it should hyphenate pe-/sadilla, pesa-/dilla, or pesadi-/lla; and if you allow it after 3 characters, pesa-/dilla or pesadi-/lla. Isn't it? However, it should never hyphenate as pes-/adilla because it splits a syllable between two lines, which is not correct. Isn’t that a spellcheck dictionary error? That’s why I suggested changing the dictionary...

In any case, thank you so much once again.

[JSWolf]

Originally Posted by serser

Alright. But if you allow hyphenation after either 2 or 3 characters, shouldn’t the dictionary only permit correct word hyphenation in that language given that condition? For example, if you allow hyphenation after 2 characters, it should hyphenate pe-/sadilla, pesa-/dilla, or pesadi-/lla; and if you allow it after 3 characters, pesa-/dilla or pesadi-/lla. Isn't it? However, it should never hyphenate as pes-/adilla because it splits a syllable between two lines, which is not correct. Isn’t that a spellcheck dictionary error? That’s why I suggested changing the dictionary...

In any case, thank you so much once again.

It's nothing to do with a spellcheck dictionary. It's a hyphenation dictionary. I don't know what would have to be deleted to fix the 2 character left hyphenation.

[serser]

Originally Posted by JSWolf

It's nothing to do with a spellcheck dictionary. It's a hyphenation dictionary. I don't know what would have to be deleted to fix the 2 character left hyphenation.

Ok, I mean: the dictionary tells the Kobo how to split words in accordance with the rules of each language (e.g., in this language, words can be split this way or that way) and the parameters you establish (i.e., hyphenation after 3 characters). A correct Spanish dictionary, however, should not suggest splitting a syllable in half, as it does with pes-/adilla. That’s why I suggested trying another base dictionary to create the hyphenation dictionary.

Anyway, thank you once again.

[belle39234]

first of all, thank you for this. is there instructions how to create these for other languages? I'd like one for slovenian language, it's very niche but I need it and I'm totally happy to create it myself!

[JSWolf]

Originally Posted by belle39234

first of all, thank you for this. is there instructions how to create these for other languages? I'd like one for slovenian language, it's very niche but I need it and I'm totally happy to create it myself!

I have no directions written on how to do this. It's not all that difficult to do. What can be the most difficult is to find the hyphenation dictionary to use. I've been using hyphenation dictionaries from Libre Office. If you can find a hyphenation dictionary, link it here and I'll create a hyphenation dictionary for you to try.

[belle39234]

Yes please! Would this be okay? https://extensions.libreoffice.org/e...ictionary-pack

[JSWolf]

Originally Posted by belle39234

Yes please! Would this be okay? https://extensions.libreoffice.org/e...ictionary-pack

I'll have a look at this.

[shamanNS]

Here's a KoboRoot.tgz I created that adds "hyphen_sr.dic" for Serbian language. The hyphenation dictionary contains both serbian cyrillic script and serbian latin script hyphenation rules ( = the same approach that KOReader / crengine uses for Serbian books) in a single file so it should work for both cyrillic and latin script books.

serbian_Hyphens_Kobo.zip

[dcbrow2]

This post concerns a possible method to avoid hyphenation in the word thing ("th-ing").

I did a brief study on the format of the dictionary and while I'm not really clear in my mind about it, I can say that the numbers tell the software whether a given word location is a good place to put a hyphen (if needed) or a bad place to put a hyphen. Basically odd numbers favor hyphens (higher numbers are more favorable) and even numbers discourage hyphens (high numbers are more dis-favorable). Using this idea and the information here, I tested adding these 2 lines to the dictionary file provided by JSWolf (the English one):

t8h8i8n8g8
t8h8i8n8g8s8

I put those 2 lines near the top, above the first pattern (ie above the .a2 line). I believe this is basically tells the hyphenation engine to not hyphenation "thing" and "things".

I found a place in a book where "thing" was hyphenated as "th-ing" using a dictionary without those lines and was not hyphenated using a dictionary with those lines (no other changes between the two dictionaries). But it is not rigorously tested.

Also, thing is a challenging pattern of letters because "cloth-ing" is a valid hyphenation place but "th-ing" is not.

YMMV, but other in this thread reported seeing "th-ing" so I wanted to report my tests. Good luck!

If it's not clear, you have to unpack the dictionary so you can see the usr\local\kobo\hyphenDicts folder, edit the dictionary with a text editor and then repackage it to KoboRoot.tgz and drop it into the hidden .kobo folder on your device.