check spelling

[kovidgoyal]

@jackie_w: calibre uses the ICU word break iteration algorithm, which as far as I recall, splits up most hyphenated words into two words (the details are language dependent), so, for example, abc-def will show up in the words list as two words, abc and def

See http://userguide.icu-project.org/boundaryanalysis for details

[jackie_w]

Oh, well. I was hoping this would be one of the Sigil features which would be replicated. A quick list of all hyphenated words is useful when dealing with a scanned source which has not been 'de-hyphenated' very well.

The new spellchecker is a welcome addition, though. Thanks

[BetterRed]

Originally Posted by kovidgoyal

@jackie_w: calibre uses the ICU word break iteration algorithm, which as far as I recall, splits up most hyphenated words into two words (the details are language dependent), so, for example, abc-def will show up in the words list as two words, abc and def

See http://userguide.icu-project.org/boundaryanalysis for details

that's a valuable link - explains some things I've been puzzling about wrt leading & trailing apostrophes.

At http://www.unicode.org/reports/tr29/#WB14 there is this with respect to word boundaries and hyphens

The correct interpretation of hyphens in the context of word boundaries is challenging ... it is better overall to keep the hyphen out of the default definition

Is that to be interpreted as... a hyphen should or should not constitute a word boundary... I'm inclined to read it as should not.

BR

[kovidgoyal]

@BR: You'll have to take it up with the developers of ICU, this is one swamp I have no intention of wading into.

[mrmikel]

Kovid, when you create a user dictionary, is it active on all documents, or is it possible to select ones to activate?

[arspr]

Hi Kovid, in the release notes for 1.33.0 it says:

It comes with builtin dictionaries for English and Spanish.

But when I try to run the check on Spanish books I get:

calibre, version 1.33.0
ERROR: Failed to check spelling: Failed to check spelling, click "Show details" for the full error information.

Traceback (most recent call last):
  File "M:\DVD Personal 55\Calibre\SourceCode\src\calibre\gui2\tweak_book\spell.py", line 1102, in get_words
  File "M:\DVD Personal 55\Calibre\SourceCode\src\calibre\gui2\tweak_book\spell.py", line 1102, in <dictcomp>
  File "M:\DVD Personal 55\Calibre\SourceCode\src\calibre\spell\dictionary.py", line 339, in recognized
  File "M:\DVD Personal 55\Calibre\SourceCode\src\calibre\spell\dictionary.py", line 205, in dictionary_for_locale
  File "M:\DVD Personal 55\Calibre\SourceCode\src\calibre\spell\dictionary.py", line 174, in load_dictionary
IOError: [Errno 2] No such file or directory: u'M:\\DVD Personal 55\\Calibre\\SourceCode\\resources\\dictionaries\\es-ES\\es-AR.dic'

(I've also tested with "official" 1.33 and I get the same error but loading from

IOError: [Errno 2] No such file or directory: u'C:\\Program Files\\Calibre2\\resources\\dictionaries\\es-ES\\es-AR.dic'

)





What am I doing wrong?

(Nevertheless if this is the "default" warning you get if you don't have the needed dictionary installed, I do really think it's veeeeeery un-userfriendly).

[BetterRed]

Originally Posted by kovidgoyal

@BR: You'll have to take it up with the developers of ICU, this is one swamp I have no intention of wading into.

Before I put on me wellies on and get me canoe out - I note that Sigil appears to be using a later version of the icu libraries than the one's that calibre uses.

BR

[kovidgoyal]

@BR: I have ICU 52.1 on my system and hyphenated words are still split up by the break iterator. IIRC the break iterator has been present for a very long time in ICU and I doubt the algorithm for english has changed anytime in the recent past. Not to mention, that I highly doubt sigil uses ICU for word iteration. I'd guess the only reason sigil includes ICU is because WebKit requires it.

@mrmikel: You can mark a user dictionary active or inactive. Active ones apply to all documents, inactive ones to none.

[kovidgoyal]

@arspr: https://github.com/kovidgoyal/calibr...b6581d78c46f01

[arspr]

Originally Posted by kovidgoyal

@arspr: https://github.com/kovidgoyal/calibr...b6581d78c46f01

Fixed

Another doubt or possible feature. Looking in the Manage Dictionaries window I can see that there are three built-in ones:

• British English
• US English
• Spanish (whatever flavour)

And I can see that, as simple example, Philippines English is associated to US English dictionary, and New Zealand to British one. But I cannot change that association (or I don't know how). Could it be a good idea? I mean, OK, I don't have a Kiwi English dictionary installed, why should I use the British one instead of the US one? Who has decided that Australian English is closer to British than to US one?

(Please keep in mind that I'm hypothetically speaking. In fact I don't really know if Ghana English is SO much closer to the British flavour that the previous possible choice just makes no sense at all. No offence meant in any way).

[kovidgoyal]

The associations come from the maintainers of the dictionaries. I am not going to get involved in second guessing them. If you find you prefer a US english dictionary for say ghana instead of the GB one, then just install a custom US english one and set it to be the one used for ghana. Or change your book to use the language code en-US instead of en-GH or whatever.

[Perkin]

@arspr, I think which association has to do with some of the root words, and their spelling.
(think colour/color or honour/honor)

[Divingduck]

Hi Kovid,
Thank you very much for integrating the import function. I made a first test it with my wordlists. Now I am missing some additional things (sorry).
Exporting words out of a user dictionary to clipboard and/or File. I can select more than one word, but I cannot copy them together to clipboard (only the first one in a list of more than one word).
Sometimes there is a need to make corrections in the word list. Is it possible to implement a function for edit a word, change of language for selected words and delete selected words?

[kovidgoyal]

There are already buttons to add and remove words. If you wish to edit, remove and then re-add the changed word.

[jbacelar]

The feature Import list of words works correctly, word by word and paste from the clipboard.
But ... simultaneously it opens the attached error window:



Thanks for all Kovid.