check spelling - Page 2

theducks · 04-18-2014, 09:08 AM

Quote:

Originally Posted by kovidgoyal

@Divingduck: Attach a file where you cannot jump to the next occurrence.

As for importing wordlists as user dictionaries, I can certainly implement that. However, in calibre a user dictionary contains not just a list of words but every word also has an associated language. Therefore, they cannot be exported as simple word lists without losing some information.

I dont understand why you need to turn off dictionaries. The spell check is perfectly capable of handling multilingual documents based on the language specified in the lang HTML attribute.

I 'correct' scanned Fantasy and Science Fiction. I sometimes create Universe specific dictionaries that do not belong elsewhere.

Having the ability to invoke a specific set of dictionaries is a plus.

kovidgoyal · 04-18-2014, 09:28 AM

@DrChiper: The spell check most definitely is aware of tags. In fact it will only check text in tags (other than script and style) and in the attributes alt and title (which are used for tooltips/descriptions). If you are getting words from places other than tags, then it is likely a bug, attach the file.

@theducks: That's what the user dictionaries are for.

@BR: I suggest setting up a user dictionary with your abbreviations. The correct solution for not flagging some words is not to simply ignore an entire file.

roger64 · 04-18-2014, 10:11 AM

Hi

This was the last big item left to do. How amazing this huge task has been completed in so few months! Congratulations and thank you.

French

Here is a link for a good French dictionnary in oxt format. Of course it's hunspell all along. It's used by Grammalecte, a grammatical extension to OpenOffice and LibreOffice. It is updated quite frequently. The 5.0.2 version is from January 2014. It works pretty well with the Editor.
http://www.dicollecte.org/grammalecte/telecharger.php

Elided forms

Take a word: articulateur. You also can find: l'articulateur. The editor will display it two times, in two separate places once as "articulateur", the other as "l'articulateur" because elided forms seem to be counted as independent words (same for d' and others). It is grammatically the same word.

GUI

Even on a full page window, I can, at most visualize 20 words. I should at least be able to see a healthy 50 and it seems really necessary to be able to narrow the fields.

Several languages together

What exactly should I write and where if I have a book in French with lots of English words inside? I start from here.
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="fr-FR">

mrmikel · 04-18-2014, 10:56 AM

I have been on the lookout for a British-american dictionary, since I work with a lot of British works and also american or works about WWII where both are used at the same time.

kovidgoyal · 04-18-2014, 10:59 AM

@roger64: Regarding treatment of elided forms, as I am not a french speaker, it is rather difficult for me to write rules for combining words. At the moment all distinctly spelled words are treated differently. Patches are welcome to change that.

If you want multiple languages, you do something like this:

<div lang="en">Some english text<span lang="fr">some french words</span></div>

Basically when the lang attribute exists on a tag, the contents of that tag are interpreted in the specified language. You can put a lang attribute on any tag, and they can be nested.

roger64 · 04-18-2014, 11:23 AM

Thanks

I will get in touch with the Grammalecte author about elided forms.

Divingduck · 04-18-2014, 11:27 AM

Quote:

Originally Posted by kovidgoyal

As for importing wordlists as user dictionaries, I can certainly implement that. However, in calibre a user dictionary contains not just a list of words but every word also has an associated language. Therefore, they cannot be exported as simple word lists without losing some information.

I saw the prefs.json and understand what you mean. Maybe you can add in an import dialog a possibility for choosing a language of a user dictionary. Then there is the flexibility using your optimized setting and in addition with an import functionality. For export it maybe a good idea to add this in a default filename (e.g. user_de.txt)

Quote:

I dont understand why you need to turn off dictionaries. The spell check is perfectly capable of handling multilingual documents based on the language specified in the lang HTML attribute.

Well, I try to explain it on a workflow. Germans read many translated books. These book are often using phrases or words what might be spelled correct and sometimes not. Sometimes they use colloquial words or an author makes a decision to write a word in a special way. This makes it sometimes a bit difficult to understand whether a word was spell correct or not. (Especially if a check uses a mixed language database)
I often digitize books what use these kinds of mixed things. For this my best workflow is to use only the native spelling language and in addition one specialized (user) dictionary build up only for this particular book what apply to all other used words. Sometimes I use a second one for book series what apply only to special words used in a particular series.

Anyway, I can stay as well with the current solution as I saw there is a work around possible by deleting the standard dictionaries.

Divingduck · 04-18-2014, 11:37 AM

Quote:

Originally Posted by kovidgoyal

@roger64:
<div lang="en">Some english text<span lang="fr">some french words</span></div>
Basically when the lang attribute exists on a tag, the contents of that tag are interpreted in the specified language. You can put a lang attribute on any tag, and they can be nested.

Interesting and good to know for future projects. Please put this hint as a side note in your help file too.

kovidgoyal · 04-18-2014, 11:59 AM

Quote:

Originally Posted by Divingduck

Interesting and good to know for future projects. Please put this hint as a side note in your help file too.

It is in the manual, was it not clear?

Quote:

Words are shown with the number of times they occur in the book and the
language the word belongs to. Language information is taken from the books
metadata and from ``lang`` attributes in the HTML files. This allows the spell
checker to work well even with books that contain text in multiple languages.

mrmikel · 04-18-2014, 12:00 PM

An example might be helpful to those of us who are not so knowledgeable. It is not so common a construction. Your previous cited example was perfect in that regard.

kovidgoyal · 04-18-2014, 12:04 PM

I have added an example.

Regarding the show next occurrence not working, I have found a bug that was causing it. Unfortunately, fixing it is not as simple as I hoped, I need to think about it a little.

mrmikel · 04-18-2014, 12:21 PM

As it happens, if you press control c when on the misspelled word, it does copy, and can be pasted to the find area of search and located that way, so it is not critical.

DrChiper · 04-18-2014, 12:30 PM

Interestingly, sometimes the same assumed "typo" is only caught once, like in this example:

Code:

            <div class="images">
               <img src="images/drawing1.jpg" alt="Image drawing1.jpg: Illustration: Triangles (Fig. 1 - 3)" class="calibre21"/>
               <p class="image-caption">Fig. 1, 2 and 3.</p>
            </div>

in which the first occurring "drawing1" is caught, but the 2nd not (at least it is not counted in the list). It looks like the preceding character might have something to do with this behavior.

Divingduck · 04-18-2014, 12:31 PM

Kovid,
Yes, your example was a kind eye opener for me. I didn’t saw this way of switching a language within a text any time before.

mrmikel,
Thanks for your explanations to Kovid. It was exactly what I thought.

kovidgoyal · 04-18-2014, 12:42 PM

@mrmikel: Turns out, it was easy to fix after all: https://github.com/kovidgoyal/calibr...a6ef8453c57139

@DrChiper: calibre does not check the src attribute, since spelling is meaningless htere. The title and alt attributes are checked, however.

04-18-2014, 10:11 AM	#18
roger64 Wizard Posts: 2,624 Karma: 3120635 Join Date: Jan 2009 Device: Kindle PW3 (wifi)	Hi This was the last big item left to do. How amazing this huge task has been completed in so few months! Congratulations and thank you. French Here is a link for a good French dictionnary in oxt format. Of course it's hunspell all along. It's used by Grammalecte, a grammatical extension to OpenOffice and LibreOffice. It is updated quite frequently. The 5.0.2 version is from January 2014. It works pretty well with the Editor. http://www.dicollecte.org/grammalecte/telecharger.php Elided forms Take a word: articulateur. You also can find: l'articulateur. The editor will display it two times, in two separate places once as "articulateur", the other as "l'articulateur" because elided forms seem to be counted as independent words (same for d' and others). It is grammatically the same word. GUI Even on a full page window, I can, at most visualize 20 words. I should at least be able to see a healthy 50 and it seems really necessary to be able to narrow the fields. Several languages together What exactly should I write and where if I have a book in French with lots of English words inside? I start from here. <html xmlns="http://www.w3.org/1999/xhtml" xml:lang="fr-FR"> Last edited by roger64; 04-18-2014 at 12:48 PM.

04-18-2014, 12:30 PM	#28
DrChiper Bookish Posts: 1,008 Karma: 2003162 Join Date: Jun 2011 Device: PC, t1, t2, t3, Clara BW, Clara HD, Libra 2, Libra Color, Nxtpaper 11	Interestingly, sometimes the same assumed "typo" is only caught once, like in this example: Code: <div class="images"> <img src="images/drawing1.jpg" alt="Image drawing1.jpg: Illustration: Triangles (Fig. 1 - 3)" class="calibre21"/> <p class="image-caption">Fig. 1, 2 and 3.</p> </div> in which the first occurring "drawing1" is caught, but the 2nd not (at least it is not counted in the list). It looks like the preceding character might have something to do with this behavior.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Spelling anomalies	DMB	General Discussions	71	06-19-2012 07:55 AM
Are DRM books with check in/check out allowed?	i8abug	Library Management	4	05-31-2012 02:27 PM
Spelling errors and such	starrlamia	General Discussions	29	11-29-2010 03:59 AM
Seriously thoughtful Spelling contractions	SameOldStory	Lounge	47	09-08-2010 09:08 PM
Spelling Macro	PieOPah	Workshop	36	12-13-2008 02:27 AM

04-18-2014, 09:28 AM	#17
kovidgoyal creator of calibre Posts: 45,219 Karma: 27110894 Join Date: Oct 2006 Location: Mumbai, India Device: Various	@DrChiper: The spell check most definitely is aware of tags. In fact it will only check text in tags (other than script and style) and in the attributes alt and title (which are used for tooltips/descriptions). If you are getting words from places other than tags, then it is likely a bug, attach the file. @theducks: That's what the user dictionaries are for. @BR: I suggest setting up a user dictionary with your abbreviations. The correct solution for not flagging some words is not to simply ignore an entire file.

04-18-2014, 10:56 AM	#19
mrmikel Color me gone Posts: 2,089 Karma: 1445295 Join Date: Apr 2008 Location: Central Oregon Coast Device: PRS-300	I have been on the lookout for a British-american dictionary, since I work with a lot of British works and also american or works about WWII where both are used at the same time.

04-18-2014, 10:59 AM	#20
kovidgoyal creator of calibre Posts: 45,219 Karma: 27110894 Join Date: Oct 2006 Location: Mumbai, India Device: Various	@roger64: Regarding treatment of elided forms, as I am not a french speaker, it is rather difficult for me to write rules for combining words. At the moment all distinctly spelled words are treated differently. Patches are welcome to change that. If you want multiple languages, you do something like this: <div lang="en">Some english text<span lang="fr">some french words</span></div> Basically when the lang attribute exists on a tag, the contents of that tag are interpreted in the specified language. You can put a lang attribute on any tag, and they can be nested.

04-18-2014, 11:23 AM	#21
roger64 Wizard Posts: 2,624 Karma: 3120635 Join Date: Jan 2009 Device: Kindle PW3 (wifi)	Thanks I will get in touch with the Grammalecte author about elided forms.

04-18-2014, 12:00 PM	#25
mrmikel Color me gone Posts: 2,089 Karma: 1445295 Join Date: Apr 2008 Location: Central Oregon Coast Device: PRS-300	An example might be helpful to those of us who are not so knowledgeable. It is not so common a construction. Your previous cited example was perfect in that regard.

04-18-2014, 12:04 PM	#26
kovidgoyal creator of calibre Posts: 45,219 Karma: 27110894 Join Date: Oct 2006 Location: Mumbai, India Device: Various	I have added an example. Regarding the show next occurrence not working, I have found a bug that was causing it. Unfortunately, fixing it is not as simple as I hoped, I need to think about it a little.

04-18-2014, 12:21 PM	#27
mrmikel Color me gone Posts: 2,089 Karma: 1445295 Join Date: Apr 2008 Location: Central Oregon Coast Device: PRS-300	As it happens, if you press control c when on the misspelled word, it does copy, and can be pasted to the find area of search and located that way, so it is not critical.

04-18-2014, 12:31 PM	#29
Divingduck Wizard Posts: 1,166 Karma: 1410083 Join Date: Nov 2010 Location: Germany Device: Sony PRS-650	Kovid, Yes, your example was a kind eye opener for me. I didn’t saw this way of switching a language within a text any time before. mrmikel, Thanks for your explanations to Kovid. It was exactly what I thought.

04-18-2014, 12:42 PM	#30
kovidgoyal creator of calibre Posts: 45,219 Karma: 27110894 Join Date: Oct 2006 Location: Mumbai, India Device: Various	@mrmikel: Turns out, it was easy to fix after all: https://github.com/kovidgoyal/calibr...a6ef8453c57139 @DrChiper: calibre does not check the src attribute, since spelling is meaningless htere. The title and alt attributes are checked, however.