Using more than one dictionary at the same time

[mcdummy]

As far as I understand the dictionary handling of Sigil, it is possible to install dictionaries for several languages, but only one of them is used for checking an ebook.

Since I have have several ebooks that include passages in several languages, it would be helpful if the spell checking and highlighting of incorrect words in the editor would be able to switch languages within the text using, for instance, the lang-tag.

Regards,

McDummy

[BetterRed]

The calibre editor has support for spell checking multi-lingual text.

Someone (not Kevin or Doug), attempted add multi-lingual spell checking to Sigil but it never saw the light of day. I spell check in the calibre-editor, but do other editing in Sigil.

BR

[BeckyEbook]

Archival RoadMap: https://code.google.com/archive/p/si...s/RoadMap.wiki
Issue: https://github.com/Sigil-Ebook/Sigil/issues/218

This is not a trivial task. Patience is a virtue.

[KevinH]

The issue is *not* doing a final spellcheck in multiple languages if the xml:lang or lang attribute is properly used using a dedicated spellcheck dialog. That is doable.

The issue is showing on the fly spelling mistakes (with red squiggly underlines) as you are editing the code itself in CodeView. The existence of potentially incomplete or broken code and the need to walk the tree back up the parent path to determine the language on the fly makes things hard to determine what language to check the just completed word in (and do it quickly).

If people could just live with looking up the word on the fly in *all* used languages during editing in CV but when using the Spell Check dialog, being able to choose the specific language to spellcheck in, that would greatly simplify things.

[BetterRed]

Kevin, I'd prefer to only have red squiggly's under the primary language errors, and a sortable language column in spell checker.

A spell checker bonus would be a list selector of languages found in the book (default would be 'All') so that I could focus on a particular language.

BR

[DNSB]

And if there was some way to force the authors to wrap foreign language bits properly so Sigil wouldn't be guessing what dictionary to use. Is it that hard to wrap your foreign language quotes in <span lang="fr">À chacun son goût</span> or <span class="el">Ημετέρα φύσις αι καθ' έξιν πράξεις εστί. Αριστεία ουν έξις (εστί)</span>?

[BetterRed]

I wouldn't expect Sigil to guess the language if its not marked up, but if it could, quelle surprise, it should be an settable option

BR

[Doitsu]

Originally Posted by mcdummy

Since I have have several ebooks that include passages in several languages, it would be helpful if the spell checking and highlighting of incorrect words in the editor would be able to switch languages within the text using, for instance, the lang-tag.

Many years ago, when varlog worked on implementing multi-language spell checking, I created a simple proof-of-concept validation plugin that checks foreign language segments wrapped in <span> tags with lang/xml:lang attributes.
It works like the EPUBCheck validation plugin. I.e., it'll display potential spelling errors in the validation window. (It won't add red wavy lines!)
If you're interested in testing it, PM me, and I'll send you a download link.

[BetterRed]

@Doitsu - so what happened to varlog's work. I did a little bit of testing for him, but I lost track - might have been when my old Linux laptop bit the dust.

BR

[Doitsu]

Originally Posted by BetterRed

@Doitsu - so what happened to varlog's work. I did a little bit of testing for him, but I lost track - might have been when my old Linux laptop bit the dust.

If I understand his development thread correctly, varlog didn't achieve his goal:

Originally Posted by varlog

Kevin: forget it. It is obscure feature: even the initiator wasn't motivated enough to check it out. I haven't achieved my original goal but had my fun anyway. Concentrate on epub 3. It is way more important.

Since the code is still available in the mlspell branch of the Sigil Github repository, maybe another developer will give it another try in the future.

[KevinH]

FWIW, This will be my next project sometime after our next release. But real time during editing (ie. red squiggly) will be either all languages or main language limited. I am thinking, the main spellcheck dialogue will be the same as now but with a filter by language pull down added.

Something along those lines ...

[mcdummy]

Originally Posted by KevinH

FWIW, This will be my next project sometime after our next release. But real time during editing (ie. red squiggly) will be either all languages or main language limited. I am thinking, the main spellcheck dialogue will be the same as now but with a filter by language pull down added.

Something along those lines ...

The following might not apply to sigil (so you can forget the rest), but I noticed in calbire that a problem could occur when a text in a first language (say lang1) contains passages in a second language (lang2) and the spell ckecker offers the options to jump to incorrectly spelled words.

For instance, assume that a word (e.g., "word-example") is spelled correctly in the second language, but not in the first language, and appears in the following manner:

<p lang="lang1">... <span lang="lang2">word-example</span> ... word-example ... </p>

When a simple search is applied which only uses "word-example" as search item to jump to the incorrectly spelled word, this search algorithm would also jump to the <span lang="lang"> instance, in which it is correctly spelled.

[KevinH]

Yes, it should be rare that the same word would appear in two-different languages side by side.
What most people do not recognize is that once parsed, the text sits at the end of the node chain and so search must be used. Character offsets to specify a file position are simply messed up as soon as a replacement or change is performed.

Even a epub3 cfi needs and uses search/pattern matching to identify text segments.

[mcdummy]

Originally Posted by KevinH

Yes, it should be rare that the same word would appear in two-different languages side by side.

That's actually more common that one might think, though the more general situation is

<p lang="lang1">... word-example ...</p>

...

<p lang="lang1">... <span lang="lang2">word-example</span> ...</p>

I've read many ebooks that are written in a first language, but contain terms or citations from other languages. If the creator of the ebook does not mark all instances of a word in the proper language, you end up with this situation.

Since my ebook reader applies hyphenation, reading the text gets quite annyoing, if words are not assigned to their proper language.

[Tex2002ans]

Originally Posted by mcdummy

That's actually more common that one might think,

Overlaps definitely occur more often than expected.

Latest example I've been using is:

"die" (German) = "the" (English)
"die" (English) = "sterben" (German)

It also has enormous overlap when mixing en-US (American) + en-GB (British) within the same book. (One of the latest books I converted was collection of 20 articles, about half and half.)

Originally Posted by mcdummy

Since my ebook reader applies hyphenation, reading the text gets quite annyoing, if words are not assigned to their proper language.

And which ereader are you using that applies proper hyphenation?

Does it work at the per-word level too? Or only works on a per-book's-language level?