Bug 106640 - Improve the Hunspell engine
Summary: Improve the Hunspell engine
Status: RESOLVED NOTOURBUG
Alias: None
Product: LibreOffice
Classification: Unclassified
Component: Linguistic (show other bugs)
Version:
(earliest affected)
5.4.0.0.alpha0+
Hardware: All All
: medium enhancement
Assignee: Not Assigned
URL:
Whiteboard:
Keywords:
Depends on:
Blocks: Dictionaries
  Show dependency treegraph
 
Reported: 2017-03-19 10:43 UTC by Marco A.G.Pinto
Modified: 2017-06-26 10:05 UTC (History)
5 users (show)

See Also:
Crash report or crash signature:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Marco A.G.Pinto 2017-03-19 10:43:20 UTC
Hello Németh,

Could you improve the Hunspell engine to make it more accurate just like in M$ Office?

I have this idea.

Imagine someone types a words with an hyphen and the word is not in the speller.

What happens currently is that if the word before and after the hyphen are in the speller, no typo is reported.

I would like an enhancement that would work like this:

Imagine I typed:
on-line

If Hunspell didn't find this exact match, it would search for the word without the hyphen.

If found without the hyphen, "online", it would mark the current word as a typo and suggest the one in the speller.

This is just a small example to explain because "on-line" and "online" are both in the spellers.

Thank you!
Comment 1 Marco A.G.Pinto 2017-03-19 11:08:13 UTC
[11:03] <MarkHung> marcoagpinto: regarding tdf#106640, if you're reporting that issue, at least you should highlight the difference between LibreOffice and MS Office, and state which part do you want.

[11:04] <MarkHung> Otherwise it is very likely to be a invalid bug.

[11:05] <marcoagpinto> What I mean is "abc-def", if the words "abc" and "def" exist in the speller, Hunspell won't mark as a typo, but if it doesn't find "abc-def" and finds "abcdef" would suggest "abcdef"

[11:06] <marcoagpinto> tml__: that is how M$ Word works

So, M$ Word is better in this case
Comment 2 Marco A.G.Pinto 2017-03-19 11:10:52 UTC
[11:07] <MarkHung> Search without hyphen if the target with hyphen is concrete but "more accurate just like in M$ Office" is vague.

[11:10] <marcoagpinto> MarkHung: what I mean is that if no word with hyphen is found in the speller, but it checks without the hyphen, and if it finds, it will suggest without the hyphen
Comment 3 Xisco Faulí 2017-03-19 11:20:27 UTC
Hello Marco,

Huspell is an independent project and its bugs should be reported here: http://hunspell.github.io/

Regards
Comment 4 László Németh 2017-06-26 09:56:36 UTC
Hi Marco,

Conditional checking of the words with hyphens is a good idea. Thanks for the suggestion, I'll check it.

By the way, you can forbid  "on-line" with the recent Hunspell, too, if needed, using

BREAK 0

in the affix file, see tests/breakoff* in the Hunspell source distribution. But this will forbid all non-dictionary words with hyphens.

Best regards,
Laszlo
Comment 5 Marco A.G.Pinto 2017-06-26 10:05:11 UTC
(In reply to László Németh from comment #4)
> Best regards,
> Laszlo

Németh!

Thanks for replying!

Kind regards,