Bug 99999 - Thesaurus in LO handling of antonyms dependent on WordNet 2.1 tagging used in Hunspell and mythes bundled externals, need better data (comment 2)
Summary: Thesaurus in LO handling of antonyms dependent on WordNet 2.1 tagging used in...
Status: NEW
Alias: None
Product: LibreOffice
Classification: Unclassified
Component: Linguistic (show other bugs)
Version:
(earliest affected)
Inherited From OOo
Hardware: All All
: medium enhancement
Assignee: Not Assigned
URL:
Whiteboard:
Keywords:
: 56565 99683 (view as bug list)
Depends on:
Blocks: Thesaurus
  Show dependency treegraph
 
Reported: 2016-05-23 04:16 UTC by Aron Budea
Modified: 2018-10-24 20:45 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 Aron Budea 2016-05-23 04:16:38 UTC
In English (USA) language, type the word "normally", and see right click->Synonyms, in particular: "remarkably" and "unremarkably", which are opposites.

Checking thesaurus via right click->Synonyms->Thesaurus reveals that only "unremarkably" is a synonym, "remarkably" is an antonym.
So the information is there, but antonyms aren't shown separately, they're given as synonyms.
Comment 1 Aron Budea 2016-05-24 21:39:27 UTC
To be precise, the bug is that synonyms show antonyms as well, showing antonyms separately would be an enhancement.

Technically a solution would be to rid the thesaurus of antonyms, but that is far from elegant. If there's a fix, other languages need to be considered as well, I don't know how antonyms are listed in other thesauri.
Comment 2 V Stuart Foote 2016-05-25 01:29:41 UTC
Believe this is a simple mistagging issue with the original WordNet 2.1 definitions file, th_en_US_new.dat from 2005, used for the HunSpell projects thesaurus MyThes--so not really a bug, just incorrectly formatted data.

We then tweak that listing a bit and end up with:
https://cgit.freedesktop.org/libreoffice/dictionaries/tree/en/th_en_US_v2.dat

In a LibreOffice Writer session , if you look at "normal" rather than "normally", where the entries are correctly tagged, you'll see what is supposed to be shown. Antonyms are not displayed on the context menu. And shifting into the Thesaurus with <ctrl>+F7 you'll see the full record including the type annotation.

As to adding a separate LibreOffice context menu for antonyms--I don't think that would be of much value as there are just too few formally defined to make it worth while.

I believe coverage in WordNet 3/3.1 is better, but looks like the data format changed quite a bit so the processing scripts (Daniel N.'s wn2ooo) used for MyThes would have to be refactored. Also, don't know if the licensing allows needed MPL redistribution for use with HunSpell or Language Tools project(s).

=-ref-=
http://hunspell.github.io/
https://cgit.freedesktop.org/libreoffice/dictionaries/tree/en
http://opengrok.libreoffice.org/xref/core/external/mythes/
Comment 3 Stanislav Horacek 2016-07-11 08:26:40 UTC
*** Bug 99683 has been marked as a duplicate of this bug. ***
Comment 4 Stanislav Horacek 2016-07-11 08:27:05 UTC
*** Bug 56565 has been marked as a duplicate of this bug. ***
Comment 5 Yousuf Philips (jay) (retired) 2016-07-12 00:43:02 UTC
Stanislav: I maybe confused but we dont close old bugs as duplicates of new bugs.
Comment 6 V Stuart Foote 2016-07-12 01:22:41 UTC
(In reply to Yousuf (Jay) Philips from comment #5)
> Stanislav: I maybe confused but we dont close old bugs as duplicates of new
> bugs.

Yeah, I was leaning toward saying the same, but then I realized that I'd provided the correct details regards the WordNet 2.1 shortcomings on this issue. The topic is better handled and on point here.  Stanislav, has it right and the others are correctly dup'ed to this.