Bug 1490 - First part of Ani+Ani/Anipart-compound should be akk/gen
Summary: First part of Ani+Ani/Anipart-compound should be akk/gen
Status: ASSIGNED
Alias: None
Product: sme lexicon
Classification: Unclassified
Component: Continuation lexica (show other bugs)
Version: unspecified
Hardware: All All
: P5 - Later enhancement
Assignee: Thomas Omma
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2012-10-31 09:43 CET by Børre Gaup
Modified: 2018-05-29 10:55 CEST (History)
5 users (show)

See Also:

Attachments
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Børre Gaup 2012-10-31 09:43:40 CET 
Both usme and usmeNorm accept the compound luossačuopma, while the correct form is luosačuopma.

For compounds consisting of Ani+Ani/Anipart/Food we should only accept Akk/Gen of the first part of the compound.

~ $ usme
0%>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>100%
luosačuopma
luosačuopma     luossa+Ani+N+SgGenCmp+Cmp#čuopma+N+Sg+Nom
luosačuopma     luosačuopma+N+Sg+Nom

luossačuopma
luossačuopma    luossa+Ani+N+SgNomCmp+Cmp#čuopma+N+Sg+Nom

~ $ usmeNorm
0%>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> 100%
luossačuopma
luossačuopma    luossa+Ani+N+SgNomCmp+Cmp#čuopma+N+Sg+Nom

luosačuopma
luosačuopma     luossa+Ani+N+SgGenCmp+Cmp#čuopma+N+Sg+Nom
luosačuopma     luosačuopma+N+Sg+Nom
Comment 1 Thomas Omma 2012-10-31 09:47:23 CET 
I think this is a great way to make use of the semantic tags for the speller. We have already got +Ani and +Food tags, we could introduce +AniPart
Comment 2 Børre Gaup 2012-10-31 11:34:43 CET 
(In reply to comment #1)
> I think this is a great way to make use of the semantic tags for the speller.
> We have already got +Ani and +Food tags, we could introduce +AniPart

Who should write the rule?
Where should they be written?
Comment 3 Thomas Omma 2014-11-07 11:24:38 CET 
add Linda, one two three
Comment 4 Thomas Omma 2014-11-18 11:11:39 CET 
hello, this is the bug that i was talking about in uppsala
at the beginning i thought it was a thing for the grammarchecker, but if it is possible i rather have it in speller

this should always be in SgGen:
boazu+CmpN/SgN+CmpN/SgG+CmpN/PlG+Sem/Ani:boah'cu BOAZU "reindeer N" ;

when one of these is in second part:
juolgi+Sem/Body:juol'gi AIGI ;
juolut+CmpN/SgN+CmpN/SgG+CmpN/DefPlGen+Sem/Plant:juoluh GAHPIRLONGSHORT ;
márfi+Sem/Food:már'fi GOAHTI-I ;


Compoundtagging?  Like CmpN/AniSgG

Or flags?
Comment 5 Thomas Omma 2015-06-02 10:00:09 CEST 
this is an amazing bug.

CmpN/LeftAniSgG

This is how we do it!?

boazu + juolgi

When "joulgi" has the tag CmpN/LeftAniSgG, the speller rules out *boazojuolgi, and only suggest bohccojuolgi
Comment 6 Thomas Omma 2015-06-02 10:00:55 CEST 
Sjur, what is the status of compound-tags in hfst-speller?
Comment 7 Thomas Omma 2015-06-02 10:10:33 CEST 
(In reply to comment #6)
> Sjur, what is the status of compound-tags in hfst-speller?


ok, I found this:

So — the only reasonable way to handle this is by using flag diacritics. But
these tags are not flags, and can't be turned into flags either (that would
break the PLX conversion).

What is needed is a flag diacritic system parallel to the existing tag system,
implementing the same semantics that way. It will NOT be pretty, and we need to
devote some time to it to get it right. But it is the only practical way to
solve this that I can see.
Comment 8 Sjur Nørstebø Moshagen 2015-09-21 10:38:24 CEST 
No need to have Biret Ánne, Berit Merete, Inga and Ritva on the CC list anymore. Added Sandra.