search

giellalt / bugzilla-dummy

0 stars 0 forks source link

sme: Flytting av G3 til N+G3 (Bugzilla Bug 1238) #1337

Closed albbas closed 12 years ago

albbas commented 12 years ago

This issue was created automatically with bugzilla2github

Bugzilla Bug 1238

Date: 2011-12-30T14:04:35+01:00 From: Lene Antonsen <> To: Sjur Nørstebø Moshagen <> CC: sjur.n.moshagen, tomi.k.pieski

Last updated: 2012-02-23T12:34:54+01:00

albbas commented 12 years ago

Comment 5518

Date: 2011-12-30 14:04:35 +0100 From: Lene Antonsen <>

Analyser med G3 bør ha N+G3 og ikke motsatt. Dette er delvis gjort i common/src/usage-tags-remove.regex men fungerer ikke for lemmaer som er merka med ektra tagger, f.eks. semantiske tagger: vuorru vuorru+N+G3+Sg+Nom men oarri oarri+G3+Ani+N+Sg+Nom

fordi denne har en ekstra tag mellom lemma og N: oarri+G3+CmpN/SgN+CmpN/SgG+CmpN/PlG+Ani:oarri¤ ALBMI ;

Dette er veldig viktig for generering for Oahpa, og at ordklasse skal bli gjenkjent i oppgavene. Men dette gjør det også enklere å generer til dict osv

albbas commented 12 years ago

Comment 5612

Date: 2012-01-16 16:44:18 +0100 From: Lene Antonsen <>

Minner om denne bugen, dette er viktig for Oahpa. Nå mister vi en god del lemmaer til Morf-programmane. Hvis den ikke kan fikses på en elegant måte, så fikser jeg den bit for bit.

albbas commented 12 years ago

Comment 5613

Date: 2012-01-16 23:49:42 +0100 From: Trond Trosterud <>

Eg trudde denne vart fiksa då vi såg på dette sist, bra du minner meg på det

albbas commented 12 years ago

Comment 5692

Date: 2012-01-31 09:09:05 +0100 From: Lene Antonsen <>

Det er fremdeles 14 substantiver som ikke blir generert til Oahpa pga av G3+N.

albbas commented 12 years ago

Comment 5707

Date: 2012-02-03 15:49:51 +0100 From: Sjur Nørstebø Moshagen <>

No er buggen fiksa i svn r53576:

Log: Added support for reordering +N after +G3, with an arbitrary number of symbols in between. Now all cases of +G3 … +N are reordered to +N+G3, for all transducers (it is done as part of the lexc compilation).

albbas commented 12 years ago

Comment 5827

Date: 2012-02-23 09:58:42 +0100 From: Lene Antonsen <>

usme:

láđđegákti láđđegákti láđđe#gákti+N+Sg+Nom láđđegákti láđđi+N+SgNomCmp+Cmp#gákti+N+Sg+Nom láđđegákti láđđi+N+G3+SgNomCmp+Cmp#gákti+Sg+Nom <======

Her forsvinner N etter Cmp, noe som ikke skal skje.

albbas commented 12 years ago

Comment 5831

Date: 2012-02-23 11:46:51 +0100 From: Sjur Nørstebø Moshagen <>

Fixed in svn r54366:

$ usme láđđegákti láđđegákti láđđe#gákti+N+Sg+Nom láđđegákti láđđi+N+SgNomCmp+Cmp#gákti+N+Sg+Nom láđđegákti láđđi+N+G3+SgNomCmp+Cmp#gákti+N+Sg+Nom

Log: Added a restriction on the G3 tag moving thing to ensure that moving does not extend across word boundaries.

albbas commented 12 years ago

Comment 5835

Date: 2012-02-23 12:09:24 +0100 From: Tomi Pieski <>

I saw the change in makefile and I was also working with this one. I would suggest to change the left context into:

%+G3 \%#* _

This addresses the complement of word boundary mark. Both solutions work though..

albbas commented 12 years ago

Comment 5837

Date: 2012-02-23 12:34:54 +0100 From: Sjur Nørstebø Moshagen <>

(In reply to comment #7)

I saw the change in makefile and I was also working with this one. I would suggest to change the left context into:

%+G3 \%#* _

This addresses the complement of word boundary mark. Both solutions work though..

Sorry for the confusion. Whatever you think is more elegant is fine with me:)