I'm working on it. The problem in the disambiguation of "una" in Spanish (the indefinite article!), that is in a component of Apertium that is common with the very used apertium-spa-cat (that's why I want to test well the patch in order not to break anything).
My problem is: can a new version of apertium-es-gl be done when this would be fixed? Missatge de Francis Tyers <fty...@prompsit.com> del dia dc., 17 de jul. 2019 a les 14:55: > > > -------- Mensaje Original -------- > Asunto: Re: Still problems with "una" > Fecha: 2019-07-17 11:43 > De: Elisardo <elisard...@gmail.com> > Destinatario: Francis Tyers <fty...@prompsit.com> > > Sorry for bother you, but problem persists... :( > > I translated "Una casa en la playa. Una persona en el mar." and the > result was "Una casa na praia. Una persoa no mar." The correct > translation is "Unha casa na praia. Unha persoa no mar." > > Additionally I tested all this sentences: > una vez. dar una caminata. es la una. dar una vuelta. sacar una buena > nota. ver una película. hacer una pregunta. a la una. tengo una > pregunta. echar una carta. hay una. había una vez. dar una fiesta. una > vez allí. es una pena. tener una cita. tomar una ducha. es la una y > media. escribir una carta. poner una multa > > And sentences with wrong translation were "ter una cita" and "tomar una > ducha". > > With this test I think that the problem should be in rule > *uno<num><f><sp>, in feminine it should be translated to "unha". > > Could you fix it, please? > > Saúdos, > Elisardo > > O sáb., 13/07/2019 ás 07:09, Francis Tyers (<fty...@prompsit.com>) > escribiu: > > > El 2019-07-12 16:43, Elisardo escribió: > >> Ok, I think that it's correct, "una" always should be translated to > >> "unha". > >> > >> Where are the changes to tell it to Tino? > >> > >> Bye, > >> Elisardo > >> > >> O ven., 12/07/2019 ás 13:14, Francis Tyers (<fty...@prompsit.com>) > >> escribiu: > >> > >>> El 2019-07-12 09:55, Elisardo escribió: > >>>> Francis the problem with "una" still persists :( > >>>> > >>>> I tryed to translate: "Voy para casa y llevo una caja. Para los > >>>> elementos de esta forma no necesitamos una estructura compleja." > >>> and > >>>> the result was "Vou para casa e levo unha caixa. Para os elementos > >>>> desta forma non necesitamos una estrutura complexa.", the last > >>> "una" > >>>> persist untranslated... > >>>> > >>>> Tino told me the rules about "una", could you help me to review > >>> it? > >>>> I'm not linguist, so I don't know all the details of this > >>> problem... > >>>> > >>>> He told me that the rules asociated to "una" are: > >>>> *uno<num><f><sp> > >>>> *uno<prn><tn><f><sg> > >>>> > >>>> *uno<det><ind><f><sg> > >>>> > >>>> *unir<vblex><prs><p3><sg> > >>>> > >>>> *unir<vblex><prs><p1><sg> > >>>> > >>>> *unir<vblex><imp><p3><sg> > >>>> > >>>> I understand that uno<num> is the numeral uno, and the femenine > >>> in > >>>> galician is always "unha" > >>>> > >>>> I understand that uno<prn> is the pronoun, and the femenine in > >>>> galician is always "unha" > >>>> > >>>> I understand that uno<det> is the determinant, and the feminine in > >>>> galician is always "unha" > >>>> > >>>> The other rules about "unir" are a verb, the form "una", present > >>> of > >>>> subjunctive, is the same in galician, but the rules shouldn't mix > >>> up a > >>>> verb with the other forms... > >>>> > >>>> Could you help me with that? > >>>> > >>> > >>> Hi Elisardo, > >>> > >>> Tino is right that those are the possible analyses of "una" in > >>> Spanish. > >>> > >>> The problems are when the tagger is choosing a verb reading instead > >>> of > >>> one of the other readings. > >>> > >>> Here is the sentence you gave: > >>> > >>> "<Para>" > >>> "para" pr > >>> ; "parar" vblex pri p3 sg REMOVE:321:r_para_2 > >>> ; "parar" vblex imp p2 sg REMOVE:321:r_para_2 > >>> ; "parir" vblex prs p3 sg REMOVE:321:r_para_2 > >>> ; "parir" vblex prs p1 sg REMOVE:321:r_para_2 > >>> ; "parir" vblex imp p3 sg REMOVE:321:r_para_2 > >>> "<los>" > >>> "el" det def m pl > >>> "lo" prn pro p3 m pl > >>> "<elementos>" > >>> "elemento" n m pl > >>> "<de>" > >>> "de" pr > >>> "<esta>" > >>> "este" prn tn f sg > >>> "este" det dem f sg > >>> "<forma>" > >>> "forma" n f sg > >>> ; "formar" vblex pri p3 sg REMOVE:255:r_verb_1 > >>> ; "formar" vblex imp p2 sg REMOVE:255:r_verb_1 > >>> "<no>" > >>> "no" adv > >>> "no" ij > >>> "<necesitamos>" > >>> "necesitar" vblex pri p1 pl > >>> ; "necesitar" vblex ifi p1 pl REMOVE:1189:presente_pasado0 > >>> "<una>" > >>> "uno" num f sp > >>> "uno" prn tn f sg > >>> "uno" det ind f sg > >>> "unir" vblex prs p3 sg > >>> "unir" vblex prs p1 sg > >>> "unir" vblex imp p3 sg > >>> "<estructura>" > >>> "estructura" n f sg > >>> "estructurar" vblex pri p3 sg > >>> ; "estructurar" vblex imp p2 sg REMOVE:426:imp_1 > >>> "<compleja>" > >>> "complejo" adj f sg > >>> "<.>" > >>> "." sent > >>> > >>> Notice that "una" is undisambiguated. > >>> > >>> We have a rule that says: > >>> > >>> "<una>" REMOVE Vblex IF (-1C Fin) (1C N) ; > >>> > >>> = For the word "una" remove the verb reading if it is preceeded by > > a > >>> > >>> finite verb and followed by something that can only be a noun. > >>> > >>> If I change it to: > >>> > >>> "<una>" REMOVE Vblex IF (-1C Fin) (1C N OR Fin) ; > >>> > >>> Allowing the following word to be a finite-verb too, then it works. > >>> > >>> $ echo "Voy para casa y llevo una caja. Para los elementos de esta > >>> forma > >>> no necesitamos una estructura compleja." | apertium -d . es-gl > >>> Vou para casa e levo unha caixa. Para os elementos desta forma non > >>> necesitamos unha estrutura complexa. > >>> > >>> I've pushed the changes. > >>> > >>> F. > > > > > > There is a persistent problem with 'una' in the Spanish tagger. It keeps > on getting selected as the imperative or subjunctive of 'unir' > > > ^una/uno<num><f><sp>/uno<prn><tn><f><sg>/uno<det><ind><f><sg>/unir<vblex><prs><p3><sg>/unir<vblex><prs><p1><sg>/unir<vblex><imp><p3><sg>$ > > I don't have time to work on the Spanish CG, so what do people think > about > just removing those two analyses without context? > > E.g. > > Una = "<una>" ; > > REMOVE (0 Unir) (0 Una) ; > > If there are no complains I will implement this. > > Fran > > > _______________________________________________ > Apertium-stuff mailing list > Apertium-stuff@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/apertium-stuff >
_______________________________________________ Apertium-stuff mailing list Apertium-stuff@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/apertium-stuff