; This routine converts from 'morfeusz' dialect of NKJP tagset (more genders) ; into the 'proper' dialect of KIPI tagset (less genders). ; The conversion is rought and some important details may be missing. [convert] tagset_from=sgjp tagset_to=kipi check=ignore ; make the gender set smaller as defined in the "two tagsets" paper override=n1:n override=n2:n override=n3:n override=p1:m1 override=p2:n override=p3:n ; cast the new indeclinable stuff as qubs override=interj:qub override=burk:qub ; reduce brev:any to xxx, can't do any better override=brev:xxx ; cast complementisers as conjunctions override=comp:conj ; should be adj:sg:nom:m1.m2.m3:pos override=adjc:adj ; numcol as num? optionally ; override=numcol:num ; degree value override=com:comp late-check=true [tag] ; cast adverbs not specified for degree as qubs pre=adv:deg post=qub [tag] ; should be m1.m2.m3 but tag-to-set conversion currently not supported ; (and m1 seems most likely for the adjc class) pre=adj:cas post=adj:sg:nom:m1:pos [remove_duplicates]