first version of MorphoDictHrv extracted from wiktionary; TODO better use of PN and V forms

src/croatian/NounHrv.gf

@@ -30,7 +30,7 @@ lin
 
     DefArt = {s = \\_,_,_ => []} ;
     IndefArt = {s = \\_,_,_ => []} ;
-    NumPl = {s = \\_,_ => [] ; size = NS_20_} ; ---- size
+    NumPl = {s = \\_,_ => [] ; size = NS_2_4} ; ---- size
     NumSg = {s = \\_,_ => [] ; size = NS_1} ;
 
     UsePron pron = {

src/croatian/ParadigmsHrv.gf

@@ -13,6 +13,8 @@ oper
     = Masc Anim ;
   mascInanimate : Gender
     = Masc Inanim ;
+  masculine : Gender
+    = Masc Inanim ;
   feminine : Gender
     = Fem ;
   neuter : Gender
@@ -132,12 +134,27 @@ oper
 	  compar = velikA comp ;
 	  superl = superlAForms (velikA comp)
 	  } ;
+    mkA : (posit : AForms) -> (compar : Str) -> A
+      = \posit,compar -> lin A {
+          posit = posit ;
+	  compar = velikA compar ;
+	  superl = superlAForms (velikA compar)
+	  } ;
     mkA : (posit, compar : AForms) -> A
       = \posit,compar -> lin A {
           posit = posit ;
 	  compar = compar ;
 	  superl = superlAForms compar
 	  } ;
+    mkA : (posit : AForms) -> A
+      = \posit ->
+          let
+            compar = regComparAForms posit
+          in lin A {
+            posit = posit ;
+	    compar = compar ;
+	    superl = superlAForms compar
+	    } ;
     } ;
 
   invarA : Str -> A

src/croatian/ResHrv.gf

@@ -338,12 +338,12 @@ voicing : Str -> Str = \s -> case s of {
     msins : Str ;        -- nsins, pdat, ploc, pins = msins
     fsins : Str ;               -- no o/e variation like in msdat
     mpnom : Str ;               -- mpvoc = mpnom
-    pgen : Str ;                --
+    mpgen : Str ;               --
     } ;
 
 invarAdjForms : Str -> AdjForms = \s -> {
     msnom, fsnom, nsnom, msgen, fsgen, msdat,
-    fsdat, fsacc, msins, fsins, mpnom, pgen = s ;
+    fsdat, fsacc, msins, fsins, mpnom, mpgen = s ;
     } ;
 
 -- used in PositA but will also work in Compar and Superl by calling their record fields
@@ -368,7 +368,7 @@ adjFormsAdjective : AdjForms -> Adjective = \afs -> {
       | <Pl,Dat|Loc|Ins, _>      => afs.msins ;
     <Sg, Ins, Fem>               => afs.fsins ;
     <Pl, Nom|Voc, Masc _>        => afs.mpnom ;
-    <Pl, Gen,_>                  => afs.pgen
+    <Pl, Gen,_>                  => afs.mpgen
     }
     } ;
 
@@ -395,7 +395,7 @@ adjFormsAdjective : AdjForms -> Adjective = \afs -> {
       msins   = velk + "im" ;
       fsins   = velk + "om" ;
       mpnom   = velk + "i" ;
-      pgen    = velk + "ih" ;
+      mpgen   = velk + "ih" ;
       } ;
 
   regComparAForms : AdjForms -> AdjForms

src/croatian/wiktionary/extract.py

@@ -40,13 +40,11 @@ ADJ_FORMS = {
          'singular': {
             'nominative': 'msnom',
             'genitive': 'msgen',
-            'dative': 'msdat',
+            'dative': 'msdat'
-            'locative': 'msloc',
-            'instrumental': 'msins'
             },
         'plural': {
             'nominative': 'mpnom',
-            'genitive': 'pgen'
+            'genitive': 'mpgen'
             }
         },
     'feminine': {
@@ -54,7 +52,8 @@ ADJ_FORMS = {
             'nominative': 'fsnom',
             'genitive': 'fsgen',
             'dative': 'fsdat',
-            'accusative': 'fsacc'
+            'accusative': 'fsacc',
+            'instrumental': 'fsins'
             }
         },
     'neuter': {
@@ -122,7 +121,7 @@ def unaccent(word):
 
 cyrillic = 'ЀЈЉЊЋЍЏАБВГДЕЖЗИКЛМНОПРСТУФХЦЧШабвгдежзиклмнопрстуфхцчшыѐђјљњћѝџӣӯ'
 
-def get_forms(pos, forms):
+def get_forms(pos, forms, word):
     dict = {}
     if pos == 'noun':
         for f in forms:
@@ -157,6 +156,9 @@ def get_forms(pos, forms):
                                 for c in ADJ_FORMS[g][n]:
                                     if c in tags:
                                         dict[ADJ_FORMS[g][n][c]] = unaccent(f['form'])
+            elif all([t in tags for t in [
+                  'comparative', 'masculine', 'singular', 'nominative']]):
+                dict['cmsnom'] = unaccent(f['form'])
     elif pos == 'verb':
         for f in forms:
             tags = f.get('tags', [])
@@ -167,6 +169,8 @@ def get_forms(pos, forms):
                             for g in VERB_FORMS[t][n]:
                                if g in tags:
                                    dict[VERB_FORMS[t][n][g]] = unaccent(f['form'])
+        if dict:
+            dict['infin'] = unaccent(word)
 
     else:
         dict['forms'] = forms[:10] ####
@@ -178,7 +182,7 @@ def get_forms(pos, forms):
 
 def lexinfo(data):
     return data['word'], {
-        'pos': data['pos'], 'forms': get_forms(data['pos'], data['forms'])}
+        'pos': data['pos'], 'forms': get_forms(data['pos'], data['forms'], data['word'])}
 
 
 # write morphology of mylang in m.json
@@ -219,7 +223,7 @@ def print_gf_code(data, i):
     cats = {
         'name': ('PN', 7),
         'noun': ('N', 11),
-        'adj': ('A', 12),
+        'adj': ('A', 13),
         'verb': ('V', 12)
         }    
     pos = data[lemma]['pos']
@@ -230,7 +234,10 @@ def print_gf_code(data, i):
         else:
             s = '{'
             for f in fs:
-                s += f + ' = ' + '"' + str(fs[f]) + '"' + ' ; '
+                if f == 'gender':
+                    s += f + ' = P.' + str(fs[f]) + ' ; '
+                else:
+                    s += f + ' = ' + '"' + str(fs[f]) + '"' + ' ; '
             return s[:-3] + '}'  # removing last ;
         
     if pos in cats: