moved wikt-specific paradigms to a separate file (for the moment)

src/arabic/ParadigmsAra.gf

@@ -868,6 +868,8 @@ formV : (root : Str) -> VerbForm -> V = \s,f -> case f of {
 param VerbForm =
   FormI | FormII | FormIII | FormIV | FormV | FormVI | FormVII | FormVIII | FormX | FormXI ;
 
+
+{- temporarily moved to wiktionary/MoreAra.gf
 -- paradigms for Wiktionary extraction
 ---- TODO: better usage of information in Wiktionary
 
@@ -894,55 +896,55 @@ oper
   wmkA = overload {
     wmkA : {root : Str} -> A
       = \r -> mkA r.root ;
-    mkA : {masc_sg : Str; fem_pl : Str; root : Str} -> A
+    wmkA : {masc_sg : Str; fem_pl : Str; root : Str} -> A
       = \r -> mkA r.root ;
-    mkA : {masc_sg : Str; fem_sg : Str; fem_pl : Str; root : Str} -> A
+    wmkA : {masc_sg : Str; fem_sg : Str; fem_pl : Str; root : Str} -> A
       = \r -> mkA r.root ;
-    mkA : {masc_sg, fem_sg, masc_pl, fem_pl, root, sg_patt, pl_patt : Str} -> A
+    wmkA : {masc_sg, fem_sg, masc_pl, fem_pl, root, sg_patt, pl_patt : Str} -> A
       = \r -> mkA r.root r.sg_patt r.pl_patt ;
-    mkA : {masc_sg, fem_sg, masc_pl, root, sg_patt, pl_patt : Str} -> A
+    wmkA : {masc_sg, fem_sg, masc_pl, root, sg_patt, pl_patt : Str} -> A
       = \r -> mkA r.root r.sg_patt r.pl_patt ;
-    mkA : {fem_pl : Str; fem_sg : Str; masc_sg : Str; root : Str; sg_patt : Str} -> A
+    wmkA : {fem_pl : Str; fem_sg : Str; masc_sg : Str; root : Str; sg_patt : Str} -> A
       = \r -> mkA r.root r.sg_patt ;
-    mkA : {fem_pl : Str; fem_sg : Str; masc_sg, masc_pl, root, sg_patt : Str} -> A
+    wmkA : {fem_pl : Str; fem_sg : Str; masc_sg, masc_pl, root, sg_patt : Str} -> A
       = \r -> mkA r.root r.sg_patt ;
-    mkA : {masc_sg, root, sg_patt : Str} -> A
+    wmkA : {masc_sg, root, sg_patt : Str} -> A
       = \r -> mkA r.root r.sg_patt ;
-    mkA : {masc_sg, masc_pl, root, sg_patt : Str} -> A
+    wmkA : {masc_sg, masc_pl, root, sg_patt : Str} -> A
       = \r -> mkA r.root r.sg_patt ;
-    mkA : {masc_sg, fem_sg, masc_pl, fem_pl, root, pl_patt : Str} -> A
+    wmkA : {masc_sg, fem_sg, masc_pl, fem_pl, root, pl_patt : Str} -> A
       = \r -> mkA r.root ; ----
-    mkA : {masc_sg, fem_sg, masc_pl, fem_pl, root : Str} -> A
+    wmkA : {masc_sg, fem_sg, masc_pl, fem_pl, root : Str} -> A
       = \r -> mkA r.root ; ----
-    mkA : {masc_sg, fem_sg, root : Str} -> A
+    wmkA : {masc_sg, fem_sg, root : Str} -> A
       = \r -> mkA r.root ; ----
-    mkA : {masc_sg, fem_sg, masc_pl, fem_pl, pl_patt : Str} -> A
+    wmkA : {masc_sg, fem_sg, masc_pl, fem_pl, pl_patt : Str} -> A
       = \r -> mkA r.masc_sg ; ----
-    mkA : {masc_sg : Str; fem_sg : Str; fem_pl : Str} -> A
+    wmkA : {masc_sg : Str; fem_sg : Str; fem_pl : Str} -> A
       = \r -> mkA r.masc_sg ; ----
-    mkA : {masc_sg : Str; fem_sg : Str; root : Str ; sg_patt : Str} -> A
+    wmkA : {masc_sg : Str; fem_sg : Str; root : Str ; sg_patt : Str} -> A
       = \r -> mkA r.root r.sg_patt ;
-    mkA : {masc_sg : Str; fem_sg : Str} -> A
+    wmkA : {masc_sg : Str; fem_sg : Str} -> A
       = \r -> mkA r.masc_sg ; ----
-    mkA : {masc_sg : Str; masc_pl : Str; fem_sg : Str; fem_pl : Str} -> A
+    wmkA : {masc_sg : Str; masc_pl : Str; fem_sg : Str; fem_pl : Str} -> A
       = \r -> mkA r.masc_sg ; ----
-    mkA : {masc_sg : Str; masc_pl : Str; fem_sg : Str; root : Str} -> A
+    wmkA : {masc_sg : Str; masc_pl : Str; fem_sg : Str; root : Str} -> A
       = \r -> mkA r.root ;
-    mkA : {masc_sg : Str; masc_pl : Str; fem_sg : Str} -> A
+    wmkA : {masc_sg : Str; masc_pl : Str; fem_sg : Str} -> A
       = \r -> mkA r.masc_sg ; ----
-    mkA : {masc_sg : Str; masc_pl : Str; root : Str} -> A
+    wmkA : {masc_sg : Str; masc_pl : Str; root : Str} -> A
       = \r -> mkA r.root ;
-    mkA : {masc_sg : Str; masc_pl, pl_patt : Str; root : Str} -> A
+    wmkA : {masc_sg : Str; masc_pl, pl_patt : Str; root : Str} -> A
       = \r -> mkA r.root ;
-    mkA : {masc_sg : Str; masc_pl, pl_patt, sg_patt : Str; root : Str} -> A
+    wmkA : {masc_sg : Str; masc_pl, pl_patt, sg_patt : Str; root : Str} -> A
       = \r -> mkA r.sg_patt r.pl_patt ;
-    mkA : {masc_sg : Str; masc_pl : Str} -> A
+    wmkA : {masc_sg : Str; masc_pl : Str} -> A
       = \r -> mkA r.masc_sg ; ----
-    mkA : {masc_sg : Str; masc_pl, pl_patt : Str} -> A
+    wmkA : {masc_sg : Str; masc_pl, pl_patt : Str} -> A
       = \r -> mkA r.masc_sg ; ----
-    mkA : {masc_sg : Str; root : Str} -> A
+    wmkA : {masc_sg : Str; root : Str} -> A
       = \r -> mkA r.root ;
-    mkA : {masc_sg : Str} -> A
+    wmkA : {masc_sg : Str} -> A
       = \r -> mkA r.masc_sg ; ----
     } ;
 
@@ -960,5 +962,5 @@ oper
     wmkV : {imperfect : Str} -> V
       = \r -> variants {} ; ---- mkV r.imperfect ;
     } ;
-
+-}
 } ;

src/arabic/wiktionary/Makefile

@@ -4,4 +4,4 @@ all:
 	python3 read_wiktionary.py gf-map >source_of_MorphoDictAra.jsonl 
 	gf -make MorphoDictAra.gf
 	python3 read_wiktionary.py eval-funs >1-eval.txt 
-	python3 to_wordnet.py >wornet-arabic.jsonl
+	python3 to_wordnet.py >wordnet-arabic.jsonl

src/arabic/wiktionary/read_wiktionary.py

@@ -71,6 +71,10 @@ CONCRETE_MODULE = 'MorphoDictAra'
 
 # concrete syntax file, to debug sources of linearizations
 CONCRETE_FILE = CONCRETE_MODULE + '.gf'
+
+# evaluation result file, created with mode eval-funs
+EVAL_FILE = 'eval.jsonl'
+
             
 # read a gzipped jsonl file (one object per line),
 # showing lines where one of a list of languages is present
@@ -93,6 +97,17 @@ if MODE == 'raw':
     get_gzip_json(WIKTIONARY_DUMP, 1, [EXTRACTED_LANGUAGE])
     exit()
 
+    
+if MODE == 'error-analysis':
+    evals = {}
+    with open(EVAL_FILE) as file:
+        for line in file:
+            row = json.loads(line)
+            if labels := row.get('labels', None):
+                verdict = row['verdict']
+                evals[(labels, verdict)] = evals.get((labels, verdict), 0) + 1
+    for labverdict, n in sorted(list(evals.items())):
+        print(labverdict, n)
 
 # https://en.wikipedia.org/wiki/Buckwalter_transliteration
 buckwalter_dict = {
@@ -378,7 +393,7 @@ def find_root(s):
 if MODE == 'gf-abs':
     print('abstract MorphoDictAraAbs = Cat ** {')    
 if MODE == 'gf-cnc':
-    print('concrete MorphoDictAra of MorphoDictAraAbs = CatAra ** open ParadigmsAra in {') 
+    print('concrete MorphoDictAra of MorphoDictAraAbs = CatAra ** open ParadigmsAra, MoreAra in {') 
 
 # go through the Arabic Wiktionary entries
 # generate functions with unique names
@@ -552,9 +567,9 @@ def eval_grammar(pgffile, concretename, mapfile, show=True, verbose=False):
                 totals[cat][rep] = totals[cat].get(rep, 0) + 1 
 
             if show:
-                print(report)
+                print(json.dumps(report, ensure_ascii=False))
 
-    print(totals)
+        print(json.dumps(totals, ensure_ascii=False))
     
 
 if MODE.startswith('eval'):