translation

[iramuteq] / corpus.py

diff --git a/corpus.py b/corpus.py
index 8518b25..011206f 100644 (file)
--- a/corpus.py
+++ b/corpus.py
@@ -172,6 +172,16 @@ class Corpus :
         res = self.cformes.execute(query)
         return list(set(list(itertools.chain(*[[int(val) for val in row[0].split()] if not isinstance(row[0], int) else [row[0]] for row in res]))))
 
+    def gettgenst(self, tgen):
+        formesid = ', '.join([`val` for lem in tgen for val in self.lems[lem].formes if lem in self.lems])
+        query = 'SELECT uces FROM uces where id IN (%s) ORDER BY id' % formesid
+        res = self.cformes.execute(query)
+        return list(set(list(itertools.chain(*[[int(val) for val in row[0].split()] if not isinstance(row[0], int) else [row[0]] for row in res]))))
+    
+    def gettgentxt(self, tgen):
+        sts = self.gettgenst(tgen)
+        return list(set([self.getucefromid(val).uci for val in sts]))
+
     def getlemucis(self, lem) :
         uces = self.getlemuces(lem)
         return list(set([self.getucefromid(val).uci for val in uces]))
@@ -391,6 +401,26 @@ class Corpus :
                         tgenoccurrences[t][etoiles[i]] += sum([lemuceeff[uce] for uce in concern])
         return tgenoccurrences, totoccurrences
     
+    def make_tgen_profile(self, tgen, ucecl, uci = False) :
+        log.info('tgen/classes')
+        if uci :
+            tab = [[lem] + [len(set(self.gettgentxt(tgen[lem])).intersection(classe)) for classe in ucecl] for lem in tgen]
+        else :
+            tab = [[lem] + [len(set(self.gettgenst(tgen[lem])).intersection(classe)) for classe in ucecl] for lem in tgen]
+        tab = [[line[0]] + [val for val in line[1:]] for line in tab if sum(line[1:]) >= 3]
+        return tab
+        #i = 0
+        #nam = 'total'
+        #while nam + `i` in tgen :
+        #    i += 1
+        #nam = nam + `i`
+        #last = [nam] + [`len(classe)` for classe in ucecl]
+        #tab += [last]
+        #line0 = ['tgen'] + ['_'.join(['cluster', `i+1`]) for i in range(len(ucecl))]
+        #tab = [line0] + tab
+        #with open(fileout, 'w') as f :
+        #    f.write('\n'.join(['\t'.join(line) for line in tab]).encode(self.parametres['syscoding']))        
+    
     def make_efftype_from_etoiles(self, etoiles) :
         dtype = {}
         etuces = [[] for et in etoiles]
@@ -756,6 +786,28 @@ class Corpus :
                 except IndexError :
                     det[et[0]] = 1
         return det
+    
+    def make_theme_dict(self):
+        themes = [val for uci in self.ucis for val in uci.paras]
+        det = {}
+        for theme in themes :
+            th = theme.split('_')
+            if th[0] in det :
+                try :
+                    endth = '_'.join(th[1:])
+                    if theme in det[th[0]] :
+                        det[th[0]][theme] += 1
+                    else :
+                        det[th[0]][theme] = 1
+                except IndexError :
+                    det[th[0]] += 1
+            else :
+                try :
+                    endth = '_'.join(th[1:])
+                    det[th[0]] = {theme:1}
+                except IndexError :
+                    det[th[0]] = 1
+        return det
 
     def make_etline(self, listet) :
         etuces = [[] for et in listet]
@@ -1545,6 +1597,7 @@ class Builder :
                 filein = parametres['dictionary']
             else :
                 filein = None
+            dial.Destroy()
             ReadLexique(self.parent, lang = parametres['lang'], filein = filein)
             if parametres['lang'] != 'other' and  os.path.exists(self.parent.DictPath.get(parametres['lang']+'_exp', 'french_exp')):
                 self.parent.expressions = ReadDicoAsDico(self.parent.DictPath.get(parametres['lang']+'_exp', 'french_exp'))
@@ -1552,9 +1605,9 @@ class Builder :
                 self.parent.expressions = {}
             self.parametres = parametres
         else :
+            dial.Destroy()
             if self.dlg is not None :
                 self.dlg.Destroy()
-        dial.Destroy()
 
     def doanalyse(self) :
         return BuildFromAlceste(self.parent.filename, self.parametres, self.parent.lexique, self.parent.expressions, dlg = self.dlg).corpus