Modified Suggestion function with argument word and threshold

SRT.py

@@ -405,7 +405,25 @@ class SRT_script():
                     seg.source_text = seg.source_text.lower().replace(word, term_enzh_dict.get(word))
                     logging.info("replace term: " + word + " --> " + term_enzh_dict.get(word) + " in time stamp {}".format(i+1))
                     logging.info("source text becomes: " + seg.source_text)
-
+                    
+    
+    comp_dict = []
+    
+    def fetchfunc(self,word,threshold):
+        import enchant
+        result = word;
+        threshold = threshold*len(word)
+        if len(self.comp_dict)==0:
+            with open("./finetune_data/dict_enzh.csv", 'r', encoding='utf-8') as f:
+                    self.comp_dict = {rows[0]: rows[1] for rows in reader(f)}
+        temp = ""
+        for matched in self.comp_dict:
+            if enchant.utils.levenshtein(word, matched)<enchant.utils.levenshtein(word, temp):
+                temp = matched
+        if enchant.utils.levenshtein(word, temp) < threshold:
+            result = temp
+        return result    
+    
     def spell_check_term(self):
         ## known bug: I've will be replaced because i've is not in the dict
         logging.info("performing spell check")
@@ -419,17 +437,21 @@ class SRT_script():
                 word = ready_words[i]
                 [real_word, pos] = self.get_real_word(word)
                 if not dict.check(word[:pos]):
-                    suggest = term_spellDict.suggest(real_word)
-                    if suggest and enchant.utils.levenshtein(word, suggest[0]) < (len(word)+len(suggest[0]))/4:  # relax spell check
+                    new_word = word.replace(word[:pos],self.fetchfunc(word[:pos],0.5))
+                    
+                    logging.info(word + "\t" + self.fetchfunc(word[:pos],0.5) + "\t" + str(enchant.utils.levenshtein(word, self.fetchfunc(word[:pos],0.5)))+'\n')
+                                
+                    #suggest = term_spellDict.suggest(real_word)
+                    #if suggest and enchant.utils.levenshtein(word, suggest[0]) < (len(word)+len(suggest[0]))/4:  # relax spell check
                         
                         # with open("dislog.log","a") as log:
                         #     if not os.path.exists("dislog.log"):
                         #         log.write("word \t suggest \t levenshtein \n")
-                        logging.info(word + "\t" + suggest[0] + "\t" + str(enchant.utils.levenshtein(word, suggest[0]))+'\n')                            
-                        #print(word + ":" + suggest[0] + ":---:levenshtein:" + str(enchant.utils.levenshtein(word, suggest[0])))                                                
-                        new_word = word.replace(word[:pos],suggest[0])
-                    else:
-                        new_word = word
+                    #    logging.info(word + "\t" + suggest[0] + "\t" + str(enchant.utils.levenshtein(word, suggest[0]))+'\n')                            
+                    #    #print(word + ":" + suggest[0] + ":---:levenshtein:" + str(enchant.utils.levenshtein(word, suggest[0])))                                                
+                    #    new_word = word.replace(word[:pos],suggest[0])
+                    #else:
+                    #    new_word = word
                 else: 
                     new_word = word
                 ready_words[i] = new_word