~further changes

evaluation.py

 from model import *
 from peewee import fn
 import json
+from difflib import SequenceMatcher
+
 
 punctuation = ['!', '?', '.', ',', ';', ':', '-', '_', '"', "'", '…', '—', '#', '@']
-semantic = ['e.v.', 'ev', 'ev.', 'e.v', 'herr', 'hr', 'frau', 'fr', 'ag', 'gmbh', '& co kg']
+semantic = ['e.v.', 'ev.', 'e.v', 'ev', 'herr', 'hr', 'frau', 'fr', 'ag', 'gmbh', '& co kg', 'dr.', 'dr']
 
 def remove_punctuation(text):
     for char in punctuation:
@@ -16,9 +18,9 @@ def remove_semantics(text):
     return text
 
 def replace_umlautes(text):
-    text = text.replace("oe", "ö")
+    text = text.replace("ö", "oe")
     text = text.replace("ß", "ss")
-    text = text.replace("ae", "ä")
+    text = text.replace("ä", "ae")
     text = text.replace("ue", "ü")
     return text
 
@@ -62,7 +64,35 @@ def update_answer_from_file():
         jzon = json.load(file)
         update_answer_text_batch(jzon)
 
-def generate_answer_classes():
+def generate_answer_classes(treshhold):
+    questions = Question.select()
+    for q in questions:
+        answers = Answer.select().where(Answer.question == q)
+        answer_list = []
+        for a in answers:
+            answer_list.append(a)
+        
+        length = len(answer_list)
+        classes = []
+        for i in range(length):
+            answer = answer_list[i]
+            if answer is None:
+                    continue
+            current_class = [answer.text]
+            for j in range(i+1, length):
+                bnswer = answer_list[j]
+                if bnswer is None:
+                    continue
+                r = SequenceMatcher(None, answer.text, bnswer.text).ratio()
+                if r > treshhold:
+                    current_class.append(bnswer.text)
+                    answer_list[j] = None
+            current_class = list(set(current_class))
+            current_class = sorted(current_class)
+            classes.append(current_class)
+        
+        with open("q%d-classes.json" % q.id, "w") as file:
+            json.dump(classes, file, indent=4, sort_keys=True)
 
 
 def aggregate_answer_file():

model/__init__.py

@@ -3,9 +3,9 @@ from peewee import Model
 from peewee import MySQLDatabase
 import json
 
-mysql_creds = json.load(open("./creds.json"))
+#mysql_creds = json.load(open("./creds.json"))
 
-sqlite = SqliteDatabase("cfd.sqlite", pragmas={'foreign_keys': 1})
+sqlite = SqliteDatabase("model/cfd.sqlite", pragmas={'foreign_keys': 1})
 sqlite.connect()
 #sqlite = MySQLDatabase(mysql_creds["db"], user=mysql_creds["user"], password=mysql_creds['password'],
 #                         host=mysql_creds['host'], port=int(mysql_creds['port']),)