Merge pull request #54 from simnotes/de_quotecleanup

removed unnecessary quotationmarks from de-lang sentences

src/corporacreator/preprocessors/de.py

@@ -1,3 +1,35 @@
+import re
+QUOTE_PATTERN = re.compile(r'^\"{3}(.*)\"{2}(.*)\"{1}$')
+QUOTE_PATTERN_2 = re.compile(r'^\"{1}(.*)\"{2}(.*)\"{2}(.*)\"{1}$')
+QUOTE_PATTERN_3 = re.compile(r'^\"{1}(.*)\"{1}$')
+
+def _change_multi_quotes(sentence):
+    """Changes all quotes from patterns like 
+    [\"""content""content"] to ["content"content] or
+    ["content""content""content"] to [content"content"content] or
+    ["content" to content]
+    
+    Args:
+      sentence (str): Sentence to be cleaned up.
+      
+    Returns:
+      (str): Cleaned up sentence. Returns the sentence 'as-is', if matching
+      did not work as expected
+    """
+    matches = QUOTE_PATTERN.match(sentence) # pattern: \"\"\"content\"\"content\"
+    matches2 = QUOTE_PATTERN_2.match(sentence) # pattern: \"content\"\"content\"\"content\"
+    matches3 = QUOTE_PATTERN_3.match(sentence) # patter: \"content\"
+
+    if matches != None and matches.lastindex == 2:
+        return "\"{}\"{}".format(matches.group(1), matches.group(2))
+    elif matches2 != None and matches2.lastindex == 3:
+        return "{}\"{}\"{}".format(matches2.group(1), matches2.group(2), matches2.group(3))
+    elif matches3 != None and matches3.lastindex == 1:
+        return "{}".format(matches3.group(1))
+
+    return sentence
+
+
 def de(client_id, sentence):
     """Cleans up the passed sentence, removing or reformatting invalid data.
 
@@ -8,5 +40,7 @@ def de(client_id, sentence):
     Returns:
       (str): Cleaned up sentence. Returning None or a `str` of whitespace flags the sentence as invalid.
     """
+    sentence = _change_multi_quotes(sentence)
+
     # TODO: Clean up de data
     return sentence