mginoben commited on
Commit
fe9ff70
1 Parent(s): 6912dca

Added no profanity handle

Browse files
Files changed (1) hide show
  1. app.py +11 -6
app.py CHANGED
@@ -72,7 +72,6 @@ def fuzzy_lookup(tweet):
72
  for prof_variant in prof_varations:
73
  tweet = tweet.replace(prof_variant, profanity)
74
 
75
- print('Fuzzy Returns:', tweet)
76
  return tweet, matches
77
 
78
 
@@ -131,24 +130,30 @@ def query(payload):
131
  def predict(tweet):
132
 
133
  fuzzy_text, matches = fuzzy_lookup(tweet)
134
- output = query(preprocess(fuzzy_text))
135
-
 
 
136
  if 'error' in output:
137
- return output['error'], 'Error occured. Try again later.', {"error": "error"}
 
 
138
  else:
139
  output = [tuple(i.values()) for i in output[0]]
140
  output = dict((x, y) for x, y in output)
141
  predicted_label = list(output.keys())[0]
142
 
143
  if predicted_label == 'Abusive':
 
144
  for base_word, _ in matches.items():
145
-
146
  tweet = tweet.replace(base_word, re.sub("[a-zA-Z0-9@]", "*", base_word))
147
-
148
  return output, tweet, json.dumps(matches)
149
  else:
150
  return output, tweet, json.dumps(matches)
151
 
 
 
 
152
  hf_writer = gr.HuggingFaceDatasetSaver('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', 'tagalog-profanity-feedbacks')
153
 
154
 
 
72
  for prof_variant in prof_varations:
73
  tweet = tweet.replace(prof_variant, profanity)
74
 
 
75
  return tweet, matches
76
 
77
 
 
130
  def predict(tweet):
131
 
132
  fuzzy_text, matches = fuzzy_lookup(tweet)
133
+ processed_text = preprocess(fuzzy_text)
134
+ output = query(processed_text)
135
+ match_profanities = set(processed_text.split()) & set(list(profanities.keys()))
136
+
137
  if 'error' in output:
138
+ return output['error'], 'Error occured. Try again later.', {}
139
+ elif len(match_profanities) == 0:
140
+ return 'No Profanity Found.', '', {}
141
  else:
142
  output = [tuple(i.values()) for i in output[0]]
143
  output = dict((x, y) for x, y in output)
144
  predicted_label = list(output.keys())[0]
145
 
146
  if predicted_label == 'Abusive':
147
+ # Censor
148
  for base_word, _ in matches.items():
 
149
  tweet = tweet.replace(base_word, re.sub("[a-zA-Z0-9@]", "*", base_word))
 
150
  return output, tweet, json.dumps(matches)
151
  else:
152
  return output, tweet, json.dumps(matches)
153
 
154
+ # output, tweet, matches = predict('Sama ng ugali mo pre')
155
+ # print(output, '\n', tweet, '\n', matches)
156
+
157
  hf_writer = gr.HuggingFaceDatasetSaver('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', 'tagalog-profanity-feedbacks')
158
 
159