Update app.py

app.py

@@ -110,39 +110,43 @@ def classify(df, new_column = True):
   formatted_sentences = []
   for sentence in sentencesMCTIList_xp8:
     formatted_sentences.append(json.loads(sentence.replace("'",'"')))
-  del sentencesMCTIList_xp8
-
-  print("Transformado em W2V")
-  words = list(reloaded_w2v_model.wv.vocab) 
-  item_shape = np.shape(reloaded_w2v_model.wv[words[0]])
-  # print(formatted_sentences)
-
-  MCTIinput_vector = []
-  for sentence in formatted_sentences:
-    aux_vector = []
-    for word in sentence:
-      if word in reloaded_w2v_model.wv.vocab:
-        aux_vector.append(reloaded_w2v_model.wv[word])
-      else:
-        aux_vector.append(np.zeros(item_shape))
-      MCTIinput_vector.append(aux_vector)
-  del formatted_sentences
-  print("Convertido W2V")
-  MCTIinput_padded = pad_sequences(MCTIinput_vector, maxlen=2726, padding='pre')
-  del MCTIinput_vector
-  print("Sentenças com Padding")
-  print(len(MCTIinput_padded))
-  print(len(MCTIinput_padded[0]))
-  predictions = reconstructed_model_CNN.predict(MCTIinput_padded)
-  del MCTIinput_padded
-  print(predictions)
-
-  cleaned_up_predictions = []
-  for prediction in predictions:
-    cleaned_up_predictions.append(1 if prediction >= 0.5 else 0);
-  del predictions
-
-  df['classification'] = cleaned_up_predictions
+  # del sentencesMCTIList_xp8
+
+  print(sentencesMCTIList_xp8[0])
+  print("##########################")
+  print(formatted_sentences[0])
+
+  # print("Transformado em W2V")
+  # words = list(reloaded_w2v_model.wv.vocab) 
+  # item_shape = np.shape(reloaded_w2v_model.wv[words[0]])
+  # # print(formatted_sentences)
+
+  # MCTIinput_vector = []
+  # for sentence in formatted_sentences:
+  #   aux_vector = []
+  #   for word in sentence:
+  #     if word in reloaded_w2v_model.wv.vocab:
+  #       aux_vector.append(reloaded_w2v_model.wv[word])
+  #     else:
+  #       aux_vector.append(np.zeros(item_shape))
+  #     MCTIinput_vector.append(aux_vector)
+  # del formatted_sentences
+  # print("Convertido W2V")
+  # MCTIinput_padded = pad_sequences(MCTIinput_vector, maxlen=2726, padding='pre')
+  # del MCTIinput_vector
+  # print("Sentenças com Padding")
+  # print(len(MCTIinput_padded))
+  # print(len(MCTIinput_padded[0]))
+  # predictions = reconstructed_model_CNN.predict(MCTIinput_padded)
+  # del MCTIinput_padded
+  # print(predictions)
+
+  # cleaned_up_predictions = []
+  # for prediction in predictions:
+  #   cleaned_up_predictions.append(1 if prediction >= 0.5 else 0);
+  # del predictions
+
+  # df['classification'] = cleaned_up_predictions
   return df
 
 def gen_output(data):