update check embeddings to resolve error in which vectors were being …

…generated even when they were not requested

examples/featurize.py

@@ -45,7 +45,7 @@
 		output_file_path_chat_level = "./jury_TINY_output_chat_level.csv",
 		output_file_path_user_level = "./jury_TINY_output_user_level.csv",
 		output_file_path_conv_level = "./jury_TINY_output_conversation_level.csv",
-		turns = False,
+		turns = False
 	)
 	tiny_juries_feature_builder.featurize(col="message")
 

src/team_comm_tools/feature_dict.py

@@ -23,7 +23,7 @@
     "vect_data": False,
     "bert_sentiment_data": False
   },
-  "Positivity (BERT)": {
+  "Positivity (RoBERTa)": {
     "columns": ["positive_bert", "negative_bert", "neutral_bert"],
     "file": "./utils/check_embeddings.py",
     "level": "Chat",

src/team_comm_tools/utils/check_embeddings.py

@@ -48,25 +48,34 @@ def check_embeddings(chat_data, vect_path, bert_path, need_sentence, need_sentim
     :return: None
     :rtype: None
     """
-    if regenerate_vectors or (not os.path.isfile(vect_path)):
+    if (regenerate_vectors or (not os.path.isfile(vect_path))) and need_sentence:
         generate_vect(chat_data, vect_path, message_col)
-    if regenerate_vectors or (not os.path.isfile(bert_path)):
+    if (regenerate_vectors or (not os.path.isfile(bert_path))) and need_sentiment:
         generate_bert(chat_data, bert_path, message_col)
     if (not os.path.isfile(Path(__file__).resolve().parent.parent/"features/lexicons/certainty.txt")):
         # unpickle certainty
         unpickle_certainty()
 
-    vector_df = pd.read_csv(vect_path)
-    bert_df = pd.read_csv(bert_path)
-    # check is given vector and bert data matches length of chat data 
-    if len(vector_df) != len(chat_data):
-        print("ERROR: The length of the vector data does not match the length of the chat data.")
-        generate_vect(chat_data, vect_path, message_col)
-
-    if len(bert_df) != len(chat_data):
-        print("ERROR: The length of the bert data does not match the length of the chat data.")
-        generate_bert(chat_data, bert_path, message_col)
-
+    try:
+        vector_df = pd.read_csv(vect_path)
+        # check whether the given vector and bert data matches length of chat data 
+        if len(vector_df) != len(chat_data):
+            print("ERROR: The length of the vector data does not match the length of the chat data. Regenerating...")
+            generate_vect(chat_data, vect_path, message_col)
+    except FileNotFoundError: # It's OK if we don't have the path, if the sentence vectors are not necessary
+        if need_sentence:
+            generate_vect(chat_data, vect_path, message_col)
+
+    try:
+        bert_df = pd.read_csv(bert_path)
+        if len(bert_df) != len(chat_data):
+            print("ERROR: The length of the sentiment data does not match the length of the chat data. Regenerating...")
+            generate_bert(chat_data, bert_path, message_col)
+    except FileNotFoundError:
+        if need_sentiment: # It's OK if we don't have the path, if the sentiment features are not necessary
+            generate_bert(chat_data, bert_path, message_col)
+
+    # Get the lexicon pickle
     current_script_directory = Path(__file__).resolve().parent
     LEXICON_PATH_STATIC = current_script_directory.parent/"features/lexicons_dict.pkl"
     if (not os.path.isfile(LEXICON_PATH_STATIC)):
@@ -135,7 +144,7 @@ def generate_vect(chat_data, output_path, message_col):
     :rtype: None
     """
 
-    print(f"Generating sentence vectors....")
+    print(f"Generating SBERT sentence vectors...")
 
     embedding_arr = [row.tolist() for row in model_vect.encode(chat_data[message_col])]
     embedding_df = pd.DataFrame({'message': chat_data[message_col], 'message_embedding': embedding_arr})
@@ -158,7 +167,7 @@ def generate_bert(chat_data, output_path, message_col):
     :return: None
     :rtype: None
     """
-    print(f"Generating BERT sentiments....")
+    print(f"Generating RoBERTa sentiments...")
 
     messages = chat_data[message_col]
     sentiments = messages.apply(get_sentiment)