Update app.py

app.py

@@ -1,4 +1,4 @@
-# Changed now everyting from session_index to session_id 
+# Changed now everyting from session_index to PROLIFIC_PID 
 
 # Original code from https://huggingface.co/spaces/huggingface-projects/llama-2-7b-chat
 # Modified for trust game purposes
@@ -86,7 +86,7 @@ if torch.cuda.is_available():
     tokenizer = AutoTokenizer.from_pretrained(model_id)
     tokenizer.use_default_system_prompt = False
     
-def fetch_personalized_data(session_id):
+def fetch_personalized_data(PROLIFIC_PID):
     try:
         # Connect to the database
         with mysql.connector.connect(
@@ -97,13 +97,13 @@ def fetch_personalized_data(session_id):
         ) as conn:
             # Create a cursor object
             with conn.cursor() as cursor:
-                # Query to fetch relevant data from both tables based on session_id
+                # Query to fetch relevant data from both tables based on PROLIFIC_PID
                 query = """
                     SELECT e5390g37581_core.playerNr, 
                            e5390g37581_core.groupNrStart, 
                            e5390g37581_core.subjectNr, 
                            e5390g37581_core.onPage, 
-                           e5390g37581_decisions.session_id, 
+                           e5390g37581_decisions.PROLIFIC_PID, 
                            e5390g37581_decisions.transfer1,
                            e5390g37581_decisions.tripledAmount1,
                            e5390g37581_decisions.keptForSelf1,
@@ -123,13 +123,13 @@ def fetch_personalized_data(session_id):
                     FROM e5390g37581_core
                     JOIN e5390g37581_decisions ON 
                         e5390g37581_core.playerNr = e5390g37581_decisions.playerNr
-                    WHERE e5390g37581_decisions.session_id = %s
+                    WHERE e5390g37581_decisions.PROLIFIC_PID = %s
                     UNION ALL
                     SELECT e5390g37581_core.playerNr, 
                            e5390g37581_core.groupNrStart, 
                            e5390g37581_core.subjectNr, 
                            e5390g37581_core.onPage, 
-                           e5390g37581_decisions.session_id, 
+                           e5390g37581_decisions.PROLIFIC_PID, 
                            e5390g37581_decisions.transfer1,
                            e5390g37581_decisions.tripledAmount1,
                            e5390g37581_decisions.keptForSelf1,
@@ -154,17 +154,17 @@ def fetch_personalized_data(session_id):
                         FROM e5390g37581_core
                         JOIN e5390g37581_decisions 
                             ON e5390g37581_core.playerNr = e5390g37581_decisions.playerNr
-                        WHERE e5390g37581_decisions.session_id = %s
-                    ) AND e5390g37581_decisions.session_id != %s
+                        WHERE e5390g37581_decisions.PROLIFIC_PID = %s
+                    ) AND e5390g37581_decisions.PROLIFIC_PID != %s
                 """
-                cursor.execute(query,(session_id, session_id, session_id))
+                cursor.execute(query,(PROLIFIC_PID, PROLIFIC_PID, PROLIFIC_PID))
                 # Fetch data row by row
                 data = [{
                     'playerNr': row[0], 
                     'groupNrStart': row[1], 
                     'subjectNr': row[2], 
                     'onPage': row[3],
-                    'session_id': row[4],
+                    'PROLIFIC_PID': row[4],
                     'transfer1': row[5],
                     'tripledAmount1': row[6],
                     'keptForSelf1': row[7],
@@ -277,18 +277,18 @@ def generate(
     #print(params)
     
     # Assuming params = request.query_params is the dictionary containing the query parameters    
-    # Extract the value of the 'session_id' parameter
-    session_id = params.get('session_id')
+    # Extract the value of the 'PROLIFIC_PID' parameter
+    PROLIFIC_PID = params.get('PROLIFIC_PID')
     
-    # Check if session_id_value is None or contains a value
-    if session_id is not None:
-        print("Session index:", session_id)
+    # Check if PROLIFIC_PID_value is None or contains a value
+    if PROLIFIC_PID is not None:
+        print("Session index:", PROLIFIC_PID)
     else:
-        session_id = 'no_session_id'
+        PROLIFIC_PID = 'no_PROLIFIC_PID'
         print("Session index not found or has no value.")
 
     # Fetch personalized data
-    personalized_data = fetch_personalized_data(session_id)
+    personalized_data = fetch_personalized_data(PROLIFIC_PID)
     
     # Construct the input prompt using the functions from the system_prompt_config module
     input_prompt = construct_input_prompt(chat_history, message, personalized_data)
@@ -355,52 +355,79 @@ def generate(
     
     # Iterate over each dictionary in the list
     for entry in personalized_data:
-        # Check if the session_id matches the value in session_id variable
-        if entry['session_id'] == session_id: 
+        # Check if the PROLIFIC_PID matches the value in PROLIFIC_PID variable
+        if entry['PROLIFIC_PID'] == PROLIFIC_PID: 
             # If a match is found, retrieve the onPage value
             onPage = entry['onPage']
             break  # Break the loop since we found the desired entry
             
-    # Check if onPage is still None (i.e., no onPage found or session_id is None)
+    # Check if onPage is still None (i.e., no onPage found or PROLIFIC_PID is None)
     if onPage is None:
         onPage = "no_onPage"
+
+    # Find playerNr and groupNrStart variables in personalized_data to add it to the .csv filename to record 
+    # Initialize onPage variable to None
+    playerNr = None
+    groupNrStart = None 
     
+    # Iterate over each dictionary in the list
+    for entry in personalized_data:
+        # Check if the PROLIFIC_PID matches the value in PROLIFIC_PID variable
+        if entry['PROLIFIC_PID'] == PROLIFIC_PID: 
+            # If a match is found, retrieve the onPage value
+            playerNr = entry['playerNr']
+            groupNrStart = entry['groupNrStart']
+            break  # Break the loop since we found the desired entry
+            
+    # Check if playerNr, groupNrStart is still None (i.e., no onPage found or PROLIFIC_PID is None)
+    if playerNr is None:
+        playerNr = "no_playerNr"
+        
+    if groupNrStart is None:
+        groupNrStart = "groupNrStart"
+        
     # Print the onPage value
-    print("onPage for session_id =", session_id, ":", onPage)
+    print("onPage for PROLIFIC_PID =", PROLIFIC_PID, ":", onPage)
     onPage_filename, onPage_prompt = map_onPage(onPage)
     print("onPage_filename:", onPage_filename)
     print("onPage_prompt:", onPage_prompt)
 
+    # Print the playerNr, groupNrStart values
+    print(groupNrStart)
+    print(playerNr)
+
     # Save chat history to .csv file on HuggingFace Hub 
     # Generate filename with bot id and session id
-    filename = f"{session_id}_{onPage}_{DATA_FILENAME}"
+    filename = f"{groupNrStart}_{playerNr}_{PROLIFIC_PID}_{onPage}_{DATA_FILENAME}"
     data_file = os.path.join(DATA_DIRECTORY, filename)
-    
+
     # Generate timestamp
     timestamp = datetime.datetime.now()
     
+    # Create a DataFrame for the current conversation turn
+    turn_data = {
+        "turn_id": len(existing_data) + 1 if existing_data is not None else 1,
+        "role": "user",
+        "content": user_message,
+        "ip_address": request.client.host,
+        "readable_sentence": user_readable_sentence,
+        "timestamp": timestamp,
+    }
+    turn_df = pd.DataFrame([turn_data])
+    
     # Check if the file already exists
     if os.path.exists(data_file):
         # If file exists, load existing data
         existing_data = pd.read_csv(data_file)
-        
-        # Add timestamp column
-        conversation_df = pd.DataFrame(conversation)
-        conversation_df['ip_address'] = request.client.host
-        conversation_df['readable_sentence'] = readable_sentence
-        conversation_df['timestamp'] = timestamp
-        
-        # Append new conversation to existing data
-        updated_data = pd.concat([existing_data, conversation_df], ignore_index=True)
-        updated_data.to_csv(data_file, index=False)
+        # Append the new conversation turn to the existing data
+        updated_data = pd.concat([existing_data, turn_df], ignore_index=True)
     else:
-        # If file doesn't exist, create new file with conversation data
-        conversation_df = pd.DataFrame(conversation)
-        conversation_df['ip_address'] = request.client.host
-        conversation_df['readable_sentence'] = readable_sentence
-        conversation_df['timestamp'] = timestamp
-        conversation_df.to_csv(data_file, index=False)
-
+        # If file doesn't exist, create new DataFrame with the current conversation turn
+        updated_data = turn_df
+    
+    # Write the updated data to the CSV file
+    updated_data.to_csv(data_file, index=False)
+    
     print("Updating .csv")
     repo.push_to_hub(blocking=False, commit_message=f"Updating data at {timestamp}")