Update app.py

app.py

@@ -1773,6 +1773,8 @@ def get_top_scoring_statements(dataset):
     return top_statements
 
 last_top_scoring_statements = None 
+
+
 def chatbot_response(message, history):
     """
     Generate chatbot response dynamically using selected dataset, user input, and maintaining history.
@@ -1828,17 +1830,20 @@ def chatbot_response(message, history):
         else:
             return f"Error: Invalid dataset selection '{selected_dataset_ai}'."
 
-        # Prepare top-scoring statements
-        top_scoring_statements = "### Top Scoring Statements ###\n\n"
-        for bucket, statements in trust_data.items():
-            top_scoring_statements += f"**{bucket}**:\n"
-            for statement in statements:
-                top_scoring_statements += f"- {statement}\n"
-            top_scoring_statements += "\n"
-        last_top_scoring_statements = top_scoring_statements
-
-        # Combine predefined prompt, top-scoring statements, and user input
-        combined_prompt = "\n\n### Top-Scoring Statements for Integration ###\n" + top_scoring_statements
+        # If top-scoring statements have not been generated yet, generate them
+        if not last_top_scoring_statements:
+            top_scoring_statements = "### Top Scoring Statements ###\n\n"
+            for bucket, statements in trust_data.items():
+                top_scoring_statements += f"**{bucket}**:\n"
+                for statement in statements:
+                    top_scoring_statements += f"- {statement}\n"
+                top_scoring_statements += "\n"
+            last_top_scoring_statements = top_scoring_statements  # Save this for future use
+        else:
+            top_scoring_statements = ""  # Don't include the top-scoring statements in subsequent messages
+
+        # Combine predefined prompt, top-scoring statements (only the first time), and user input
+        combined_prompt = top_scoring_statements  # Include top-scoring statements only once
         combined_prompt += "\n\nUser Input:\n" + message
         trust_tip, suggestion = get_trust_tip_and_suggestion()
         trust_tip_and_suggestion = f"\n\n---\n\n**Trust Tip**: {trust_tip}\n\n**Suggestion**: {suggestion}"
@@ -1866,7 +1871,9 @@ def chatbot_response(message, history):
 
         # Prepare the final response
         response = f"**Selected Dataset: {selected_dataset_ai}**\n\n"
-        response += f"{top_scoring_statements}\n"
+        # Only include top-scoring statements once
+        if top_scoring_statements:
+            response += f"{top_scoring_statements}\n"
         response += f"\n{agent_output['output']}"
         response += trust_tip_and_suggestion
 
@@ -1886,6 +1893,119 @@ def chatbot_response(message, history):
         logger.error(f"Unexpected error: {e}")
         return "Error occurred during response generation."
 
+# def chatbot_response(message, history):
+#     """
+#     Generate chatbot response dynamically using selected dataset, user input, and maintaining history.
+#     """
+#     global selected_dataset_ai, last_top_scoring_statements, chat_history
+
+#     try:
+#         # Ensure a dataset is selected
+#         if not selected_dataset_ai:
+#             return "Error: No dataset selected. Please select a dataset and try again."
+
+#         # Define datasets and corresponding trust buckets
+#         datasets = {
+#             "VW Owners.xlsx": {
+#                 "Development": [
+#                     "We bring together the world's best talent in many disciplines to create your cars. (25%)",
+#                     "Building great and affordable cars is our foundation. (22%)",
+#                     "Our beginnings are a unique combination of investors and unions. (18%)",
+#                 ],
+#                 "Benefit": [
+#                     "We bring together the world's best talent in many disciplines to create your cars. (23%)",
+#                     "We strongly focus on keeping and nurturing our team and have a 99.5% retention rate. (18%)",
+#                     "Employees are provided with extensive continuous training. (16%)",
+#                 ],
+#                 "Vision": [
+#                     "Our brands are ranked No. 2 and 5 in the reliability rankings. (27%)",
+#                     "Our technology and manufacturing capabilities are second to none. (22%)",
+#                     "We produce almost 9 million cars per year. (15%)",
+#                 ],
+#             },
+#             "Volkswagen Non Customers.xlsx": {
+#                 "Stability": [
+#                     "We work with our unions in our restructuring and future plans. (21%)",
+#                     "We have learned from our mistakes in the Diesel Affair and we have made fundamental changes. (19%)",
+#                     "Building great and affordable cars is our foundation. (18%)",
+#                 ],
+#                 "Relationship": [
+#                     "We put a lot of emphasis on the interior experience and two of our cars have been ranked in the top 10. (24%)",
+#                     "We are at the forefront of technology to deliver better cars and driving experiences. (17%)",
+#                     "Our beginnings are a unique combination of investors and unions and today 9 of our 20 board members are staff representatives. (17%)",
+#                 ],
+#                 "Competence": [
+#                     "At the heart of our decision-making is the long-term quality of life for all of us. (20%)",
+#                     "We put a lot of emphasis on the interior experience and two of our cars have been ranked in the top 10. (19%)",
+#                     "We are one of the longest-established car companies. (18%)",
+#                 ],
+#             },
+#         }
+
+#         # Retrieve relevant data for the selected dataset
+#         if selected_dataset_ai in datasets:
+#             trust_data = datasets[selected_dataset_ai]
+#         else:
+#             return f"Error: Invalid dataset selection '{selected_dataset_ai}'."
+
+#         # Prepare top-scoring statements
+#         top_scoring_statements = "### Top Scoring Statements ###\n\n"
+#         for bucket, statements in trust_data.items():
+#             top_scoring_statements += f"**{bucket}**:\n"
+#             for statement in statements:
+#                 top_scoring_statements += f"- {statement}\n"
+#             top_scoring_statements += "\n"
+#         last_top_scoring_statements = top_scoring_statements
+
+#         # Combine predefined prompt, top-scoring statements, and user input
+#         combined_prompt = "\n\n### Top-Scoring Statements for Integration ###\n" + top_scoring_statements
+#         combined_prompt += "\n\nUser Input:\n" + message
+#         trust_tip, suggestion = get_trust_tip_and_suggestion()
+#         trust_tip_and_suggestion = f"\n\n---\n\n**Trust Tip**: {trust_tip}\n\n**Suggestion**: {suggestion}"
+
+#         # Validate chat history
+#         validated_chat_history = []
+#         for entry in history:
+#             if isinstance(entry, dict) and "role" in entry and "content" in entry:
+#                 validated_chat_history.append(entry)
+#             else:
+#                 logger.warning(f"Invalid chat history entry skipped: {entry}")
+
+#         # Include validated history in the prompt
+#         for entry in validated_chat_history:
+#             combined_prompt += f"\n{entry['role']}: {entry['content']}"
+
+#         # Structured input for agent execution
+#         structured_input = {
+#             "input": combined_prompt,
+#             "chat_history": validated_chat_history,
+#         }
+
+#         # Generate AI output using the agent pipeline (replace with actual logic)
+#         agent_output = agent_executor.invoke(structured_input)
+
+#         # Prepare the final response
+#         response = f"**Selected Dataset: {selected_dataset_ai}**\n\n"
+#         response += f"{top_scoring_statements}\n"
+#         response += f"\n{agent_output['output']}"
+#         response += trust_tip_and_suggestion
+
+#         # Append interaction to history
+#         validated_chat_history.append({"role": "user", "content": message})
+#         validated_chat_history.append({"role": "assistant", "content": agent_output["output"]})
+
+#         return response
+
+#     except KeyError as ke:
+#         logger.error(f"KeyError encountered: {ke}")
+#         return "An unexpected error occurred. Please try again."
+#     except ValueError as ve:
+#         logger.error(f"ValueError encountered: {ve}")
+#         return "An unexpected value was encountered. Please refine your input."
+#     except Exception as e:
+#         logger.error(f"Unexpected error: {e}")
+#         return "Error occurred during response generation."
+
 def read_ai_dataset_selection():
     global selected_dataset_ai
     return selected_dataset_ai