Update app.py

app.py

@@ -259,89 +259,89 @@ class RAGPipeline:
     #         return message
 
     def process_query(self, query: str, placeholder) -> str:
-    try:
-        # Preprocess query
-        query = self.preprocess_query(query)
-        logging.info(f"Processing query: {query}")
+        try:
+            # Preprocess query
+            query = self.preprocess_query(query)
+            logging.info(f"Processing query: {query}")
         
-        # Show retrieval status
-        status = placeholder.empty()
-        status.write("🔍 Finding relevant information...")
+            # Show retrieval status
+            status = placeholder.empty()
+            status.write("🔍 Finding relevant information...")
         
-        # Get embeddings and search
-        query_embedding = self.retriever.encode([query])
-        similarities = F.cosine_similarity(query_embedding, self.retriever.doc_embeddings)
-        scores, indices = torch.topk(similarities, k=min(self.k, len(self.documents)))
+            # Get embeddings and search
+            query_embedding = self.retriever.encode([query])
+            similarities = F.cosine_similarity(query_embedding, self.retriever.doc_embeddings)
+            scores, indices = torch.topk(similarities, k=min(self.k, len(self.documents)))
         
-        # Log similarity scores
-        for idx, score in zip(indices.tolist(), scores.tolist()):
-            logging.info(f"Score: {score:.4f} | Document: {self.documents[idx][:100]}...")
+            # Log similarity scores
+            for idx, score in zip(indices.tolist(), scores.tolist()):
+                logging.info(f"Score: {score:.4f} | Document: {self.documents[idx][:100]}...")
         
-        relevant_docs = [self.documents[idx] for idx in indices.tolist()]
+            relevant_docs = [self.documents[idx] for idx in indices.tolist()]
         
-        # Update status
-        status.write("💭 Generating response...")
+            # Update status
+            status.write("💭 Generating response...")
         
-        # Prepare context and prompt
-        context = "\n".join(relevant_docs[:3])
-        prompt = f"""Context information is below:
-        {context}
+            # Prepare context and prompt
+            context = "\n".join(relevant_docs[:3])
+            prompt = f"""Context information is below:
+            {context}
         
-        Given the context above, please answer the following question:
-        {query}
-
-        Guidelines:
-        - If you cannot answer based on the context, say so politely
-        - Keep the response concise and focused
-        - Only include sports-related information
-        - No dates or timestamps in the response
-        - Use clear, natural language
+            Given the context above, please answer the following question:
+            {query}
+
+            Guidelines:
+            - If you cannot answer based on the context, say so politely
+            - Keep the response concise and focused
+            - Only include sports-related information
+            - No dates or timestamps in the response
+            - Use clear, natural language
         
-        Answer:"""
+            Answer:"""
         
-        # Generate response
-        response_placeholder = placeholder.empty()
+            # Generate response
+            response_placeholder = placeholder.empty()
         
-        try:
-            # Add logging for model state
-            logging.info("Model state check - Is None?: " + str(self.llm is None))
+            try:
+                # Add logging for model state
+                logging.info("Model state check - Is None?: " + str(self.llm is None))
             
-            # Directly use Llama model
-            response = self.llm(
-                prompt,
-                max_tokens=512,
-                temperature=0.4,
-                top_p=0.95,
-                echo=False,
-                stop=["Question:", "\n\n"]
-            )
+                # Directly use Llama model
+                response = self.llm(
+                    prompt,
+                    max_tokens=512,
+                    temperature=0.4,
+                    top_p=0.95,
+                    echo=False,
+                    stop=["Question:", "\n\n"]
+                )
             
-            logging.info(f"Raw model response: {response}")
+                logging.info(f"Raw model response: {response}")
             
-            if response and isinstance(response, dict) and 'choices' in response:
-                generated_text = response['choices'][0].get('text', '').strip()
-                if generated_text:
-                    final_response = self.postprocess_response(generated_text)
-                    response_placeholder.markdown(final_response)
-                    return final_response
+                if response and isinstance(response, dict) and 'choices' in response:
+                    generated_text = response['choices'][0].get('text', '').strip()
+                    if generated_text:
+                        final_response = self.postprocess_response(generated_text)
+                        response_placeholder.markdown(final_response)
+                        return final_response
             
-            message = "No relevant answer found. Please try rephrasing your question."
-            response_placeholder.warning(message)
-            return message
+                message = "No relevant answer found. Please try rephrasing your question."
+                response_placeholder.warning(message)
+                return message
+            
+            except Exception as e:
+                logging.error(f"Generation error: {str(e)}")
+                logging.error(f"Full error details: ", exc_info=True)
+                message = f"Had some trouble generating the response: {str(e)}"
+                response_placeholder.warning(message)
+                return message
             
         except Exception as e:
-            logging.error(f"Generation error: {str(e)}")
+            logging.error(f"Process error: {str(e)}")
             logging.error(f"Full error details: ", exc_info=True)
-            message = f"Had some trouble generating the response: {str(e)}"
-            response_placeholder.warning(message)
+            message = f"Something went wrong: {str(e)}"
+            placeholder.warning(message)
             return message
-            
-    except Exception as e:
-        logging.error(f"Process error: {str(e)}")
-        logging.error(f"Full error details: ", exc_info=True)
-        message = f"Something went wrong: {str(e)}"
-        placeholder.warning(message)
-        return message
 
 @st.cache_resource(show_spinner=False)
 def initialize_rag_pipeline():