Update app.py

app.py

@@ -78,31 +78,34 @@ def update_vectors(files, parser):
     
     return f"Vector store updated successfully. Processed {total_chunks} chunks from {len(files)} files using {parser}."
 
-def generate_chunked_response(prompt, model, max_tokens=1000, max_chunks=5, temperature=0.2):
+def generate_chunked_response(prompt, model, max_tokens=1000, num_calls=3, temperature=0.2):
     client = InferenceClient(
         model,
         token=huggingface_token,
     )
     
-    full_response = ""
+    full_responses = []
     messages = [{"role": "user", "content": prompt}]
     
-    try:
-        for message in client.chat_completion(
-            messages=messages,
-            max_tokens=max_tokens,
-            temperature=temperature,
-            stream=True,
-        ):
-            chunk = message.choices[0].delta.content
-            if chunk:
-                full_response += chunk
-        
-    except Exception as e:
-        print(f"Error in generating response: {str(e)}")
+    for _ in range(num_calls):
+        try:
+            response = ""
+            for message in client.chat_completion(
+                messages=messages,
+                max_tokens=max_tokens,
+                temperature=temperature,
+                stream=True,
+            ):
+                if message.choices and message.choices[0].delta and message.choices[0].delta.content:
+                    chunk = message.choices[0].delta.content
+                    response += chunk
+            full_responses.append(response)
+        except Exception as e:
+            print(f"Error in generating response: {str(e)}")
     
-    # Clean up the response
-    clean_response = re.sub(r'<s>\[INST\].*?\[/INST\]\s*', '', full_response, flags=re.DOTALL)
+    # Combine and clean up the responses
+    combined_response = " ".join(full_responses)
+    clean_response = re.sub(r'<s>\[INST\].*?\[/INST\]\s*', '', combined_response, flags=re.DOTALL)
     clean_response = clean_response.replace("Using the following context:", "").strip()
     clean_response = clean_response.replace("Using the following context from the PDF documents:", "").strip()