fix: handle StopIteration in HF Inference API streaming

llm_handler.py

@@ -112,19 +112,30 @@ class LLMHandler:
             stream = self.client.text_generation(
                 prompt=full_prompt,
                 model=self.model,
-                max_new_tokens=1024,
+                max_new_tokens=512,
                 temperature=0.7,
                 top_p=0.95,
                 stream=True,
                 details=False,
             )
             
+            # Collect tokens and yield them
+            has_content = False
             for token in stream:
-                yield token
+                if token:  # Only yield non-empty tokens
+                    has_content = True
+                    yield token
+            
+            # If no content was generated, yield a fallback message
+            if not has_content:
+                yield "I apologize, but I couldn't generate an answer based on the provided context. Please try rephrasing your question."
         
+        except StopIteration:
+            # Handle empty generator
+            yield "I apologize, but I couldn't generate an answer. The model returned an empty response."
         except Exception as e:
             logger.error(f"Error streaming answer: {e}")
-            raise
+            yield f"\n\n❌ Error: {str(e)}"
 
 
 def format_response(