Fix gemma chat template fallback

app/models/huggingface_local.py

@@ -73,7 +73,9 @@ class HuggingFaceLocal(BaseLLM):
         if not self._initialized:
             raise RuntimeError(f"[{self.name}] Model not initialized")
         
-        # Format prompt
+        formatted_prompt = None
+        
+        # Format prompt from chat messages
         if chat_messages:
             try:
                 formatted_prompt = self.tokenizer.apply_chat_template(
@@ -81,15 +83,16 @@ class HuggingFaceLocal(BaseLLM):
                     tokenize=False,
                     add_generation_prompt=True
                 )
-            except Exception:
-                # Fallback to raw prompt
-                if prompt:
-                    formatted_prompt = prompt
-                else:
-                    raise ValueError("Cannot generate without valid prompt")
-        elif prompt:
+            except Exception as e:
+                print(f"[{self.name}] apply_chat_template failed: {e}, using fallback")
+                # Fallback: manually format chat messages
+                formatted_prompt = self._format_chat_fallback(chat_messages)
+        
+        # Use raw prompt if provided and no chat_messages
+        if formatted_prompt is None and prompt:
             formatted_prompt = prompt
-        else:
+        
+        if formatted_prompt is None:
             raise ValueError("Either prompt or chat_messages required")
         
         # Generate
@@ -115,13 +118,35 @@ class HuggingFaceLocal(BaseLLM):
                 response = full_text
             
             # Clean up special tokens
-            if response.endswith("<|im_end|>"):
-                response = response[:-len("<|im_end|>")]
+            for token in ["<|im_end|>", "<end_of_turn>", "<eos>", "</s>"]:
+                if response.endswith(token):
+                    response = response[:-len(token)]
             
             return response.strip()
         
         return ""
     
+    def _format_chat_fallback(self, chat_messages: List[Dict[str, str]]) -> str:
+        """
+        Fallback chat formatting for models without proper chat template.
+        Works with Gemma and other models.
+        """
+        formatted = ""
+        for msg in chat_messages:
+            role = msg.get("role", "user")
+            content = msg.get("content", "")
+            
+            if role == "system":
+                formatted += f"{content}\n\n"
+            elif role == "user":
+                formatted += f"User: {content}\n"
+            elif role == "assistant":
+                formatted += f"Assistant: {content}\n"
+        
+        # Add generation prompt
+        formatted += "Assistant:"
+        return formatted
+    
     def get_info(self) -> Dict[str, Any]:
         """Return model info."""
         return {