fix incorrect LLM message

app/llm.py

@@ -102,18 +102,21 @@ class LLMClient:
         """
         logger.info(f"[LLM] generate_text - provider: {self.provider}, prompt: {prompt}")
         try:
+            result = None
             if self.provider == "openai":
-                return await self._generate_openai(prompt, system_prompt, **kwargs)
+                result = await self._generate_openai(prompt, system_prompt, **kwargs)
             elif self.provider == "huggingface":
-                return await self._generate_huggingface(prompt, **kwargs)
+                result = await self._generate_huggingface(prompt, **kwargs)
             elif self.provider == "local":
-                return await self._generate_local(prompt, **kwargs)
+                result = await self._generate_local(prompt, **kwargs)
             elif self.provider == "custom":
-                return await self._generate_custom(prompt, **kwargs)
+                result = await self._generate_custom(prompt, **kwargs)
             elif self.provider == "hfs":
-                return await self._generate_hfs(prompt, **kwargs)
+                result = await self._generate_hfs(prompt, **kwargs)
             else:
                 raise ValueError(f"Unsupported provider: {self.provider}")
+            logger.info(f"[LLM] generate_text - provider: {self.provider}\n\t result: {result}")
+            return result
         except Exception as e:
             logger.error(f"[LLM] Error generating text with {self.provider}: {e}")
             raise
@@ -169,7 +172,8 @@ class LLMClient:
         headers = {}
         if self.api_key:
             headers["Authorization"] = f"Bearer {self.api_key}"
-        response = await call_endpoint_with_retry(self._client, endpoint, payload, 3, 300, headers=headers)
+        response = await call_endpoint_with_retry(self._client, endpoint, payload, 3, 500, headers=headers)
+        logger.info(f"[LLM] generate_text - provider: {self.provider}\n\t response: {response}")
         if response is not None:
             data = response.json()
             if 'response' in data:

app/main.py

@@ -430,8 +430,6 @@ async def format_search_results(question: str, matches: List[Dict[str, Any]]) ->
         f"\n\nCâu hỏi của người dùng: {question}\n"
     )
 
-    logger.info(f"[DEBUG] prompt:\n {prompt}")
-
     # Gọi LLM để sinh câu trả lời, fallback nếu lỗi
     try:
         answer = await llm_client.generate_text(prompt)