chore(tests): fix responses and vector_io tests

llama_stack/core/routers/inference.py

@@ -65,7 +65,7 @@ from llama_stack.providers.datatypes import HealthResponse, HealthStatus, Routin
 from llama_stack.providers.utils.inference.inference_store import InferenceStore
 from llama_stack.providers.utils.telemetry.tracing import get_current_span
 
-logger = get_logger(name=__name__, category="core")
+logger = get_logger(name=__name__, category="inference")
 
 
 class InferenceRouter(Inference):
@@ -854,4 +854,5 @@ class InferenceRouter(Inference):
                     model=model.identifier,
                     object="chat.completion",
                 )
+                logger.debug(f"InferenceRouter.completion_response: {final_response}")
                 await self.store.store_chat_completion(final_response, messages)

llama_stack/core/routing_tables/models.py

@@ -63,6 +63,8 @@ class ModelsRoutingTable(CommonRoutingTableImpl, Models):
 
     async def get_provider_impl(self, model_id: str) -> Any:
         model = await lookup_model(self, model_id)
+        if model.provider_id not in self.impls_by_provider_id:
+            raise ValueError(f"Provider {model.provider_id} not found in the routing table")
         return self.impls_by_provider_id[model.provider_id]
 
     async def register_model(

llama_stack/log.py

@@ -32,6 +32,7 @@ CATEGORIES = [
     "tools",
     "client",
     "telemetry",
+    "openai_responses",
 ]
 
 # Initialize category levels with default level

llama_stack/models/llama/llama3/chat_format.py

@@ -236,6 +236,7 @@ class ChatFormat:
                     arguments_json=json.dumps(tool_arguments),
                 )
             )
+            content = ""
 
         return RawMessage(
             role="assistant",

llama_stack/providers/inline/agents/meta_reference/openai_responses.py

@@ -488,8 +488,12 @@ class OpenAIResponsesImpl:
             # Convert collected chunks to complete response
             if chat_response_tool_calls:
                 tool_calls = [chat_response_tool_calls[i] for i in sorted(chat_response_tool_calls.keys())]
+
+                # when there are tool calls, we need to clear the content
+                chat_response_content = []
             else:
                 tool_calls = None
+
             assistant_message = OpenAIAssistantMessageParam(
                 content="".join(chat_response_content),
                 tool_calls=tool_calls,

llama_stack/providers/remote/inference/fireworks/fireworks.py

@@ -235,6 +235,7 @@ class FireworksInferenceAdapter(ModelRegistryHelper, Inference, NeedsRequestProv
 
         llama_model = self.get_llama_model(request.model)
         if isinstance(request, ChatCompletionRequest):
+            # TODO: tools are never added to the request, so we need to add them here
             if media_present or not llama_model:
                 input_dict["messages"] = [
                     await convert_message_to_openai_dict(m, download=True) for m in request.messages
@@ -378,6 +379,7 @@ class FireworksInferenceAdapter(ModelRegistryHelper, Inference, NeedsRequestProv
         # Fireworks chat completions OpenAI-compatible API does not support
         # tool calls properly.
         llama_model = self.get_llama_model(model_obj.provider_resource_id)
+
         if llama_model:
             return await OpenAIChatCompletionToLlamaStackMixin.openai_chat_completion(
                 self,
@@ -431,4 +433,5 @@ class FireworksInferenceAdapter(ModelRegistryHelper, Inference, NeedsRequestProv
             user=user,
         )
 
+        logger.debug(f"fireworks params: {params}")
         return await self._get_openai_client().chat.completions.create(model=model_obj.provider_resource_id, **params)

llama_stack/providers/utils/memory/openai_vector_store_mixin.py

@@ -614,7 +614,7 @@ class OpenAIVectorStoreMixin(ABC):
                 )
                 vector_store_file_object.status = "completed"
         except Exception as e:
-            logger.error(f"Error attaching file to vector store: {e}")
+            logger.exception("Error attaching file to vector store")
             vector_store_file_object.status = "failed"
             vector_store_file_object.last_error = VectorStoreFileLastError(
                 code="server_error",