fix: extract llama-stack params from model_extra, not as explicit fields

llama_stack/apis/vector_io/vector_io.py

@@ -477,9 +477,6 @@ class OpenAICreateVectorStoreRequestWithExtraBody(BaseModel, extra="allow"):
     :param expires_after: (Optional) Expiration policy for the vector store
     :param chunking_strategy: (Optional) Strategy for splitting files into chunks
     :param metadata: Set of key-value pairs that can be attached to the vector store
-    :param embedding_model: (Optional) The embedding model to use for this vector store
-    :param embedding_dimension: (Optional) The dimension of the embedding vectors (default: 384)
-    :param provider_id: (Optional) The ID of the provider to use for this vector store
     """
 
     name: str | None = None
@@ -487,9 +484,6 @@ class OpenAICreateVectorStoreRequestWithExtraBody(BaseModel, extra="allow"):
     expires_after: dict[str, Any] | None = None
     chunking_strategy: dict[str, Any] | None = None
     metadata: dict[str, Any] | None = None
-    embedding_model: str | None = None
-    embedding_dimension: int | None = 384
-    provider_id: str | None = None
 
 
 # extra_body can be accessed via .model_extra

llama_stack/core/routers/vector_io.py

@@ -126,11 +126,15 @@ class VectorIORouter(VectorIO):
         self,
         params: Annotated[OpenAICreateVectorStoreRequestWithExtraBody, Body(...)],
     ) -> VectorStoreObject:
-        logger.debug(f"VectorIORouter.openai_create_vector_store: name={params.name}, provider_id={params.provider_id}")
+        # Extract llama-stack-specific parameters from extra_body
+        extra = params.model_extra or {}
+        embedding_model = extra.get("embedding_model")
+        embedding_dimension = extra.get("embedding_dimension", 384)
+        provider_id = extra.get("provider_id")
+
+        logger.debug(f"VectorIORouter.openai_create_vector_store: name={params.name}, provider_id={provider_id}")
 
         # If no embedding model is provided, use the first available one
-        embedding_model = params.embedding_model
-        embedding_dimension = params.embedding_dimension
         if embedding_model is None:
             embedding_model_info = await self._get_first_embedding_model()
             if embedding_model_info is None:
@@ -143,22 +147,13 @@ class VectorIORouter(VectorIO):
             vector_db_id=vector_db_id,
             embedding_model=embedding_model,
             embedding_dimension=embedding_dimension,
-            provider_id=params.provider_id,
+            provider_id=provider_id,
             provider_vector_db_id=vector_db_id,
             vector_db_name=params.name,
         )
         provider = await self.routing_table.get_provider_impl(registered_vector_db.identifier)
 
-        # Update params with resolved values
+        # Pass params as-is to provider - it will extract what it needs from model_extra
-        params.embedding_model = embedding_model
-        params.embedding_dimension = embedding_dimension
-        params.provider_id = registered_vector_db.provider_id
-
-        # Add provider_vector_db_id to extra_body if not already there
-        if params.model_extra is None:
-            params.model_extra = {}
-        params.model_extra["provider_vector_db_id"] = registered_vector_db.provider_resource_id
-
         return await provider.openai_create_vector_store(params)
 
     async def openai_list_vector_stores(

llama_stack/providers/utils/memory/openai_vector_store_mixin.py

@@ -347,30 +347,32 @@ class OpenAIVectorStoreMixin(ABC):
         """Creates a vector store."""
         created_at = int(time.time())
 
-        # Extract provider_vector_db_id from extra_body if present
+        # Extract llama-stack-specific parameters from extra_body
-        provider_vector_db_id = None
+        extra = params.model_extra or {}
-        if params.model_extra and "provider_vector_db_id" in params.model_extra:
+        provider_vector_db_id = extra.get("provider_vector_db_id")
-            provider_vector_db_id = params.model_extra["provider_vector_db_id"]
+        embedding_model = extra.get("embedding_model")
+        embedding_dimension = extra.get("embedding_dimension", 384)
+        provider_id = extra.get("provider_id")
 
         # Derive the canonical vector_db_id (allow override, else generate)
         vector_db_id = provider_vector_db_id or generate_object_id("vector_store", lambda: f"vs_{uuid.uuid4()}")
 
-        if params.provider_id is None:
+        if provider_id is None:
             raise ValueError("Provider ID is required")
 
-        if params.embedding_model is None:
+        if embedding_model is None:
             raise ValueError("Embedding model is required")
 
         # Embedding dimension is required (defaulted to 384 if not provided)
-        if params.embedding_dimension is None:
+        if embedding_dimension is None:
             raise ValueError("Embedding dimension is required")
 
         # Register the VectorDB backing this vector store
         vector_db = VectorDB(
             identifier=vector_db_id,
-            embedding_dimension=params.embedding_dimension,
+            embedding_dimension=embedding_dimension,
-            embedding_model=params.embedding_model,
+            embedding_model=embedding_model,
-            provider_id=params.provider_id,
+            provider_id=provider_id,
             provider_resource_id=vector_db_id,
             vector_db_name=params.name,
         )
@@ -404,8 +406,8 @@ class OpenAIVectorStoreMixin(ABC):
 
         # Add provider information to metadata if provided
         metadata = params.metadata or {}
-        if params.provider_id:
+        if provider_id:
-            metadata["provider_id"] = params.provider_id
+            metadata["provider_id"] = provider_id
         if provider_vector_db_id:
             metadata["provider_vector_db_id"] = provider_vector_db_id
         store_info["metadata"] = metadata