Merge branch 'main' into fix/issue-2584-llama4-tool-calling

docs/_static/llama-stack-spec.html

@@ -11340,6 +11340,9 @@
                     },
                     "embedding_dimension": {
                         "type": "integer"
+                    },
+                    "vector_db_name": {
+                        "type": "string"
                     }
                 },
                 "additionalProperties": false,
@@ -13590,10 +13593,6 @@
                     "provider_id": {
                         "type": "string",
                         "description": "The ID of the provider to use for this vector store."
-                    },
-                    "provider_vector_db_id": {
-                        "type": "string",
-                        "description": "The provider-specific vector database ID."
                     }
                 },
                 "additionalProperties": false,
@@ -15634,6 +15633,10 @@
                         "type": "string",
                         "description": "The identifier of the provider."
                     },
+                    "vector_db_name": {
+                        "type": "string",
+                        "description": "The name of the vector database."
+                    },
                     "provider_vector_db_id": {
                         "type": "string",
                         "description": "The identifier of the vector database in the provider."

docs/_static/llama-stack-spec.yaml

@@ -7984,6 +7984,8 @@ components:
           type: string
         embedding_dimension:
           type: integer
+        vector_db_name:
+          type: string
       additionalProperties: false
       required:
         - identifier
@@ -9494,10 +9496,6 @@ components:
           type: string
           description: >-
             The ID of the provider to use for this vector store.
-        provider_vector_db_id:
-          type: string
-          description: >-
-            The provider-specific vector database ID.
       additionalProperties: false
       required:
         - name
@@ -10945,6 +10943,9 @@ components:
         provider_id:
           type: string
           description: The identifier of the provider.
+        vector_db_name:
+          type: string
+          description: The name of the vector database.
         provider_vector_db_id:
           type: string
           description: >-

llama_stack/apis/vector_dbs/vector_dbs.py

@@ -19,6 +19,7 @@ class VectorDB(Resource):
 
     embedding_model: str
     embedding_dimension: int
+    vector_db_name: str | None = None
 
     @property
     def vector_db_id(self) -> str:
@@ -70,6 +71,7 @@ class VectorDBs(Protocol):
         embedding_model: str,
         embedding_dimension: int | None = 384,
         provider_id: str | None = None,
+        vector_db_name: str | None = None,
         provider_vector_db_id: str | None = None,
     ) -> VectorDB:
         """Register a vector database.
@@ -78,6 +80,7 @@ class VectorDBs(Protocol):
         :param embedding_model: The embedding model to use.
         :param embedding_dimension: The dimension of the embedding model.
         :param provider_id: The identifier of the provider.
+        :param vector_db_name: The name of the vector database.
         :param provider_vector_db_id: The identifier of the vector database in the provider.
         :returns: A VectorDB.
         """

llama_stack/apis/vector_io/vector_io.py

@@ -346,7 +346,6 @@ class VectorIO(Protocol):
         embedding_model: str | None = None,
         embedding_dimension: int | None = 384,
         provider_id: str | None = None,
-        provider_vector_db_id: str | None = None,
     ) -> VectorStoreObject:
         """Creates a vector store.
 
@@ -358,7 +357,6 @@ class VectorIO(Protocol):
         :param embedding_model: The embedding model to use for this vector store.
         :param embedding_dimension: The dimension of the embedding vectors (default: 384).
         :param provider_id: The ID of the provider to use for this vector store.
-        :param provider_vector_db_id: The provider-specific vector database ID.
         :returns: A VectorStoreObject representing the created vector store.
         """
         ...

llama_stack/distribution/configure.py

@@ -17,7 +17,7 @@ from llama_stack.distribution.distribution import (
     builtin_automatically_routed_apis,
     get_provider_registry,
 )
-from llama_stack.distribution.stack import replace_env_vars
+from llama_stack.distribution.stack import cast_image_name_to_string, replace_env_vars
 from llama_stack.distribution.utils.config_dirs import EXTERNAL_PROVIDERS_DIR
 from llama_stack.distribution.utils.dynamic import instantiate_class_type
 from llama_stack.distribution.utils.prompt_for_config import prompt_for_config
@@ -164,7 +164,8 @@ def upgrade_from_routing_table(
 def parse_and_maybe_upgrade_config(config_dict: dict[str, Any]) -> StackRunConfig:
     version = config_dict.get("version", None)
     if version == LLAMA_STACK_RUN_CONFIG_VERSION:
-        return StackRunConfig(**replace_env_vars(config_dict))
+        processed_config_dict = replace_env_vars(config_dict)
+        return StackRunConfig(**cast_image_name_to_string(processed_config_dict))
 
     if "routing_table" in config_dict:
         logger.info("Upgrading config...")
@@ -175,4 +176,5 @@ def parse_and_maybe_upgrade_config(config_dict: dict[str, Any]) -> StackRunConfi
     if not config_dict.get("external_providers_dir", None):
         config_dict["external_providers_dir"] = EXTERNAL_PROVIDERS_DIR
 
-    return StackRunConfig(**replace_env_vars(config_dict))
+    processed_config_dict = replace_env_vars(config_dict)
+    return StackRunConfig(**cast_image_name_to_string(processed_config_dict))

llama_stack/distribution/routers/vector_io.py

@@ -5,6 +5,7 @@
 # the root directory of this source tree.
 
 import asyncio
+import uuid
 from typing import Any
 
 from llama_stack.apis.common.content_types import (
@@ -81,6 +82,7 @@ class VectorIORouter(VectorIO):
         embedding_model: str,
         embedding_dimension: int | None = 384,
         provider_id: str | None = None,
+        vector_db_name: str | None = None,
         provider_vector_db_id: str | None = None,
     ) -> None:
         logger.debug(f"VectorIORouter.register_vector_db: {vector_db_id}, {embedding_model}")
@@ -89,6 +91,7 @@ class VectorIORouter(VectorIO):
             embedding_model,
             embedding_dimension,
             provider_id,
+            vector_db_name,
             provider_vector_db_id,
         )
 
@@ -123,7 +126,6 @@ class VectorIORouter(VectorIO):
         embedding_model: str | None = None,
         embedding_dimension: int | None = None,
         provider_id: str | None = None,
-        provider_vector_db_id: str | None = None,
     ) -> VectorStoreObject:
         logger.debug(f"VectorIORouter.openai_create_vector_store: name={name}, provider_id={provider_id}")
 
@@ -135,17 +137,17 @@ class VectorIORouter(VectorIO):
             embedding_model, embedding_dimension = embedding_model_info
             logger.info(f"No embedding model specified, using first available: {embedding_model}")
 
-        vector_db_id = name
+        vector_db_id = f"vs_{uuid.uuid4()}"
         registered_vector_db = await self.routing_table.register_vector_db(
-            vector_db_id,
+            vector_db_id=vector_db_id,
-            embedding_model,
+            embedding_model=embedding_model,
-            embedding_dimension,
+            embedding_dimension=embedding_dimension,
-            provider_id,
+            provider_id=provider_id,
-            provider_vector_db_id,
+            provider_vector_db_id=vector_db_id,
+            vector_db_name=name,
         )
-
         return await self.routing_table.get_provider_impl(registered_vector_db.identifier).openai_create_vector_store(
-            vector_db_id,
+            name=name,
             file_ids=file_ids,
             expires_after=expires_after,
             chunking_strategy=chunking_strategy,

llama_stack/distribution/routing_tables/vector_dbs.py

@@ -36,6 +36,7 @@ class VectorDBsRoutingTable(CommonRoutingTableImpl, VectorDBs):
         embedding_dimension: int | None = 384,
         provider_id: str | None = None,
         provider_vector_db_id: str | None = None,
+        vector_db_name: str | None = None,
     ) -> VectorDB:
         if provider_vector_db_id is None:
             provider_vector_db_id = vector_db_id
@@ -62,6 +63,7 @@ class VectorDBsRoutingTable(CommonRoutingTableImpl, VectorDBs):
             "provider_resource_id": provider_vector_db_id,
             "embedding_model": embedding_model,
             "embedding_dimension": model.metadata["embedding_dimension"],
+            "vector_db_name": vector_db_name,
         }
         vector_db = TypeAdapter(VectorDBWithOwner).validate_python(vector_db_data)
         await self.register_object(vector_db)

llama_stack/distribution/server/server.py

@@ -47,6 +47,7 @@ from llama_stack.distribution.server.routes import (
     initialize_route_impls,
 )
 from llama_stack.distribution.stack import (
+    cast_image_name_to_string,
     construct_stack,
     replace_env_vars,
     validate_env_pair,
@@ -439,7 +440,7 @@ def main(args: argparse.Namespace | None = None):
                     logger.error(f"Error: {str(e)}")
                     sys.exit(1)
         config = replace_env_vars(config_contents)
-        config = StackRunConfig(**config)
+        config = StackRunConfig(**cast_image_name_to_string(config))
 
     # now that the logger is initialized, print the line about which type of config we are using.
     logger.info(log_line)

llama_stack/distribution/stack.py

@@ -267,6 +267,13 @@ def _convert_string_to_proper_type(value: str) -> Any:
     return value
 
 
+def cast_image_name_to_string(config_dict: dict[str, Any]) -> dict[str, Any]:
+    """Ensure that any value for a key 'image_name' in a config_dict is a string"""
+    if "image_name" in config_dict and config_dict["image_name"] is not None:
+        config_dict["image_name"] = str(config_dict["image_name"])
+    return config_dict
+
+
 def validate_env_pair(env_pair: str) -> tuple[str, str]:
     """Validate and split an environment variable key-value pair."""
     try:

llama_stack/providers/remote/vector_io/chroma/chroma.py

@@ -217,7 +217,6 @@ class ChromaVectorIOAdapter(VectorIO, VectorDBsProtocolPrivate):
         embedding_model: str | None = None,
         embedding_dimension: int | None = 384,
         provider_id: str | None = None,
-        provider_vector_db_id: str | None = None,
     ) -> VectorStoreObject:
         raise NotImplementedError("OpenAI Vector Stores API is not supported in Chroma")
 

llama_stack/providers/remote/vector_io/qdrant/qdrant.py

@@ -214,7 +214,6 @@ class QdrantVectorIOAdapter(VectorIO, VectorDBsProtocolPrivate):
         embedding_model: str | None = None,
         embedding_dimension: int | None = 384,
         provider_id: str | None = None,
-        provider_vector_db_id: str | None = None,
     ) -> VectorStoreObject:
         raise NotImplementedError("OpenAI Vector Stores API is not supported in Qdrant")
 

llama_stack/providers/utils/memory/openai_vector_store_mixin.py

@@ -172,8 +172,9 @@ class OpenAIVectorStoreMixin(ABC):
         provider_vector_db_id: str | None = None,
     ) -> VectorStoreObject:
         """Creates a vector store."""
-        store_id = name or str(uuid.uuid4())
         created_at = int(time.time())
+        # Derive the canonical vector_db_id (allow override, else generate)
+        vector_db_id = provider_vector_db_id or f"vs_{uuid.uuid4()}"
 
         if provider_id is None:
             raise ValueError("Provider ID is required")
@@ -181,19 +182,19 @@ class OpenAIVectorStoreMixin(ABC):
         if embedding_model is None:
             raise ValueError("Embedding model is required")
 
-        # Use provided embedding dimension or default to 384
+        # Embedding dimension is required (defaulted to 384 if not provided)
         if embedding_dimension is None:
             raise ValueError("Embedding dimension is required")
 
-        provider_vector_db_id = provider_vector_db_id or store_id
+        # Register the VectorDB backing this vector store
         vector_db = VectorDB(
-            identifier=store_id,
+            identifier=vector_db_id,
             embedding_dimension=embedding_dimension,
             embedding_model=embedding_model,
             provider_id=provider_id,
-            provider_resource_id=provider_vector_db_id,
+            provider_resource_id=vector_db_id,
+            vector_db_name=name,
         )
-        # Register the vector DB
         await self.register_vector_db(vector_db)
 
         # Create OpenAI vector store metadata
@@ -207,11 +208,11 @@ class OpenAIVectorStoreMixin(ABC):
             in_progress=0,
             total=0,
         )
-        store_info = {
+        store_info: dict[str, Any] = {
-            "id": store_id,
+            "id": vector_db_id,
             "object": "vector_store",
             "created_at": created_at,
-            "name": store_id,
+            "name": name,
             "usage_bytes": 0,
             "file_counts": file_counts.model_dump(),
             "status": status,
@@ -231,18 +232,18 @@ class OpenAIVectorStoreMixin(ABC):
         store_info["metadata"] = metadata
 
         # Save to persistent storage (provider-specific)
-        await self._save_openai_vector_store(store_id, store_info)
+        await self._save_openai_vector_store(vector_db_id, store_info)
 
         # Store in memory cache
-        self.openai_vector_stores[store_id] = store_info
+        self.openai_vector_stores[vector_db_id] = store_info
 
         # Now that our vector store is created, attach any files that were provided
         file_ids = file_ids or []
-        tasks = [self.openai_attach_file_to_vector_store(store_id, file_id) for file_id in file_ids]
+        tasks = [self.openai_attach_file_to_vector_store(vector_db_id, file_id) for file_id in file_ids]
         await asyncio.gather(*tasks)
 
         # Get the updated store info and return it
-        store_info = self.openai_vector_stores[store_id]
+        store_info = self.openai_vector_stores[vector_db_id]
         return VectorStoreObject.model_validate(store_info)
 
     async def openai_list_vector_stores(

tests/integration/vector_io/test_openai_vector_stores.py

@@ -821,6 +821,59 @@ def test_openai_vector_store_update_file(compat_client_with_empty_stores, client
     assert retrieved_file.attributes["foo"] == "baz"
 
 
+def test_create_vector_store_files_duplicate_vector_store_name(compat_client_with_empty_stores, client_with_models):
+    """
+    This test confirms that client.vector_stores.create() creates a unique ID
+    """
+    skip_if_provider_doesnt_support_openai_vector_stores(client_with_models)
+    skip_if_provider_doesnt_support_openai_vector_store_files_api(client_with_models)
+
+    if isinstance(compat_client_with_empty_stores, LlamaStackClient):
+        pytest.skip("Vector Store Files create is not yet supported with LlamaStackClient")
+
+    compat_client = compat_client_with_empty_stores
+
+    # Create a vector store with files
+    file_ids = []
+    for i in range(3):
+        with BytesIO(f"This is a test file {i}".encode()) as file_buffer:
+            file_buffer.name = f"openai_test_{i}.txt"
+            file = compat_client.files.create(file=file_buffer, purpose="assistants")
+        file_ids.append(file.id)
+
+    vector_store = compat_client.vector_stores.create(
+        name="test_store_with_files",
+    )
+    assert vector_store.file_counts.completed == 0
+    assert vector_store.file_counts.total == 0
+    assert vector_store.file_counts.cancelled == 0
+    assert vector_store.file_counts.failed == 0
+    assert vector_store.file_counts.in_progress == 0
+
+    vector_store2 = compat_client.vector_stores.create(
+        name="test_store_with_files",
+    )
+
+    vector_stores_list = compat_client.vector_stores.list()
+    assert len(vector_stores_list.data) == 2
+
+    created_file = compat_client.vector_stores.files.create(
+        vector_store_id=vector_store.id,
+        file_id=file_ids[0],
+    )
+    assert created_file.status == "completed"
+
+    _ = compat_client.vector_stores.delete(vector_store2.id)
+    created_file_from_non_deleted_vector_store = compat_client.vector_stores.files.create(
+        vector_store_id=vector_store.id,
+        file_id=file_ids[1],
+    )
+    assert created_file_from_non_deleted_vector_store.status == "completed"
+
+    vector_stores_list_post_delete = compat_client.vector_stores.list()
+    assert len(vector_stores_list_post_delete.data) == 1
+
+
 @pytest.mark.skip(reason="Client library needs to be scaffolded to support search_mode parameter")
 def test_openai_vector_store_search_modes():
     """Test OpenAI vector store search with different search modes.

tests/unit/cli/test_stack_config.py

@@ -15,6 +15,37 @@ from llama_stack.distribution.configure import (
 )
 
 
+@pytest.fixture
+def config_with_image_name_int():
+    return yaml.safe_load(
+        f"""
+        version: {LLAMA_STACK_RUN_CONFIG_VERSION}
+        image_name: 1234
+        apis_to_serve: []
+        built_at: {datetime.now().isoformat()}
+        providers:
+          inference:
+            - provider_id: provider1
+              provider_type: inline::meta-reference
+              config: {{}}
+          safety:
+            - provider_id: provider1
+              provider_type: inline::meta-reference
+              config:
+                llama_guard_shield:
+                  model: Llama-Guard-3-1B
+                  excluded_categories: []
+                  disable_input_check: false
+                  disable_output_check: false
+                enable_prompt_guard: false
+          memory:
+            - provider_id: provider1
+              provider_type: inline::meta-reference
+              config: {{}}
+    """
+    )
+
+
 @pytest.fixture
 def up_to_date_config():
     return yaml.safe_load(
@@ -125,3 +156,8 @@ def test_parse_and_maybe_upgrade_config_old_format(old_config):
 def test_parse_and_maybe_upgrade_config_invalid(invalid_config):
     with pytest.raises(KeyError):
         parse_and_maybe_upgrade_config(invalid_config)
+
+
+def test_parse_and_maybe_upgrade_config_image_name_int(config_with_image_name_int):
+    result = parse_and_maybe_upgrade_config(config_with_image_name_int)
+    assert isinstance(result.image_name, str)