Merge remote-tracking branch 'upstream/main' into api-pkg

Signed-off-by: Charlie Doern <cdoern@redhat.com>

src/llama_stack/core/build.py

@@ -4,7 +4,6 @@
 # This source code is licensed under the terms described in the LICENSE file in
 # the root directory of this source tree.
 
-import importlib.resources
 import sys
 
 from llama_stack_api.providers.datatypes import Api
@@ -13,9 +12,6 @@ from termcolor import cprint
 
 from llama_stack.core.datatypes import BuildConfig
 from llama_stack.core.distribution import get_provider_registry
-from llama_stack.core.external import load_external_apis
-from llama_stack.core.utils.exec import run_command
-from llama_stack.core.utils.image_types import LlamaStackImageType
 from llama_stack.distributions.template import DistributionTemplate
 from llama_stack.log import get_logger
 
@@ -101,64 +97,3 @@ def print_pip_install_help(config: BuildConfig):
     for special_dep in special_deps:
         cprint(f"uv pip install {special_dep}", color="yellow", file=sys.stderr)
     print()
-
-
-def build_image(
-    build_config: BuildConfig,
-    image_name: str,
-    distro_or_config: str,
-    run_config: str | None = None,
-):
-    container_base = build_config.distribution_spec.container_image or "python:3.12-slim"
-
-    normal_deps, special_deps, external_provider_deps = get_provider_dependencies(build_config)
-    normal_deps += SERVER_DEPENDENCIES
-    if build_config.external_apis_dir:
-        external_apis = load_external_apis(build_config)
-        if external_apis:
-            for _, api_spec in external_apis.items():
-                normal_deps.extend(api_spec.pip_packages)
-
-    if build_config.image_type == LlamaStackImageType.CONTAINER.value:
-        script = str(importlib.resources.files("llama_stack") / "core/build_container.sh")
-        args = [
-            script,
-            "--distro-or-config",
-            distro_or_config,
-            "--image-name",
-            image_name,
-            "--container-base",
-            container_base,
-            "--normal-deps",
-            " ".join(normal_deps),
-        ]
-        # When building from a config file (not a template), include the run config path in the
-        # build arguments
-        if run_config is not None:
-            args.extend(["--run-config", run_config])
-    else:
-        script = str(importlib.resources.files("llama_stack") / "core/build_venv.sh")
-        args = [
-            script,
-            "--env-name",
-            str(image_name),
-            "--normal-deps",
-            " ".join(normal_deps),
-        ]
-
-    # Always pass both arguments, even if empty, to maintain consistent positional arguments
-    if special_deps:
-        args.extend(["--optional-deps", "#".join(special_deps)])
-    if external_provider_deps:
-        args.extend(
-            ["--external-provider-deps", "#".join(external_provider_deps)]
-        )  # the script will install external provider module, get its deps, and install those too.
-
-    return_code = run_command(args)
-
-    if return_code != 0:
-        log.error(
-            f"Failed to build target {image_name} with return code {return_code}",
-        )
-
-    return return_code

src/llama_stack/core/library_client.py

@@ -389,6 +389,12 @@ class AsyncLlamaStackAsLibraryClient(AsyncLlamaStackClient):
         matched_func, path_params, route_path, webmethod = find_matching_route(options.method, path, self.route_impls)
         body |= path_params
 
+        # Pass through params that aren't already handled as path params
+        if options.params:
+            extra_query_params = {k: v for k, v in options.params.items() if k not in path_params}
+            if extra_query_params:
+                body["extra_query"] = extra_query_params
+
         body, field_names = self._handle_file_uploads(options, body)
 
         body = self._convert_body(matched_func, body, exclude_params=set(field_names))

src/llama_stack/core/routers/vector_io.py

@@ -247,6 +247,13 @@ class VectorIORouter(VectorIO):
         metadata: dict[str, Any] | None = None,
     ) -> VectorStoreObject:
         logger.debug(f"VectorIORouter.openai_update_vector_store: {vector_store_id}")
+
+        # Check if provider_id is being changed (not supported)
+        if metadata and "provider_id" in metadata:
+            current_store = await self.routing_table.get_object_by_identifier("vector_store", vector_store_id)
+            if current_store and current_store.provider_id != metadata["provider_id"]:
+                raise ValueError("provider_id cannot be changed after vector store creation")
+
         provider = await self.routing_table.get_provider_impl(vector_store_id)
         return await provider.openai_update_vector_store(
             vector_store_id=vector_store_id,
@@ -338,12 +345,19 @@ class VectorIORouter(VectorIO):
         self,
         vector_store_id: str,
         file_id: str,
+        include_embeddings: bool | None = False,
+        include_metadata: bool | None = False,
     ) -> VectorStoreFileContentResponse:
-        logger.debug(f"VectorIORouter.openai_retrieve_vector_store_file_contents: {vector_store_id}, {file_id}")
-        provider = await self.routing_table.get_provider_impl(vector_store_id)
-        return await provider.openai_retrieve_vector_store_file_contents(
+        logger.debug(
+            f"VectorIORouter.openai_retrieve_vector_store_file_contents: {vector_store_id}, {file_id}, "
+            f"include_embeddings={include_embeddings}, include_metadata={include_metadata}"
+        )
+
+        return await self.routing_table.openai_retrieve_vector_store_file_contents(
             vector_store_id=vector_store_id,
             file_id=file_id,
+            include_embeddings=include_embeddings,
+            include_metadata=include_metadata,
         )
 
     async def openai_update_vector_store_file(

src/llama_stack/core/routing_tables/vector_stores.py

@@ -196,12 +196,17 @@ class VectorStoresRoutingTable(CommonRoutingTableImpl):
         self,
         vector_store_id: str,
         file_id: str,
+        include_embeddings: bool | None = False,
+        include_metadata: bool | None = False,
     ) -> VectorStoreFileContentResponse:
         await self.assert_action_allowed("read", "vector_store", vector_store_id)
+
         provider = await self.get_provider_impl(vector_store_id)
         return await provider.openai_retrieve_vector_store_file_contents(
             vector_store_id=vector_store_id,
             file_id=file_id,
+            include_embeddings=include_embeddings,
+            include_metadata=include_metadata,
         )
 
     async def openai_update_vector_store_file(