phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-06-28 02:53:30 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Make Files API an optional dependency of faiss provider

Browse source 
Because we haven't updated every template to include a Files API
provider yet, this adjusts the Files API to be an optional dependency
instead of a required one of the Faiss provider.

Signed-off-by: Ben Browning <bbrownin@redhat.com>
This commit is contained in:
Ben Browning 2025-06-12 08:51:41 -04:00
parent 8a5ea57253
commit 46da232e83
4 changed files with 14 additions and 4 deletions
Download patch file Download diff file Expand all files Collapse all files

2
llama_stack/providers/inline/vector_io/faiss/__init__.py
View file

@ -16,6 +16,6 @@ async def get_provider_impl(config: FaissVectorIOConfig, deps: dict[Api, Any]):
assert isinstance(config, FaissVectorIOConfig), f"Unexpected config type: {type(config)}" assert isinstance(config, FaissVectorIOConfig), f"Unexpected config type: {type(config)}"
impl = FaissVectorIOAdapter(config, deps[Api.inference], deps[Api.files]) impl = FaissVectorIOAdapter(config, deps[Api.inference], deps.get(Api.files, None))
await impl.initialize() await impl.initialize()
return impl return impl

2
llama_stack/providers/inline/vector_io/faiss/faiss.py
View file

@ -133,7 +133,7 @@ class FaissIndex(EmbeddingIndex):
class FaissVectorIOAdapter(OpenAIVectorStoreMixin, VectorIO, VectorDBsProtocolPrivate): class FaissVectorIOAdapter(OpenAIVectorStoreMixin, VectorIO, VectorDBsProtocolPrivate):
def __init__(self, config: FaissVectorIOConfig, inference_api: Inference, files_api: Files) -> None: def __init__(self, config: FaissVectorIOConfig, inference_api: Inference, files_api: Files | None) -> None:
self.config = config self.config = config
self.inference_api = inference_api self.inference_api = inference_api
self.files_api = files_api self.files_api = files_api

4
llama_stack/providers/registry/vector_io.py
View file

@ -24,6 +24,7 @@ def available_providers() -> list[ProviderSpec]:
config_class="llama_stack.providers.inline.vector_io.faiss.FaissVectorIOConfig", config_class="llama_stack.providers.inline.vector_io.faiss.FaissVectorIOConfig",
deprecation_warning="Please use the `inline::faiss` provider instead.", deprecation_warning="Please use the `inline::faiss` provider instead.",
api_dependencies=[Api.inference], api_dependencies=[Api.inference],
optional_api_dependencies=[Api.files],
), ),
InlineProviderSpec( InlineProviderSpec(
api=Api.vector_io, api=Api.vector_io,
@ -31,7 +32,8 @@ def available_providers() -> list[ProviderSpec]:
pip_packages=["faiss-cpu"], pip_packages=["faiss-cpu"],
module="llama_stack.providers.inline.vector_io.faiss", module="llama_stack.providers.inline.vector_io.faiss",
config_class="llama_stack.providers.inline.vector_io.faiss.FaissVectorIOConfig", config_class="llama_stack.providers.inline.vector_io.faiss.FaissVectorIOConfig",
api_dependencies=[Api.inference, Api.files], api_dependencies=[Api.inference],
optional_api_dependencies=[Api.files],
), ),
# NOTE: sqlite-vec cannot be bundled into the container image because it does not have a # NOTE: sqlite-vec cannot be bundled into the container image because it does not have a
# source distribution and the wheels are not available for all platforms. # source distribution and the wheels are not available for all platforms.

10
llama_stack/providers/utils/memory/openai_vector_store_mixin.py
View file

@ -47,7 +47,7 @@ class OpenAIVectorStoreMixin(ABC):
# These should be provided by the implementing class # These should be provided by the implementing class
openai_vector_stores: dict[str, dict[str, Any]] openai_vector_stores: dict[str, dict[str, Any]]
files_api: Files files_api: Files | None
@abstractmethod @abstractmethod
async def _save_openai_vector_store(self, store_id: str, store_info: dict[str, Any]) -> None: async def _save_openai_vector_store(self, store_id: str, store_info: dict[str, Any]) -> None:
@ -425,6 +425,14 @@ class OpenAIVectorStoreMixin(ABC):
vector_store_id=vector_store_id, vector_store_id=vector_store_id,
) )
if not self.files_api:
vector_store_file_object.status = "failed"
vector_store_file_object.last_error = VectorStoreFileLastError(
code="server_error",
message="Files API is not available",
)
return vector_store_file_object
if isinstance(chunking_strategy, VectorStoreChunkingStrategyStatic): if isinstance(chunking_strategy, VectorStoreChunkingStrategyStatic):
max_chunk_size_tokens = chunking_strategy.static.max_chunk_size_tokens max_chunk_size_tokens = chunking_strategy.static.max_chunk_size_tokens
chunk_overlap_tokens = chunking_strategy.static.chunk_overlap_tokens chunk_overlap_tokens = chunking_strategy.static.chunk_overlap_tokens