phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-07-27 06:28:50 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

refactor: Align build and run provider datatypes

Browse source 
introduce the concept of a `module` for users to specify for a provider upon build time.
In order to facilitate this, align the build and run spec to use `Provider` class rather than the stringified provider_type that build currently uses.

Signed-off-by: Charlie Doern <cdoern@redhat.com>
This commit is contained in:
Charlie Doern 2025-07-06 19:56:18 -04:00
parent 21bae296f2
commit 233f8c81bf
6 changed files with 48 additions and 38 deletions
Download patch file Download diff file Expand all files Collapse all files

8
llama_stack/distribution/library_client.py
View file

@ -249,15 +249,9 @@ class AsyncLlamaStackAsLibraryClient(AsyncLlamaStackClient):
file=sys.stderr,
)
if self.config_path_or_template_name.endswith(".yaml"):
# Convert Provider objects to their types
provider_types: dict[str, str | list[str]] = {}
for api, providers in self.config.providers.items():
types = [p.provider_type for p in providers]
# Convert single-item lists to strings
provider_types[api] = types[0] if len(types) == 1 else types
build_config = BuildConfig(
distribution_spec=DistributionSpec(
providers=provider_types,
providers=self.config.providers,
),
external_providers_dir=self.config.external_providers_dir,
)