phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-08-07 11:08:20 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

feat: add provider API for listing and inspecting provider info

Browse source 
currently the `inspect` API for providers is really a `list` API. Create a new `providers` API which has a GET `providers/{provider_id}` inspect API
which returns "user friendly" configuration to the end user. Also add a GET `/providers` endpoint which returns the list of providers as `inspect/providers` does today.

This API follows CRUD and is more intuitive/RESTful.

This work is part of the RFC at #1359

Signed-off-by: Charlie Doern <cdoern@redhat.com>
This commit is contained in:
Charlie Doern 2025-03-05 15:47:49 -05:00
parent 28aade9a27
commit 520b55073a
15 changed files with 291 additions and 12 deletions
Download patch file Download diff file Expand all files Collapse all files

76
docs/_static/llama-stack-spec.html vendored
View file

@ -2642,7 +2642,81 @@
}
}
},
"/v1/inspect/providers": {
"/v1/providers": {
"get": {
"responses": {
"200": {
"description": "OK",
"content": {
"application/json": {
"schema": {
"$ref": "#/components/schemas/ListProvidersResponse"
}
}
}
},
"400": {
"$ref": "#/components/responses/BadRequest400"
},
"429": {
"$ref": "#/components/responses/TooManyRequests429"
},
"500": {
"$ref": "#/components/responses/InternalServerError500"
},
"default": {
"$ref": "#/components/responses/DefaultError"
}
},
"tags": [
"Providers"
],
"description": "",
"parameters": []
}
},
"/v1/providers/{provider_id}": {
"get": {
"responses": {
"200": {
"description": "OK",
"content": {
"application/json": {
"schema": {
"$ref": "#/components/schemas/GetProviderResponse"
}
}
}
},
"400": {
"$ref": "#/components/responses/BadRequest400"
},
"429": {
"$ref": "#/components/responses/TooManyRequests429"
},
"500": {
"$ref": "#/components/responses/InternalServerError500"
},
"default": {
"$ref": "#/components/responses/DefaultError"
}
},
"tags": [
"Providers"
],
"description": "",
"parameters": [
{
"name": "provider_id",
"in": "path",
"required": true,
"schema": {
"type": "string"
}
}
]
},
"/v1/inspect/providers": {
"get": {
"responses": {
"200": {

51
docs/_static/llama-stack-spec.yaml vendored
View file

@ -1782,6 +1782,57 @@ paths:
schema:
$ref: '#/components/schemas/RegisterModelRequest'
required: true
/v1/providers:
get:
responses:
'200':
description: OK
content:
application/json:
schema:
$ref: '#/components/schemas/ListProvidersResponse'
'400':
$ref: '#/components/responses/BadRequest400'
'429':
$ref: >-
#/components/responses/TooManyRequests429
'500':
$ref: >-
#/components/responses/InternalServerError500
default:
$ref: '#/components/responses/DefaultError'
tags:
- Providers
description: ''
parameters: []
/v1/providers/{provider_id}:
get:
responses:
'200':
description: OK
content:
application/json:
schema:
$ref: '#/components/schemas/GetProviderResponse'
'400':
$ref: '#/components/responses/BadRequest400'
'429':
$ref: >-
#/components/responses/TooManyRequests429
'500':
$ref: >-
#/components/responses/InternalServerError500
default:
$ref: '#/components/responses/DefaultError'
tags:
- Providers
description: ''
parameters:
- name: provider_id
in: path
required: true
schema:
type: string
/v1/inspect/providers:
get:
responses:

1
llama_stack/apis/datatypes.py
View file

@ -14,6 +14,7 @@ from llama_stack.schema_utils import json_schema_type
@json_schema_type
class Api(Enum):
providers = "providers"
inference = "inference"
safety = "safety"
agents = "agents"

18
llama_stack/apis/inspect/inspect.py
View file

@ -11,13 +11,6 @@ from pydantic import BaseModel
from llama_stack.schema_utils import json_schema_type, webmethod
@json_schema_type
class ProviderInfo(BaseModel):
api: str
provider_id: str
provider_type: str
@json_schema_type
class RouteInfo(BaseModel):
route: str
@ -32,14 +25,21 @@ class HealthInfo(BaseModel):
@json_schema_type
class VersionInfo(BaseModel):
version: str
class ProviderInfo(BaseModel):
api: str
provider_id: str
provider_type: str
class ListProvidersResponse(BaseModel):
data: List[ProviderInfo]
@json_schema_type
class VersionInfo(BaseModel):
version: str
class ListRoutesResponse(BaseModel):
data: List[RouteInfo]

7
llama_stack/apis/providers/__init__.py Normal file
View file

@ -0,0 +1,7 @@
# Copyright (c) Meta Platforms, Inc. and affiliates.
# All rights reserved.
#
# This source code is licensed under the terms described in the LICENSE file in
# the root directory of this source tree.
from .providers import * # noqa: F401 F403

40
llama_stack/apis/providers/providers.py Normal file
View file

@ -0,0 +1,40 @@
# Copyright (c) Meta Platforms, Inc. and affiliates.
# All rights reserved.
#
# This source code is licensed under the terms described in the LICENSE file in
# the root directory of this source tree.
from typing import List, Protocol, runtime_checkable
from pydantic import BaseModel
from llama_stack.distribution.datatypes import Provider
from llama_stack.schema_utils import json_schema_type, webmethod
@json_schema_type
class ProviderInfo(BaseModel):
api: str
provider_id: str
provider_type: str
class GetProviderResponse(BaseModel):
data: Provider | None
class ListProvidersResponse(BaseModel):
data: List[ProviderInfo]
@runtime_checkable
class Providers(Protocol):
"""
Providers API for inspecting, listing, and modifying providers and their configurations.
"""
@webmethod(route="/providers", method="GET")
async def list_providers(self) -> ListProvidersResponse: ...
@webmethod(route="/providers/{provider_id}", method="GET")
async def inspect_provider(self, provider_id: str) -> GetProviderResponse: ...

2
llama_stack/distribution/configure.py
View file

@ -62,7 +62,7 @@ def configure_api_providers(config: StackRunConfig, build_spec: DistributionSpec
if config.apis:
apis_to_serve = config.apis
else:
apis_to_serve = [a.value for a in Api if a not in (Api.telemetry, Api.inspect)]
apis_to_serve = [a.value for a in Api if a not in (Api.telemetry, Api.inspect, Api.providers)]
for api_str in apis_to_serve:
api = Api(api_str)

2
llama_stack/distribution/distribution.py
View file

@ -56,7 +56,7 @@ def builtin_automatically_routed_apis() -> List[AutoRoutedApiInfo]:
def providable_apis() -> List[Api]:
routing_table_apis = {x.routing_table_api for x in builtin_automatically_routed_apis()}
return [api for api in Api if api not in routing_table_apis and api != Api.inspect]
return [api for api in Api if api not in routing_table_apis and api != Api.inspect and api != Api.providers]
def get_provider_registry() -> Dict[Api, Dict[str, ProviderSpec]]:

59
llama_stack/distribution/providers.py Normal file
View file

@ -0,0 +1,59 @@
# Copyright (c) Meta Platforms, Inc. and affiliates.
# All rights reserved.
#
# This source code is licensed under the terms described in the LICENSE file in
# the root directory of this source tree.
from pydantic import BaseModel
from llama_stack.apis.providers import GetProviderResponse, ListProvidersResponse, ProviderInfo, Providers
from .datatypes import StackRunConfig
from .stack import redact_sensitive_fields
class ProviderImplConfig(BaseModel):
run_config: StackRunConfig
async def get_provider_impl(config, deps):
impl = ProviderImpl(config, deps)
await impl.initialize()
return impl
class ProviderImpl(Providers):
def __init__(self, config, deps):
self.config = config
self.deps = deps
async def initialize(self) -> None:
pass
async def list_providers(self) -> ListProvidersResponse:
run_config = self.config.run_config
ret = []
for api, providers in run_config.providers.items():
ret.extend(
[
ProviderInfo(
api=api,
provider_id=p.provider_id,
provider_type=p.provider_type,
)
for p in providers
]
)
return ListProvidersResponse(data=ret)
async def inspect_provider(self, provider_id: str) -> GetProviderResponse:
run_config = self.config.run_config
safe_config = StackRunConfig(**redact_sensitive_fields(run_config.model_dump()))
ret = None
for _, providers in safe_config.providers.items():
for p in providers:
if p.provider_id == provider_id:
ret = p
return GetProviderResponse(data=ret)

21
llama_stack/distribution/resolver.py
View file

@ -16,6 +16,7 @@ from llama_stack.apis.inference import Inference
from llama_stack.apis.inspect import Inspect
from llama_stack.apis.models import Models
from llama_stack.apis.post_training import PostTraining
from llama_stack.apis.providers import Providers as ProvidersAPI
from llama_stack.apis.safety import Safety
from llama_stack.apis.scoring import Scoring
from llama_stack.apis.scoring_functions import ScoringFunctions
@ -59,6 +60,7 @@ class InvalidProviderError(Exception):
def api_protocol_map() -> Dict[Api, Any]:
return {
Api.providers: ProvidersAPI,
Api.agents: Agents,
Api.inference: Inference,
Api.inspect: Inspect,
@ -247,6 +249,25 @@ def sort_providers_by_deps(
)
)
sorted_providers.append(
(
"providers",
ProviderWithSpec(
provider_id="__builtin__",
provider_type="__builtin__",
config={"run_config": run_config.model_dump()},
spec=InlineProviderSpec(
api=Api.providers,
provider_type="__builtin__",
config_class="llama_stack.distribution.providers.ProviderImplConfig",
module="llama_stack.distribution.providers",
api_dependencies=apis,
deps__=[x.value for x in apis],
),
),
)
)
logger.debug(f"Resolved {len(sorted_providers)} providers")
for api_str, provider in sorted_providers:
logger.debug(f" {api_str} => {provider.provider_id}")

1
llama_stack/distribution/server/server.py
View file

@ -368,6 +368,7 @@ def main():
apis_to_serve.add(inf.routing_table_api.value)
apis_to_serve.add("inspect")
apis_to_serve.add("providers")
for api_str in apis_to_serve:
api = Api(api_str)

2
llama_stack/distribution/stack.py
View file

@ -23,6 +23,7 @@ from llama_stack.apis.inference import Inference
from llama_stack.apis.inspect import Inspect
from llama_stack.apis.models import Models
from llama_stack.apis.post_training import PostTraining
from llama_stack.apis.providers import Providers
from llama_stack.apis.safety import Safety
from llama_stack.apis.scoring import Scoring
from llama_stack.apis.scoring_functions import ScoringFunctions
@ -44,6 +45,7 @@ logger = get_logger(name=__name__, category="core")
class LlamaStack(
Providers,
VectorDBs,
Inference,
BatchInference,

1
pyproject.toml
View file

@ -170,6 +170,7 @@ exclude = [
"^llama_stack/apis/inspect/inspect\\.py$",
"^llama_stack/apis/models/models\\.py$",
"^llama_stack/apis/post_training/post_training\\.py$",
"^llama_stack/apis/providers/providers\\.py$",
"^llama_stack/apis/resource\\.py$",
"^llama_stack/apis/safety/safety\\.py$",
"^llama_stack/apis/scoring/scoring\\.py$",

5
tests/integration/providers/__init__.py Normal file
View file

@ -0,0 +1,5 @@
# Copyright (c) Meta Platforms, Inc. and affiliates.
# All rights reserved.
#
# This source code is licensed under the terms described in the LICENSE file in
# the root directory of this source tree.

17
tests/integration/providers/test_providers.py Normal file
View file

@ -0,0 +1,17 @@
# Copyright (c) Meta Platforms, Inc. and affiliates.
# All rights reserved.
#
# This source code is licensed under the terms described in the LICENSE file in
# the root directory of this source tree.
import pytest
from llama_stack_client import LlamaStackClient
from llama_stack import LlamaStackAsLibraryClient
class TestProviders:
@pytest.mark.asyncio
def test_list(self, llama_stack_client: LlamaStackAsLibraryClient | LlamaStackClient):
provider_list = llama_stack_client.providers.list()
assert provider_list is not None