mirror of
https://github.com/BerriAI/litellm.git
synced 2026-10-12 03:48:56 +00:00
1bc85fad40
* refactor(python-bridge): ship a signature base and read the resolved call NativeCall carries base (positionals by name plus signature defaults) instead of the fully bound dict. resolved lays kwargs over base, which is what bound held, so every pre-hook read and every route host keeps seeing the same values. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(bedrock): build the transcription NativeCall with an empty base Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * test(dispatch): read the resolved call instead of bound Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * refactor(host-python): rename effective to effective_py_args and note the shallow copy Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
87 lines
3.3 KiB
Python
87 lines
3.3 KiB
Python
from __future__ import annotations
|
|
|
|
from collections.abc import Awaitable, Callable
|
|
from typing import Final
|
|
|
|
import pytest
|
|
from pydantic import TypeAdapter
|
|
|
|
import litellm
|
|
from litellm.embeddings import dispatch
|
|
from litellm.rust_bridge.bindings import NativeBinding
|
|
from litellm.rust_bridge.catalog import Route, RouteRule, Rules
|
|
from litellm.rust_bridge.configuration import Rollout
|
|
from litellm.rust_bridge.public_call import NativeCall, native_call_hook
|
|
from litellm.types.utils import EmbeddingResponse
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_public_embedding_calls_keep_the_python_result() -> None:
|
|
vector: Final = [0.1, 0.2]
|
|
|
|
sync_response: Final = litellm.embedding(model="openai/test-model", input="hello", mock_response=vector)
|
|
async_response: Final = await litellm.aembedding(model="openai/test-model", input="hello", mock_response=vector)
|
|
|
|
assert isinstance(sync_response, EmbeddingResponse)
|
|
rows: Final = TypeAdapter(list[dict[str, object]])
|
|
assert rows.validate_python(sync_response.model_dump()["data"])[0]["embedding"] == vector
|
|
assert rows.validate_python(async_response.model_dump()["data"])[0]["embedding"] == vector
|
|
|
|
|
|
def test_sync_embedding_request_projects_public_arguments() -> None:
|
|
rules: Final[Rules] = (RouteRule(Route.EMBEDDINGS, Rollout.RUST_REQUIRED),)
|
|
expected: Final = EmbeddingResponse(model="test-model", data=[])
|
|
|
|
def native(request: NativeCall) -> EmbeddingResponse:
|
|
assert request.resolved["model"] == "test-model"
|
|
assert request.resolved["input"] == "hello"
|
|
assert request.resolved["custom_llm_provider"] == "openai"
|
|
return expected
|
|
|
|
binding: Final[NativeBinding[Callable[[NativeCall], EmbeddingResponse]]] = NativeBinding(
|
|
"embedding", validate=lambda _: None
|
|
)
|
|
binding.override(native)
|
|
response: Final = dispatch._DISPATCH.run( # pyright: ignore[reportPrivateUsage] # test an explicit route decision
|
|
("test-model", "hello"),
|
|
{"custom_llm_provider": "openai", "dimensions": 8},
|
|
python=lambda *args, **kwargs: pytest.fail("required native route must handle this call"),
|
|
binding=binding,
|
|
native=native_call_hook,
|
|
rules=rules,
|
|
)
|
|
|
|
assert response is expected
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_async_embedding_falls_back_after_native_declines() -> None:
|
|
from litellm.rust_bridge.bindings import native_exception_types
|
|
|
|
native_types: Final = native_exception_types()
|
|
if native_types is None:
|
|
pytest.skip("native bridge is unavailable")
|
|
declined, _ = native_types
|
|
expected: Final = EmbeddingResponse(model="test-model", data=[])
|
|
rules: Final[Rules] = (RouteRule(Route.EMBEDDINGS, Rollout.RUST_OPT_OUT),)
|
|
|
|
async def native(request: NativeCall) -> EmbeddingResponse:
|
|
raise declined("unsupported")
|
|
|
|
async def python(*args: object, **kwargs: object) -> EmbeddingResponse:
|
|
return expected
|
|
|
|
binding: Final[NativeBinding[Callable[[NativeCall], Awaitable[EmbeddingResponse]]]] = NativeBinding(
|
|
"aembedding", validate=lambda _: None
|
|
)
|
|
binding.override(native)
|
|
response: Final = await dispatch._ADISPATCH.arun( # pyright: ignore[reportPrivateUsage] # test an explicit route decision
|
|
("test-model", "hello"),
|
|
{},
|
|
python=python,
|
|
binding=binding,
|
|
native=native_call_hook,
|
|
rules=rules,
|
|
)
|
|
|
|
assert response is expected
|