Overview
Standard unit test suite for validating Embeddings implementations, verifying initialization and environment variable configuration.
Description
This module in the langchain-tests (standard-tests) package defines two test classes: EmbeddingsTests, a base class providing the embeddings_class abstract property and a model fixture, and EmbeddingsUnitTests, which extends it with concrete unit tests. The unit tests verify basic model initialization from constructor parameters and initialization from environment variables. The init_from_env_params property can be overridden to enable environment variable initialization testing, providing a tuple of environment variables, init args, and expected attribute values.
Usage
Import EmbeddingsUnitTests when developing a custom embeddings integration and you need to verify that the model initializes correctly from both direct parameters and environment variables without making any network calls.
Code Reference
Source Location
- Repository: Langchain_ai_Langchain
- File:
libs/standard-tests/langchain_tests/unit_tests/embeddings.py
- Lines: 1-137
Signature
class EmbeddingsTests(BaseStandardTests):
"""Embeddings tests base class."""
@property
@abstractmethod
def embeddings_class(self) -> type[Embeddings]:
"""Embeddings class."""
@property
def embedding_model_params(self) -> dict[str, Any]:
"""Embeddings model parameters."""
...
@pytest.fixture
def model(self) -> Embeddings:
"""Embeddings model fixture."""
...
class EmbeddingsUnitTests(EmbeddingsTests):
"""Base class for embeddings unit tests."""
def test_init(self) -> None:
...
@property
def init_from_env_params(
self,
) -> tuple[dict[str, str], dict[str, Any], dict[str, Any]]:
...
def test_init_from_env(self) -> None:
...
Import
from langchain_tests.unit_tests.embeddings import EmbeddingsUnitTests
I/O Contract
Abstract Properties (Must Override)
| Name |
Type |
Required |
Description
|
| embeddings_class |
type[Embeddings] |
Yes |
The Embeddings subclass to test.
|
Optional Properties
| Name |
Type |
Required |
Description
|
| embedding_model_params |
dict[str, Any] |
No |
Constructor parameters for the embeddings model. Defaults to empty dict.
|
| init_from_env_params |
tuple[dict[str, str], dict[str, Any], dict[str, Any]] |
No |
Tuple of (env_vars, init_args, expected_attrs) for environment variable initialization testing. Defaults to empty dicts (test is skipped).
|
Test Methods
| Test |
Description
|
| test_init |
Verifies the embeddings model can be initialized with the provided embedding_model_params.
|
| test_init_from_env |
Verifies the model can be initialized from environment variables. Skipped if init_from_env_params returns empty dicts. Checks that instance attributes match expected values, handling SecretStr unwrapping.
|
Usage Examples
Basic Usage
from typing import Type
from langchain_tests.unit_tests.embeddings import EmbeddingsUnitTests
from my_package.embeddings import MyEmbeddingsModel
class TestMyEmbeddingsModelUnit(EmbeddingsUnitTests):
@property
def embeddings_class(self) -> Type[MyEmbeddingsModel]:
return MyEmbeddingsModel
@property
def embedding_model_params(self) -> dict:
return {"model": "model-001", "api_key": "test-key"}
With Environment Variable Testing
from typing import Tuple
from langchain_tests.unit_tests.embeddings import EmbeddingsUnitTests
from my_package.embeddings import MyEmbeddingsModel
class TestMyEmbeddingsModelUnit(EmbeddingsUnitTests):
@property
def embeddings_class(self):
return MyEmbeddingsModel
@property
def embedding_model_params(self) -> dict:
return {"model": "model-001", "api_key": "test-key"}
@property
def init_from_env_params(self) -> Tuple[dict, dict, dict]:
return (
{"MY_API_KEY": "api_key"}, # env vars to set
{"model": "model-001"}, # init args
{"my_api_key": "api_key"}, # expected attrs
)
Related Pages
Page Connections
Double-click a node to navigate. Hold to expand connections.
