Implementation:Langchain ai Langchain IndexerTests

Knowledge Sources	Langchain_ai_Langchain
Domains	Testing, Document Indexing, Standard Tests
Last Updated	2026-02-11 00:00 GMT

Overview

Standard integration test suites for validating synchronous and asynchronous implementations of the LangChain DocumentIndex abstraction.

Description

This module provides two abstract test suite classes -- DocumentIndexerTestSuite and AsyncDocumentIndexTestSuite -- in the langchain-tests (standard-tests) package. They verify that implementations of DocumentIndex from langchain-core correctly support document upsert, retrieval by ID, deletion, and overwriting operations. Implementers subclass one or both suites and provide a pytest fixture that yields an empty DocumentIndex instance.

Usage

Import DocumentIndexerTestSuite or AsyncDocumentIndexTestSuite when writing integration tests for a custom DocumentIndex implementation to ensure it conforms to the standard document indexing contract.

Code Reference

Source Location

• Repository: Langchain_ai_Langchain
• File: libs/standard-tests/langchain_tests/integration_tests/indexer.py
• Lines: 1-398

Signature

class DocumentIndexerTestSuite(ABC):
    @abstractmethod
    @pytest.fixture
    def index(self) -> Generator[DocumentIndex, None, None]: ...

class AsyncDocumentIndexTestSuite(ABC):
    @abstractmethod
    @pytest.fixture
    async def index(self) -> AsyncGenerator[DocumentIndex, None]: ...

Import

from langchain_tests.integration_tests.indexer import (
    DocumentIndexerTestSuite,
    AsyncDocumentIndexTestSuite,
)

I/O Contract

Inputs (Fixtures)

Name	Type	Required	Description
index	DocumentIndex	Yes	An empty DocumentIndex instance to test. Must be empty at the start of each test.

Outputs

Name	Type	Description
Test results	pytest outcomes	Pass/fail results for each standard test method.

Test Methods (DocumentIndexerTestSuite)

Test Method	Description
test_upsert_documents_has_no_ids	Verifies the upsert method signature does not include an "ids" parameter.
test_upsert_no_ids	Tests upserting documents that do not have pre-assigned IDs.
test_upsert_some_ids	Tests upserting a mix of documents with and without IDs.
test_upsert_overwrites	Tests that upsert overwrites existing content for the same document ID.
test_delete_missing_docs	Verifies deleting non-existent documents does not raise exceptions and returns correct response.
test_delete_semantics	Tests that delete correctly removes one document while reporting missing IDs correctly.
test_bulk_delete	Tests deletion of multiple documents at once.
test_delete_no_args	Tests that delete with no arguments raises ValueError.
test_delete_missing_content	Tests that deleting missing content does not raise exceptions.
test_get_with_missing_ids	Tests retrieval when some requested IDs do not exist.
test_get_missing	Tests retrieval when none of the requested IDs exist.

Test Methods (AsyncDocumentIndexTestSuite)

The async suite mirrors the sync suite with equivalent tests using aupsert, aget, and adelete async methods.

Usage Examples

Basic Usage

import pytest
from collections.abc import Generator
from langchain_core.indexing.base import DocumentIndex
from langchain_tests.integration_tests.indexer import DocumentIndexerTestSuite

from my_package.indexers import MyDocumentIndex


class TestMyDocumentIndex(DocumentIndexerTestSuite):
    @pytest.fixture
    def index(self) -> Generator[DocumentIndex, None, None]:
        idx = MyDocumentIndex()
        try:
            yield idx
        finally:
            idx.clear()