Add langchain standard integration tests

Author: cbornet

langchain/pyproject.toml

@@ -33,7 +33,9 @@ Issues = "https://github.com/vectorize-io/integrations-python/issues"
 dev = [
     "mypy>=1.13.0",
     "pytest>=8.3.3",
+    "pytest-asyncio>=0.26.0",
     "ruff>=0.9.0,<0.10",
+    "langchain-tests>=0.3.20",
 ]
 
 [tool.ruff.lint]
@@ -71,3 +73,5 @@ packages = ["langchain_vectorize"]
 requires = ["hatchling"]
 build-backend = "hatchling.build"
 
+[tool.pytest.ini_options]
+asyncio_mode = "auto"

langchain/tests/conftest.py

@@ -0,0 +1,180 @@
+import json
+import logging
+import os
+import time
+from collections.abc import Iterator
+from pathlib import Path
+from typing import Literal
+
+import pytest
+import urllib3
+import vectorize_client as v
+from vectorize_client import ApiClient, ApiException, RetrieveDocumentsRequest
+
+
+@pytest.fixture(scope="session")
+def api_token() -> str:
+    token = os.getenv("VECTORIZE_TOKEN", "wdcwd")
+    if not token:
+        msg = "Please set the VECTORIZE_TOKEN environment variable"
+        raise ValueError(msg)
+    return token
+
+
+@pytest.fixture(scope="session")
+def org_id() -> str:
+    org = os.getenv("VECTORIZE_ORG", "wdcd")
+    if not org:
+        msg = "Please set the VECTORIZE_ORG environment variable"
+        raise ValueError(msg)
+    return org
+
+
+@pytest.fixture(scope="session")
+def environment() -> Literal["prod", "dev", "local", "staging"]:
+    env = os.getenv("VECTORIZE_ENV", "prod")
+    if env not in ["prod", "dev", "local", "staging"]:
+        msg = "Invalid VECTORIZE_ENV environment variable."
+        raise ValueError(msg)
+    return env
+
+
+@pytest.fixture(scope="session")
+def api_client(api_token: str, environment: str) -> Iterator[ApiClient]:
+    header_name = None
+    header_value = None
+    if environment == "prod":
+        host = "https://api.vectorize.io/v1"
+    elif environment == "dev":
+        host = "https://api-dev.vectorize.io/v1"
+    elif environment == "local":
+        host = "http://localhost:3000/api"
+        header_name = "x-lambda-api-key"
+        header_value = api_token
+    else:
+        host = "https://api-staging.vectorize.io/v1"
+
+    with v.ApiClient(
+        v.Configuration(host=host, access_token=api_token, debug=True),
+        header_name,
+        header_value,
+    ) as api:
+        yield api
+
+
+@pytest.fixture(scope="session")
+def pipeline_id(api_client: v.ApiClient, org_id: str) -> Iterator[str]:
+    pipelines = v.PipelinesApi(api_client)
+
+    connectors_api = v.SourceConnectorsApi(api_client)
+    response = connectors_api.create_source_connector(
+        org_id,
+        v.CreateSourceConnectorRequest(
+            v.FileUpload(name="from api", type="FILE_UPLOAD")
+        ),
+    )
+    source_connector_id = response.connector.id
+    logging.info("Created source connector %s", source_connector_id)
+
+    uploads_api = v.UploadsApi(api_client)
+    upload_response = uploads_api.start_file_upload_to_connector(
+        org_id,
+        source_connector_id,
+        v.StartFileUploadToConnectorRequest(
+            name="research.pdf",
+            content_type="application/pdf",
+            metadata=json.dumps({"created-from-api": True}),
+        ),
+    )
+
+    http = urllib3.PoolManager()
+    this_dir = Path(__file__).parent
+    file_path = this_dir / "research.pdf"
+
+    with file_path.open("rb") as f:
+        http_response = http.request(
+            "PUT",
+            upload_response.upload_url,
+            body=f,
+            headers={
+                "Content-Type": "application/pdf",
+                "Content-Length": str(file_path.stat().st_size),
+            },
+        )
+    if http_response.status != 200:
+        msg = "Upload failed:"
+        raise ValueError(msg)
+    else:
+        logging.info("Upload successful")
+
+    ai_platforms = v.AIPlatformConnectorsApi(api_client).get_ai_platform_connectors(
+        org_id
+    )
+    builtin_ai_platform = next(
+        c.id for c in ai_platforms.ai_platform_connectors if c.type == "VECTORIZE"
+    )
+    logging.info("Using AI platform %s", builtin_ai_platform)
+
+    vector_databases = v.DestinationConnectorsApi(
+        api_client
+    ).get_destination_connectors(org_id)
+    builtin_vector_db = next(
+        c.id for c in vector_databases.destination_connectors if c.type == "VECTORIZE"
+    )
+    logging.info("Using destination connector %s", builtin_vector_db)
+
+    pipeline_response = pipelines.create_pipeline(
+        org_id,
+        v.PipelineConfigurationSchema(
+            source_connectors=[
+                v.PipelineSourceConnectorSchema(
+                    id=source_connector_id,
+                    type=v.SourceConnectorType.FILE_UPLOAD,
+                    config={},
+                )
+            ],
+            destination_connector=v.PipelineDestinationConnectorSchema(
+                id=builtin_vector_db,
+                type="VECTORIZE",
+                config={},
+            ),
+            ai_platform_connector=v.PipelineAIPlatformConnectorSchema(
+                id=builtin_ai_platform,
+                type="VECTORIZE",
+                config={},
+            ),
+            pipeline_name="Test pipeline",
+            schedule=v.ScheduleSchema(type="manual"),
+        ),
+    )
+    pipeline_id = pipeline_response.data.id
+
+    # Wait for the pipeline to be created
+    request = RetrieveDocumentsRequest(
+        question="query",
+        num_results=2,
+    )
+    start = time.time()
+    while True:
+        try:
+            response = pipelines.retrieve_documents(org_id, pipeline_id, request)
+        except ApiException as e:
+            if "503" not in str(e):
+                raise
+        else:
+            docs = response.documents
+            if len(docs) == 2:
+                break
+        if time.time() - start > 180:
+            msg = "Docs not retrieved in time"
+            raise RuntimeError(msg)
+        time.sleep(1)
+
+    logging.info("Created pipeline %s", pipeline_id)
+
+    yield pipeline_id
+
+    try:
+        pipelines.delete_pipeline(org_id, pipeline_id)
+    except Exception:
+        logging.exception("Failed to delete pipeline %s", pipeline_id)

langchain/tests/test_langchain_integration_tests.py

@@ -0,0 +1,50 @@
+from typing import Literal
+
+import pytest
+from langchain_tests.integration_tests import RetrieversIntegrationTests
+
+from langchain_vectorize import VectorizeRetriever
+
+
+class TestVectorizeRetrieverIntegration(RetrieversIntegrationTests):
+    @pytest.fixture(autouse=True)
+    def setup(
+        self,
+        environment: Literal["prod", "dev", "local", "staging"],
+        api_token: str,
+        org_id: str,
+        pipeline_id: str,
+    ) -> None:
+        self._environment = environment
+        self._api_token = api_token
+        self._org_id = org_id
+        self._pipeline_id = pipeline_id
+
+    @property
+    def retriever_constructor(self) -> type[VectorizeRetriever]:
+        return VectorizeRetriever
+
+    @property
+    def retriever_constructor_params(self) -> dict:
+        return {
+            "environment": self._environment,
+            "api_token": self._api_token,
+            "organization": self._org_id,
+            "pipeline_id": self._pipeline_id,
+        }
+
+    @property
+    def retriever_query_example(self) -> str:
+        return "What are you?"
+
+    @pytest.mark.xfail(
+        reason="VectorizeRetriever does not support k parameter in constructor"
+    )
+    def test_k_constructor_param(self) -> None:
+        raise NotImplementedError
+
+    @pytest.mark.xfail(
+        reason="VectorizeRetriever does not support k parameter in invoke"
+    )
+    def test_invoke_with_k_kwarg(self) -> None:
+        raise NotImplementedError