scheduled tests GHA (#8879)

Adding scheduled daily GHA that runs marked integration tests. To start
just marking some tests in test_openai

.github/workflows/langchain_experimental_ci.yml

@@ -1,5 +1,5 @@
 ---
-name: libs/langchain-experimental CI
+name: libs/experimental CI
 
 on:
   push:

.github/workflows/langchain_experimental_release.yml

@@ -1,5 +1,5 @@
 ---
-name: libs/langchain-experimental Release
+name: libs/experimental Release
 
 on:
   pull_request:

.github/workflows/scheduled_test.yml

@@ -0,0 +1,38 @@
+name: Scheduled tests
+
+on:
+  scheduled:
+    - cron:  '0 13 * * *'
+
+env:
+  POETRY_VERSION: "1.4.2"
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+    environment: Scheduled testing
+    strategy:
+      matrix:
+        python-version:
+          - "3.8"
+          - "3.9"
+          - "3.10"
+          - "3.11"
+    name: Python ${{ matrix.python-version }}
+    steps:
+      - uses: actions/checkout@v3
+      - name: Set up Python ${{ matrix.python-version }}
+        uses: "./.github/actions/poetry_setup"
+        with:
+          python-version: ${{ matrix.python-version }}
+          poetry-version: "1.4.2"
+          install-command: |
+            echo "Running scheduled tests, installing dependencies with poetry..."
+            poetry install -E scheduled_testing
+      - name: Run tests
+        env:
+          OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
+        run: |
+          make scheduled_tests
+        shell: bash
+    secrets: inherit

libs/experimental/tests/integration_tests/chains/test_cpal.py

@@ -7,6 +7,10 @@ from unittest import mock
 
 import pydantic
 import pytest
+from langchain import OpenAI
+from langchain.output_parsers import PydanticOutputParser
+from langchain.prompts.prompt import PromptTemplate
+
 from langchain_experimental.cpal.base import (
     CausalChain,
     CPALChain,
@@ -35,10 +39,6 @@ from langchain_experimental.cpal.templates.univariate.narrative import (
 from langchain_experimental.cpal.templates.univariate.query import (
     template as query_template,
 )
-
-from langchain import OpenAI
-from langchain.output_parsers import PydanticOutputParser
-from langchain.prompts.prompt import PromptTemplate
 from tests.unit_tests.llms.fake_llm import FakeLLM
 
 

libs/experimental/tests/integration_tests/chains/test_sql_database.py

@@ -1,12 +1,11 @@
 """Test SQL Database Chain."""
-from sqlalchemy import Column, Integer, MetaData, String, Table, create_engine, insert
-
-from langchain.chains.sql_database.base import (
+from langchain.llms.openai import OpenAI
+from langchain.utilities.sql_database import SQLDatabase
+from libs.experimental.langchain_experimental.sql.base import (
     SQLDatabaseChain,
     SQLDatabaseSequentialChain,
 )
-from langchain.llms.openai import OpenAI
-from langchain.utilities.sql_database import SQLDatabase
+from sqlalchemy import Column, Integer, MetaData, String, Table, create_engine, insert
 
 metadata_obj = MetaData()
 

libs/langchain/Makefile

@@ -59,6 +59,9 @@ test_watch:
 integration_tests:
 	poetry run pytest tests/integration_tests
 
+scheduled_tests:
+	poetry run pytest -m scheduled tests/integration_tests
+
 docker_tests:
 	docker build -t my-langchain-image:test .
 	docker run --rm my-langchain-image:test

libs/langchain/pyproject.toml

@@ -375,6 +375,10 @@ extended_testing = [
  "xmltodict",
 ]
 
+scheduled_testing = [
+ "openai",
+]
+
 [tool.ruff]
 select = [
   "E",  # pycodestyle
@@ -413,7 +417,8 @@ addopts = "--strict-markers --strict-config --durations=5 --snapshot-warn-unused
 # Registering custom markers.
 # https://docs.pytest.org/en/7.1.x/example/markers.html#registering-markers
 markers = [
-  "requires: mark tests as requiring a specific library"
+  "requires: mark tests as requiring a specific library",
+  "scheduled: mark tests to run in scheduled testing",
 ]
 
 [tool.codespell]

libs/langchain/tests/integration_tests/cache/test_gptcache.py

@@ -18,7 +18,7 @@ except ImportError:
     gptcache_installed = False
 
 
-def init_gptcache_map(cache_obj: Cache) -> None:
+def init_gptcache_map(cache_obj: Any) -> None:
     i = getattr(init_gptcache_map, "_i", 0)
     cache_path = f"data_map_{i}.txt"
     if os.path.isfile(cache_path):
@@ -30,7 +30,7 @@ def init_gptcache_map(cache_obj: Cache) -> None:
     init_gptcache_map._i = i + 1  # type: ignore
 
 
-def init_gptcache_map_with_llm(cache_obj: Cache, llm: str) -> None:
+def init_gptcache_map_with_llm(cache_obj: Any, llm: str) -> None:
     cache_path = f"data_map_{llm}.txt"
     if os.path.isfile(cache_path):
         os.remove(cache_path)

libs/langchain/tests/integration_tests/cache/test_momento_cache.py

@@ -11,7 +11,6 @@ from datetime import timedelta
 from typing import Iterator
 
 import pytest
-from momento import CacheClient, Configurations, CredentialProvider
 
 import langchain
 from langchain.cache import MomentoCache
@@ -25,6 +24,8 @@ def random_string() -> str:
 
 @pytest.fixture(scope="module")
 def momento_cache() -> Iterator[MomentoCache]:
+    from momento import CacheClient, Configurations, CredentialProvider
+
     cache_name = f"langchain-test-cache-{random_string()}"
     client = CacheClient(
         Configurations.Laptop.v1(),
@@ -40,6 +41,8 @@ def momento_cache() -> Iterator[MomentoCache]:
 
 
 def test_invalid_ttl() -> None:
+    from momento import CacheClient, Configurations, CredentialProvider
+
     client = CacheClient(
         Configurations.Laptop.v1(),
         CredentialProvider.from_environment_variable("MOMENTO_AUTH_TOKEN"),

libs/langchain/tests/integration_tests/cache/test_redis_cache.py

@@ -1,6 +1,5 @@
 """Test Redis cache functionality."""
 import pytest
-import redis
 
 import langchain
 from langchain.cache import RedisCache, RedisSemanticCache
@@ -13,6 +12,8 @@ REDIS_TEST_URL = "redis://localhost:6379"
 
 
 def test_redis_cache() -> None:
+    import redis
+
     langchain.llm_cache = RedisCache(redis_=redis.Redis.from_url(REDIS_TEST_URL))
     llm = FakeLLM()
     params = llm.dict()
@@ -31,6 +32,8 @@ def test_redis_cache() -> None:
 
 
 def test_redis_cache_chat() -> None:
+    import redis
+
     langchain.llm_cache = RedisCache(redis_=redis.Redis.from_url(REDIS_TEST_URL))
     llm = FakeChatModel()
     params = llm.dict()
@@ -69,6 +72,8 @@ def test_redis_semantic_cache() -> None:
 
 
 def test_redis_semantic_cache_chat() -> None:
+    import redis
+
     langchain.llm_cache = RedisCache(redis_=redis.Redis.from_url(REDIS_TEST_URL))
     llm = FakeChatModel()
     params = llm.dict()

libs/langchain/tests/integration_tests/document_loaders/test_unstructured.py

@@ -2,8 +2,6 @@ import os
 from contextlib import ExitStack
 from pathlib import Path
 
-from unstructured.cleaners.core import clean_extra_whitespace
-
 from langchain.document_loaders import (
     UnstructuredAPIFileIOLoader,
     UnstructuredAPIFileLoader,
@@ -14,6 +12,8 @@ EXAMPLE_DOCS_DIRECTORY = str(Path(__file__).parent.parent / "examples/")
 
 
 def test_unstructured_loader_with_post_processor() -> None:
+    from unstructured.cleaners.core import clean_extra_whitespace
+
     file_path = os.path.join(EXAMPLE_DOCS_DIRECTORY, "layout-parser-paper.pdf")
     loader = UnstructuredFileLoader(
         file_path=file_path,

libs/langchain/tests/integration_tests/llms/test_baseten.py

@@ -1,15 +1,13 @@
 """Test Baseten API wrapper."""
 import os
 
-import baseten
-import pytest
-
 from langchain.llms.baseten import Baseten
 
 
-@pytest.mark.requires(baseten)
 def test_baseten_call() -> None:
     """Test valid call to Baseten."""
+    import baseten
+
     baseten.login(os.environ["BASETEN_API_KEY"])
     llm = Baseten(model=os.environ["BASETEN_MODEL_ID"])
     output = llm("Say foo:")

libs/langchain/tests/integration_tests/llms/test_openai.py

@@ -7,16 +7,17 @@ import pytest
 from langchain.callbacks.manager import CallbackManager
 from langchain.chat_models.openai import ChatOpenAI
 from langchain.llms.loading import load_llm
-from langchain.llms.openai import OpenAI, OpenAIChat
+from langchain.llms.openai import OpenAI
 from langchain.schema import LLMResult
 from tests.unit_tests.callbacks.fake_callback_handler import (
     FakeCallbackHandler,
 )
 
 
+@pytest.mark.scheduled
 def test_openai_call() -> None:
     """Test valid call to openai."""
-    llm = OpenAI(max_tokens=10, n=3)
+    llm = OpenAI()
     output = llm("Say something nice:")
     assert isinstance(output, str)
 
@@ -86,6 +87,7 @@ def test_saving_loading_llm(tmp_path: Path) -> None:
     assert loaded_llm == llm
 
 
+@pytest.mark.scheduled
 def test_openai_streaming() -> None:
     """Test streaming tokens from OpenAI."""
     llm = OpenAI(max_tokens=10)
@@ -97,6 +99,7 @@ def test_openai_streaming() -> None:
         assert isinstance(token, str)
 
 
+@pytest.mark.scheduled
 @pytest.mark.asyncio
 async def test_openai_astream() -> None:
     """Test streaming tokens from OpenAI."""
@@ -106,6 +109,7 @@ async def test_openai_astream() -> None:
         assert isinstance(token, str)
 
 
+@pytest.mark.scheduled
 @pytest.mark.asyncio
 async def test_openai_abatch() -> None:
     """Test streaming tokens from OpenAI."""
@@ -128,6 +132,7 @@ async def test_openai_abatch_tags() -> None:
         assert isinstance(token, str)
 
 
+@pytest.mark.scheduled
 def test_openai_batch() -> None:
     """Test streaming tokens from OpenAI."""
     llm = OpenAI(max_tokens=10)
@@ -137,6 +142,7 @@ def test_openai_batch() -> None:
         assert isinstance(token, str)
 
 
+@pytest.mark.scheduled
 @pytest.mark.asyncio
 async def test_openai_ainvoke() -> None:
     """Test streaming tokens from OpenAI."""
@@ -146,6 +152,7 @@ async def test_openai_ainvoke() -> None:
     assert isinstance(result, str)
 
 
+@pytest.mark.scheduled
 def test_openai_invoke() -> None:
     """Test streaming tokens from OpenAI."""
     llm = OpenAI(max_tokens=10)
@@ -154,6 +161,7 @@ def test_openai_invoke() -> None:
     assert isinstance(result, str)
 
 
+@pytest.mark.scheduled
 def test_openai_multiple_prompts() -> None:
     """Test completion with multiple prompts."""
     llm = OpenAI(max_tokens=10)
@@ -181,6 +189,7 @@ def test_openai_streaming_multiple_prompts_error() -> None:
         OpenAI(streaming=True).generate(["I'm Pickle Rick", "I'm Pickle Rick"])
 
 
+@pytest.mark.scheduled
 def test_openai_streaming_call() -> None:
     """Test valid call to openai."""
     llm = OpenAI(max_tokens=10, streaming=True)
@@ -203,6 +212,7 @@ def test_openai_streaming_callback() -> None:
     assert callback_handler.llm_streams == 10
 
 
+@pytest.mark.scheduled
 @pytest.mark.asyncio
 async def test_openai_async_generate() -> None:
     """Test async generation."""
@@ -235,60 +245,6 @@ def test_openai_chat_wrong_class() -> None:
     assert isinstance(output, str)
 
 
-def test_openai_chat() -> None:
-    """Test OpenAIChat."""
-    llm = OpenAIChat(max_tokens=10)
-    output = llm("Say foo:")
-    assert isinstance(output, str)
-
-
-def test_openai_chat_streaming() -> None:
-    """Test OpenAIChat with streaming option."""
-    llm = OpenAIChat(max_tokens=10, streaming=True)
-    output = llm("Say foo:")
-    assert isinstance(output, str)
-
-
-def test_openai_chat_streaming_callback() -> None:
-    """Test that streaming correctly invokes on_llm_new_token callback."""
-    callback_handler = FakeCallbackHandler()
-    callback_manager = CallbackManager([callback_handler])
-    llm = OpenAIChat(
-        max_tokens=10,
-        streaming=True,
-        temperature=0,
-        callback_manager=callback_manager,
-        verbose=True,
-    )
-    llm("Write me a sentence with 100 words.")
-    assert callback_handler.llm_streams != 0
-
-
-@pytest.mark.asyncio
-async def test_openai_chat_async_generate() -> None:
-    """Test async chat."""
-    llm = OpenAIChat(max_tokens=10)
-    output = await llm.agenerate(["Hello, how are you?"])
-    assert isinstance(output, LLMResult)
-
-
-@pytest.mark.asyncio
-async def test_openai_chat_async_streaming_callback() -> None:
-    """Test that streaming correctly invokes on_llm_new_token callback."""
-    callback_handler = FakeCallbackHandler()
-    callback_manager = CallbackManager([callback_handler])
-    llm = OpenAIChat(
-        max_tokens=10,
-        streaming=True,
-        temperature=0,
-        callback_manager=callback_manager,
-        verbose=True,
-    )
-    result = await llm.agenerate(["Write me a sentence with 100 words."])
-    assert callback_handler.llm_streams != 0
-    assert isinstance(result, LLMResult)
-
-
 def test_openai_modelname_to_contextsize_valid() -> None:
     """Test model name to context size on a valid model."""
     assert OpenAI().modelname_to_contextsize("davinci") == 2049

libs/langchain/tests/integration_tests/memory/test_cosmos_db.py

@@ -6,8 +6,8 @@ from langchain.memory.chat_message_histories import CosmosDBChatMessageHistory
 from langchain.schema.messages import _message_to_dict
 
 # Replace these with your Azure Cosmos DB endpoint and key
-endpoint = os.environ["COSMOS_DB_ENDPOINT"]
-credential = os.environ["COSMOS_DB_KEY"]
+endpoint = os.environ.get("COSMOS_DB_ENDPOINT", "")
+credential = os.environ.get("COSMOS_DB_KEY", "")
 
 
 def test_memory_with_message_store() -> None:

libs/langchain/tests/integration_tests/memory/test_momento.py

@@ -10,7 +10,6 @@ from datetime import timedelta
 from typing import Iterator
 
 import pytest
-from momento import CacheClient, Configurations, CredentialProvider
 
 from langchain.memory import ConversationBufferMemory
 from langchain.memory.chat_message_histories import MomentoChatMessageHistory
@@ -23,6 +22,8 @@ def random_string() -> str:
 
 @pytest.fixture(scope="function")
 def message_history() -> Iterator[MomentoChatMessageHistory]:
+    from momento import CacheClient, Configurations, CredentialProvider
+
     cache_name = f"langchain-test-cache-{random_string()}"
     client = CacheClient(
         Configurations.Laptop.v1(),

libs/langchain/tests/integration_tests/memory/test_mongodb.py

@@ -6,7 +6,7 @@ from langchain.memory.chat_message_histories import MongoDBChatMessageHistory
 from langchain.schema.messages import _message_to_dict
 
 # Replace these with your mongodb connection string
-connection_string = os.environ["MONGODB_CONNECTION_STRING"]
+connection_string = os.environ.get("MONGODB_CONNECTION_STRING", "")
 
 
 def test_memory_with_message_store() -> None:

libs/langchain/tests/integration_tests/retrievers/docarray/fixtures.py

@@ -1,43 +1,30 @@
+from __future__ import annotations
+
 from pathlib import Path
-from typing import Any, Dict, Generator, Tuple
+from typing import TYPE_CHECKING, Any, Dict, Generator, Tuple
 
 import numpy as np
 import pytest
-from docarray import BaseDoc
-from docarray.index import (
-    ElasticDocIndex,
-    HnswDocumentIndex,
-    InMemoryExactNNIndex,
-    QdrantDocumentIndex,
-    WeaviateDocumentIndex,
-)
-from docarray.typing import NdArray
 from pydantic import Field
-from qdrant_client.http import models as rest
+
+if TYPE_CHECKING:
+    from docarray.index import (
+        ElasticDocIndex,
+        HnswDocumentIndex,
+        InMemoryExactNNIndex,
+        QdrantDocumentIndex,
+        WeaviateDocumentIndex,
+    )
+    from docarray.typing import NdArray
+    from qdrant_client.http import models as rest
 
 from langchain.embeddings import FakeEmbeddings
 
 
-class MyDoc(BaseDoc):
-    title: str
-    title_embedding: NdArray[32]  # type: ignore
-    other_emb: NdArray[32]  # type: ignore
-    year: int
-
-
-class WeaviateDoc(BaseDoc):
-    # When initializing the Weaviate index, denote the field
-    # you want to search on with `is_embedding=True`
-    title: str
-    title_embedding: NdArray[32] = Field(is_embedding=True)  # type: ignore
-    other_emb: NdArray[32]  # type: ignore
-    year: int
-
-
 @pytest.fixture
 def init_weaviate() -> (
     Generator[
-        Tuple[WeaviateDocumentIndex[WeaviateDoc], Dict[str, Any], FakeEmbeddings],
+        Tuple[WeaviateDocumentIndex, Dict[str, Any], FakeEmbeddings],
         None,
         None,
     ]
@@ -46,6 +33,19 @@ def init_weaviate() -> (
     cd tests/integration_tests/vectorstores/docker-compose
     docker compose -f weaviate.yml up
     """
+    from docarray import BaseDoc
+    from docarray.index import (
+        WeaviateDocumentIndex,
+    )
+
+    class WeaviateDoc(BaseDoc):
+        # When initializing the Weaviate index, denote the field
+        # you want to search on with `is_embedding=True`
+        title: str
+        title_embedding: NdArray[32] = Field(is_embedding=True)  # type: ignore
+        other_emb: NdArray[32]  # type: ignore
+        year: int
+
     embeddings = FakeEmbeddings(size=32)
 
     # initialize WeaviateDocumentIndex
@@ -76,12 +76,23 @@ def init_weaviate() -> (
 
 @pytest.fixture
 def init_elastic() -> (
-    Generator[Tuple[ElasticDocIndex[MyDoc], Dict[str, Any], FakeEmbeddings], None, None]
+    Generator[Tuple[ElasticDocIndex, Dict[str, Any], FakeEmbeddings], None, None]
 ):
     """
     cd tests/integration_tests/vectorstores/docker-compose
     docker-compose -f elasticsearch.yml up
     """
+    from docarray import BaseDoc
+    from docarray.index import (
+        ElasticDocIndex,
+    )
+
+    class MyDoc(BaseDoc):
+        title: str
+        title_embedding: NdArray[32]  # type: ignore
+        other_emb: NdArray[32]  # type: ignore
+        year: int
+
     embeddings = FakeEmbeddings(size=32)
 
     # initialize ElasticDocIndex
@@ -109,7 +120,16 @@ def init_elastic() -> (
 
 
 @pytest.fixture
-def init_qdrant() -> Tuple[QdrantDocumentIndex[MyDoc], rest.Filter, FakeEmbeddings]:
+def init_qdrant() -> Tuple[QdrantDocumentIndex, rest.Filter, FakeEmbeddings]:
+    from docarray import BaseDoc
+    from docarray.index import QdrantDocumentIndex
+
+    class MyDoc(BaseDoc):
+        title: str
+        title_embedding: NdArray[32]  # type: ignore
+        other_emb: NdArray[32]  # type: ignore
+        year: int
+
     embeddings = FakeEmbeddings(size=32)
 
     # initialize QdrantDocumentIndex
@@ -144,9 +164,16 @@ def init_qdrant() -> Tuple[QdrantDocumentIndex[MyDoc], rest.Filter, FakeEmbeddin
 
 
 @pytest.fixture
-def init_in_memory() -> (
-    Tuple[InMemoryExactNNIndex[MyDoc], Dict[str, Any], FakeEmbeddings]
-):
+def init_in_memory() -> Tuple[InMemoryExactNNIndex, Dict[str, Any], FakeEmbeddings]:
+    from docarray import BaseDoc
+    from docarray.index import InMemoryExactNNIndex
+
+    class MyDoc(BaseDoc):
+        title: str
+        title_embedding: NdArray[32]  # type: ignore
+        other_emb: NdArray[32]  # type: ignore
+        year: int
+
     embeddings = FakeEmbeddings(size=32)
 
     # initialize InMemoryExactNNIndex
@@ -172,7 +199,18 @@ def init_in_memory() -> (
 @pytest.fixture
 def init_hnsw(
     tmp_path: Path,
-) -> Tuple[HnswDocumentIndex[MyDoc], Dict[str, Any], FakeEmbeddings]:
+) -> Tuple[HnswDocumentIndex, Dict[str, Any], FakeEmbeddings]:
+    from docarray import BaseDoc
+    from docarray.index import (
+        HnswDocumentIndex,
+    )
+
+    class MyDoc(BaseDoc):
+        title: str
+        title_embedding: NdArray[32]  # type: ignore
+        other_emb: NdArray[32]  # type: ignore
+        year: int
+
     embeddings = FakeEmbeddings(size=32)
 
     # initialize InMemoryExactNNIndex

libs/langchain/tests/integration_tests/retrievers/docarray/test_backends.py

@@ -1,7 +1,6 @@
 from typing import Any
 
 import pytest
-from vcr.request import Request
 
 from langchain.retrievers import DocArrayRetriever
 from tests.integration_tests.retrievers.docarray.fixtures import (  # noqa: F401
@@ -17,7 +16,7 @@ from tests.integration_tests.retrievers.docarray.fixtures import (  # noqa: F401
     "backend",
     ["init_hnsw", "init_in_memory", "init_qdrant", "init_elastic", "init_weaviate"],
 )
-def test_backends(request: Request, backend: Any) -> None:
+def test_backends(request: Any, backend: Any) -> None:
     index, filter_query, embeddings = request.getfixturevalue(backend)
 
     # create a retriever

libs/langchain/tests/integration_tests/retrievers/test_weaviate_hybrid_search.py

@@ -6,7 +6,6 @@ from typing import Generator, Union
 from uuid import uuid4
 
 import pytest
-from weaviate import Client
 
 from langchain.docstore.document import Document
 from langchain.retrievers.weaviate_hybrid_search import WeaviateHybridSearchRetriever
@@ -28,6 +27,8 @@ class TestWeaviateHybridSearchRetriever:
     @pytest.fixture(scope="class", autouse=True)
     def weaviate_url(self) -> Union[str, Generator[str, None, None]]:
         """Return the weaviate url."""
+        from weaviate import Client
+
         url = "http://localhost:8080"
         yield url
 
@@ -38,6 +39,8 @@ class TestWeaviateHybridSearchRetriever:
     @pytest.mark.vcr(ignore_localhost=True)
     def test_get_relevant_documents(self, weaviate_url: str) -> None:
         """Test end to end construction and MRR search."""
+        from weaviate import Client
+
         texts = ["foo", "bar", "baz"]
         metadatas = [{"page": i} for i in range(len(texts))]
 
@@ -64,6 +67,8 @@ class TestWeaviateHybridSearchRetriever:
     @pytest.mark.vcr(ignore_localhost=True)
     def test_get_relevant_documents_with_score(self, weaviate_url: str) -> None:
         """Test end to end construction and MRR search."""
+        from weaviate import Client
+
         texts = ["foo", "bar", "baz"]
         metadatas = [{"page": i} for i in range(len(texts))]
 
@@ -87,6 +92,8 @@ class TestWeaviateHybridSearchRetriever:
     @pytest.mark.vcr(ignore_localhost=True)
     def test_get_relevant_documents_with_filter(self, weaviate_url: str) -> None:
         """Test end to end construction and MRR search."""
+        from weaviate import Client
+
         texts = ["foo", "bar", "baz"]
         metadatas = [{"page": i} for i in range(len(texts))]
 
@@ -113,6 +120,8 @@ class TestWeaviateHybridSearchRetriever:
     @pytest.mark.vcr(ignore_localhost=True)
     def test_get_relevant_documents_with_uuids(self, weaviate_url: str) -> None:
         """Test end to end construction and MRR search."""
+        from weaviate import Client
+
         texts = ["foo", "bar", "baz"]
         metadatas = [{"page": i} for i in range(len(texts))]
         # Weaviate replaces the object if the UUID already exists

libs/langchain/tests/integration_tests/vectorstores/qdrant/async_api/test_add_texts.py

@@ -2,7 +2,6 @@ import uuid
 from typing import Optional
 
 import pytest
-from qdrant_client.http import models as rest
 
 from langchain.vectorstores import Qdrant
 from tests.integration_tests.vectorstores.fake_embeddings import (
@@ -69,6 +68,7 @@ async def test_qdrant_aadd_texts_stores_ids(
 ) -> None:
     """Test end to end Qdrant.aadd_texts stores provided ids."""
     from qdrant_client import QdrantClient
+    from qdrant_client.http import models as rest
 
     ids = [
         "fa38d572-4c31-4579-aedc-1960d79df6df",
@@ -101,6 +101,7 @@ async def test_qdrant_aadd_texts_stores_embeddings_as_named_vectors(
 ) -> None:
     """Test end to end Qdrant.aadd_texts stores named vectors if name is provided."""
     from qdrant_client import QdrantClient
+    from qdrant_client.http import models as rest
 
     collection_name = uuid.uuid4().hex
 

libs/langchain/tests/integration_tests/vectorstores/qdrant/async_api/test_similarity_search.py

@@ -2,7 +2,6 @@ from typing import Optional
 
 import numpy as np
 import pytest
-from qdrant_client.http import models as rest
 
 from langchain.schema import Document
 from langchain.vectorstores import Qdrant
@@ -239,6 +238,8 @@ async def test_qdrant_similarity_search_filters_with_qdrant_filters(
     qdrant_location: str,
 ) -> None:
     """Test end to end construction and search."""
+    from qdrant_client.http import models as rest
+
     texts = ["foo", "bar", "baz"]
     metadatas = [
         {"page": i, "details": {"page": i + 1, "pages": [i + 2, -1]}}

libs/langchain/tests/integration_tests/vectorstores/qdrant/test_add_texts.py

@@ -2,7 +2,6 @@ import uuid
 from typing import Optional
 
 import pytest
-from qdrant_client.http import models as rest
 
 from langchain.schema import Document
 from langchain.vectorstores import Qdrant
@@ -81,6 +80,7 @@ def test_qdrant_add_texts_stores_duplicated_texts(vector_name: Optional[str]) ->
 def test_qdrant_add_texts_stores_ids(batch_size: int) -> None:
     """Test end to end Qdrant.add_texts stores provided ids."""
     from qdrant_client import QdrantClient
+    from qdrant_client.http import models as rest
 
     ids = [
         "fa38d572-4c31-4579-aedc-1960d79df6df",
@@ -107,6 +107,7 @@ def test_qdrant_add_texts_stores_ids(batch_size: int) -> None:
 def test_qdrant_add_texts_stores_embeddings_as_named_vectors(vector_name: str) -> None:
     """Test end to end Qdrant.add_texts stores named vectors if name is provided."""
     from qdrant_client import QdrantClient
+    from qdrant_client.http import models as rest
 
     collection_name = uuid.uuid4().hex
 

libs/langchain/tests/integration_tests/vectorstores/qdrant/test_similarity_search.py

@@ -2,7 +2,6 @@ from typing import Optional
 
 import numpy as np
 import pytest
-from qdrant_client.http import models as rest
 
 from langchain.schema import Document
 from langchain.vectorstores import Qdrant
@@ -209,6 +208,8 @@ def test_qdrant_similarity_search_filters_with_qdrant_filters(
     vector_name: Optional[str],
 ) -> None:
     """Test end to end construction and search."""
+    from qdrant_client.http import models as rest
+
     texts = ["foo", "bar", "baz"]
     metadatas = [
         {"page": i, "details": {"page": i + 1, "pages": [i + 2, -1]}}

libs/langchain/tests/integration_tests/vectorstores/test_azuresearch.py

@@ -1,7 +1,6 @@
 import os
 import time
 
-import openai
 import pytest
 from dotenv import load_dotenv
 
@@ -10,13 +9,7 @@ from langchain.vectorstores.azuresearch import AzureSearch
 
 load_dotenv()
 
-# Azure OpenAI settings
-openai.api_type = "azure"
-openai.api_base = os.getenv("OPENAI_API_BASE", "")
-openai.api_version = "2023-05-15"
-openai.api_key = os.getenv("OPENAI_API_KEY", "")
-model: str = os.getenv("OPENAI_EMBEDDINGS_ENGINE_DOC", "text-embedding-ada-002")
-
+model = os.getenv("OPENAI_EMBEDDINGS_ENGINE_DOC", "text-embedding-ada-002")
 # Vector store settings
 vector_store_address: str = os.getenv("AZURE_SEARCH_ENDPOINT", "")
 vector_store_password: str = os.getenv("AZURE_SEARCH_ADMIN_KEY", "")

libs/langchain/tests/integration_tests/vectorstores/test_deeplake.py

@@ -1,5 +1,4 @@
 """Test Deep Lake functionality."""
-import deeplake
 import pytest
 from pytest import FixtureRequest
 
@@ -53,6 +52,8 @@ def test_deeplake_with_metadatas() -> None:
 
 def test_deeplakewith_persistence() -> None:
     """Test end to end construction and search, with persistence."""
+    import deeplake
+
     dataset_path = "./tests/persist_dir"
     if deeplake.exists(dataset_path):
         deeplake.delete(dataset_path)
@@ -84,6 +85,8 @@ def test_deeplakewith_persistence() -> None:
 
 def test_deeplake_overwrite_flag() -> None:
     """Test overwrite behavior"""
+    import deeplake
+
     dataset_path = "./tests/persist_dir"
     if deeplake.exists(dataset_path):
         deeplake.delete(dataset_path)
@@ -234,6 +237,8 @@ def test_delete_dataset_by_filter(deeplake_datastore: DeepLake) -> None:
 
 def test_delete_by_path(deeplake_datastore: DeepLake) -> None:
     """Test delete dataset."""
+    import deeplake
+
     path = deeplake_datastore.dataset_path
     DeepLake.force_delete_by_path(path)
     assert not deeplake.exists(path)

libs/langchain/tests/integration_tests/vectorstores/test_elasticsearch.py

@@ -5,7 +5,6 @@ import uuid
 from typing import Generator, List, Union
 
 import pytest
-from elasticsearch import Elasticsearch
 
 from langchain.docstore.document import Document
 from langchain.embeddings import OpenAIEmbeddings
@@ -29,6 +28,8 @@ class TestElasticsearch:
     @pytest.fixture(scope="class", autouse=True)
     def elasticsearch_url(self) -> Union[str, Generator[str, None, None]]:
         """Return the elasticsearch url."""
+        from elasticsearch import Elasticsearch
+
         url = "http://localhost:9200"
         yield url
         es = Elasticsearch(hosts=url)
@@ -108,6 +109,7 @@ class TestElasticsearch:
     ) -> None:
         """This test checks the construction of a custom
         ElasticSearch index using the 'from_documents'."""
+        from elasticsearch import Elasticsearch
 
         index_name = f"custom_index_{uuid.uuid4().hex}"
         elastic_vector_search = ElasticVectorSearch.from_documents(
@@ -134,6 +136,7 @@ class TestElasticsearch:
     ) -> None:
         """This test checks the construction of a custom
         ElasticSearch index using the 'add_documents'."""
+        from elasticsearch import Elasticsearch
 
         index_name = f"custom_index_{uuid.uuid4().hex}"
         elastic_vector_search = ElasticVectorSearch(

libs/langchain/tests/integration_tests/vectorstores/test_lancedb.py

@@ -1,10 +1,10 @@
-import lancedb
-
 from langchain.vectorstores import LanceDB
 from tests.integration_tests.vectorstores.fake_embeddings import FakeEmbeddings
 
 
 def test_lancedb() -> None:
+    import lancedb
+
     embeddings = FakeEmbeddings()
     db = lancedb.connect("/tmp/lancedb")
     texts = ["text 1", "text 2", "item 3"]
@@ -24,6 +24,8 @@ def test_lancedb() -> None:
 
 
 def test_lancedb_add_texts() -> None:
+    import lancedb
+
     embeddings = FakeEmbeddings()
     db = lancedb.connect("/tmp/lancedb")
     texts = ["text 1"]

libs/langchain/tests/integration_tests/vectorstores/test_marqo.py

@@ -1,7 +1,6 @@
 """Test Marqo functionality."""
 from typing import Dict
 
-import marqo
 import pytest
 
 from langchain.docstore.document import Document
@@ -14,6 +13,8 @@ INDEX_NAME = "langchain-integration-tests"
 
 @pytest.fixture
 def client() -> Marqo:
+    import marqo
+
     # fixture for marqo client to be used throughout testing, resets the index
     client = marqo.Client(url=DEFAULT_MARQO_URL, api_key=DEFAULT_MARQO_API_KEY)
     try:
@@ -128,6 +129,8 @@ def test_marqo_weighted_query(client: Marqo) -> None:
 
 
 def test_marqo_multimodal() -> None:
+    import marqo
+
     client = marqo.Client(url=DEFAULT_MARQO_URL, api_key=DEFAULT_MARQO_API_KEY)
     try:
         client.index(INDEX_NAME).delete()

libs/langchain/tests/integration_tests/vectorstores/test_meilisearch.py

@@ -1,7 +1,6 @@
 """Test Meilisearch functionality."""
-from typing import Generator
+from typing import TYPE_CHECKING, Generator
 
-import meilisearch
 import pytest
 import requests
 
@@ -9,6 +8,9 @@ from langchain.docstore.document import Document
 from langchain.vectorstores import Meilisearch
 from tests.integration_tests.vectorstores.fake_embeddings import FakeEmbeddings
 
+if TYPE_CHECKING:
+    import meilisearch
+
 INDEX_NAME = "test-langchain-demo"
 TEST_MEILI_HTTP_ADDR = "http://localhost:7700"
 TEST_MEILI_MASTER_KEY = "masterKey"
@@ -49,7 +51,9 @@ class TestMeilisearchVectorSearch:
             task = client.index(index.uid).delete()
             client.wait_for_task(task.task_uid)
 
-    def client(self) -> meilisearch.Client:
+    def client(self) -> "meilisearch.Client":
+        import meilisearch
+
         return meilisearch.Client(TEST_MEILI_HTTP_ADDR, TEST_MEILI_MASTER_KEY)
 
     def _wait_last_task(self) -> None:

libs/langchain/tests/integration_tests/vectorstores/test_mongodb_atlas.py

@@ -3,7 +3,7 @@ from __future__ import annotations
 
 import os
 from time import sleep
-from typing import TYPE_CHECKING
+from typing import TYPE_CHECKING, Any
 
 import pytest
 
@@ -21,23 +21,27 @@ DB_NAME, COLLECTION_NAME = NAMESPACE.split(".")
 
 # Instantiate as constant instead of pytest fixture to prevent needing to make multiple
 # connections.
-TEST_CLIENT: MongoClient = MongoClient(CONNECTION_STRING)
-collection = TEST_CLIENT[DB_NAME][COLLECTION_NAME]
+
+
+@pytest.fixture
+def collection() -> Any:
+    test_client = MongoClient(CONNECTION_STRING)
+    return test_client[DB_NAME][COLLECTION_NAME]
 
 
 class TestMongoDBAtlasVectorSearch:
     @classmethod
-    def setup_class(cls) -> None:
+    def setup_class(cls, collection: Any) -> None:
         # insure the test collection is empty
         assert collection.count_documents({}) == 0  # type: ignore[index]  # noqa: E501
 
     @classmethod
-    def teardown_class(cls) -> None:
+    def teardown_class(cls, collection: Any) -> None:
         # delete all the documents in the collection
         collection.delete_many({})  # type: ignore[index]
 
     @pytest.fixture(autouse=True)
-    def setup(self) -> None:
+    def setup(self, collection: Any) -> None:
         # delete all the documents in the collection
         collection.delete_many({})  # type: ignore[index]
 

libs/langchain/tests/integration_tests/vectorstores/test_opensearch.py

@@ -1,8 +1,6 @@
 """Test OpenSearch functionality."""
 
-import boto3
 import pytest
-from opensearchpy import AWSV4SignerAuth
 
 from langchain.docstore.document import Document
 from langchain.vectorstores.opensearch_vector_search import (
@@ -219,6 +217,9 @@ def test_opensearch_with_custom_field_name_appx_false() -> None:
 
 def test_opensearch_serverless_with_scripting_search_indexing_throws_error() -> None:
     """Test to validate indexing using Serverless without Approximate Search."""
+    import boto3
+    from opensearchpy import AWSV4SignerAuth
+
     region = "test-region"
     service = "aoss"
     credentials = boto3.Session().get_credentials()
@@ -235,6 +236,9 @@ def test_opensearch_serverless_with_scripting_search_indexing_throws_error() ->
 
 def test_opensearch_serverless_with_lucene_engine_throws_error() -> None:
     """Test to validate indexing using lucene engine with Serverless."""
+    import boto3
+    from opensearchpy import AWSV4SignerAuth
+
     region = "test-region"
     service = "aoss"
     credentials = boto3.Session().get_credentials()

libs/langchain/tests/integration_tests/vectorstores/test_pinecone.py

@@ -2,16 +2,18 @@ import importlib
 import os
 import time
 import uuid
-from typing import List
+from typing import TYPE_CHECKING, List
 
 import numpy as np
-import pinecone
 import pytest
 
 from langchain.docstore.document import Document
 from langchain.embeddings import OpenAIEmbeddings
 from langchain.vectorstores.pinecone import Pinecone
 
+if TYPE_CHECKING:
+    import pinecone
+
 index_name = "langchain-test-index"  # name of the index
 namespace_name = "langchain-test-namespace"  # name of the namespace
 dimension = 1536  # dimension of the embeddings
@@ -32,10 +34,12 @@ def reset_pinecone() -> None:
 
 
 class TestPinecone:
-    index: pinecone.Index
+    index: "pinecone.Index"
 
     @classmethod
     def setup_class(cls) -> None:
+        import pinecone
+
         reset_pinecone()
 
         cls.index = pinecone.Index(index_name)

libs/langchain/tests/integration_tests/vectorstores/test_rocksetdb.py

@@ -1,9 +1,6 @@
 import logging
 import os
 
-import rockset
-import rockset.models
-
 from langchain.docstore.document import Document
 from langchain.vectorstores.rocksetdb import Rockset
 from tests.integration_tests.vectorstores.fake_embeddings import (
@@ -44,6 +41,9 @@ class TestRockset:
 
     @classmethod
     def setup_class(cls) -> None:
+        import rockset
+        import rockset.models
+
         assert os.environ.get("ROCKSET_API_KEY") is not None
         assert os.environ.get("ROCKSET_REGION") is not None
 

libs/langchain/tests/integration_tests/vectorstores/test_weaviate.py

@@ -5,7 +5,6 @@ import uuid
 from typing import Generator, Union
 
 import pytest
-from weaviate import Client
 
 from langchain.docstore.document import Document
 from langchain.embeddings.openai import OpenAIEmbeddings
@@ -29,6 +28,8 @@ class TestWeaviate:
     @pytest.fixture(scope="class", autouse=True)
     def weaviate_url(self) -> Union[str, Generator[str, None, None]]:
         """Return the weaviate url."""
+        from weaviate import Client
+
         url = "http://localhost:8080"
         yield url