Rm retriever kwargs (#7013)

Doesn't actually limit the Retriever interface but hopefully in practice
it does

tests/integration_tests/chains/test_retrieval_qa.py

@@ -7,18 +7,19 @@ from langchain.document_loaders import TextLoader
 from langchain.embeddings.openai import OpenAIEmbeddings
 from langchain.llms import OpenAI
 from langchain.text_splitter import CharacterTextSplitter
-from langchain.vectorstores import Chroma
+from langchain.vectorstores import FAISS
 
 
 def test_retrieval_qa_saving_loading(tmp_path: Path) -> None:
     """Test saving and loading."""
-    loader = TextLoader("docs/modules/state_of_the_union.txt")
+    loader = TextLoader("docs/extras/modules/state_of_the_union.txt")
     documents = loader.load()
     text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=0)
     texts = text_splitter.split_documents(documents)
     embeddings = OpenAIEmbeddings()
-    docsearch = Chroma.from_documents(texts, embeddings)
+    docsearch = FAISS.from_documents(texts, embeddings)
     qa = RetrievalQA.from_llm(llm=OpenAI(), retriever=docsearch.as_retriever())
+    qa.run("What did the president say about Ketanji Brown Jackson?")
 
     file_path = tmp_path / "RetrievalQA_chain.yaml"
     qa.save(file_path=file_path)

tests/integration_tests/retrievers/test_contextual_compression.py

@@ -1,7 +1,7 @@
 from langchain.embeddings import OpenAIEmbeddings
 from langchain.retrievers.contextual_compression import ContextualCompressionRetriever
 from langchain.retrievers.document_compressors import EmbeddingsFilter
-from langchain.vectorstores import Chroma
+from langchain.vectorstores import FAISS
 
 
 def test_contextual_compression_retriever_get_relevant_docs() -> None:
@@ -13,7 +13,7 @@ def test_contextual_compression_retriever_get_relevant_docs() -> None:
     ]
     embeddings = OpenAIEmbeddings()
     base_compressor = EmbeddingsFilter(embeddings=embeddings, similarity_threshold=0.75)
-    base_retriever = Chroma.from_texts(texts, embedding=embeddings).as_retriever(
+    base_retriever = FAISS.from_texts(texts, embedding=embeddings).as_retriever(
         search_kwargs={"k": len(texts)}
     )
     retriever = ContextualCompressionRetriever(

tests/unit_tests/retrievers/test_base.py

@@ -2,7 +2,7 @@
 
 from __future__ import annotations
 
-from typing import Any, Dict, List, Optional
+from typing import Dict, List, Optional
 
 import pytest
 
@@ -141,45 +141,48 @@ async def test_fake_retriever_v1_with_kwargs_upgrade_async(
     assert callbacks.retriever_errors == 0
 
 
+class FakeRetrieverV2(BaseRetriever):
+    def __init__(self, throw_error: bool = False) -> None:
+        self.throw_error = throw_error
+
+    def _get_relevant_documents(
+        self, query: str, *, run_manager: CallbackManagerForRetrieverRun | None
+    ) -> List[Document]:
+        assert isinstance(self, FakeRetrieverV2)
+        assert run_manager is not None
+        assert isinstance(run_manager, CallbackManagerForRetrieverRun)
+        if self.throw_error:
+            raise ValueError("Test error")
+        return [
+            Document(page_content=query),
+        ]
+
+    async def _aget_relevant_documents(
+        self, query: str, *, run_manager: AsyncCallbackManagerForRetrieverRun | None
+    ) -> List[Document]:
+        assert isinstance(self, FakeRetrieverV2)
+        assert run_manager is not None
+        assert isinstance(run_manager, AsyncCallbackManagerForRetrieverRun)
+        if self.throw_error:
+            raise ValueError("Test error")
+        return [
+            Document(page_content=f"Async query {query}"),
+        ]
+
+
 @pytest.fixture
 def fake_retriever_v2() -> BaseRetriever:
-    class FakeRetrieverV2(BaseRetriever):
-        def _get_relevant_documents(
-            self,
-            query: str,
-            *,
-            run_manager: CallbackManagerForRetrieverRun | None = None,
-            **kwargs: Any,
-        ) -> List[Document]:
-            assert isinstance(self, FakeRetrieverV2)
-            assert run_manager is not None
-            assert isinstance(run_manager, CallbackManagerForRetrieverRun)
-            if "throw_error" in kwargs:
-                raise ValueError("Test error")
-            return [
-                Document(page_content=query, metadata=kwargs),
-            ]
-
-        async def _aget_relevant_documents(
-            self,
-            query: str,
-            *,
-            run_manager: AsyncCallbackManagerForRetrieverRun | None = None,
-            **kwargs: Any,
-        ) -> List[Document]:
-            assert isinstance(self, FakeRetrieverV2)
-            assert run_manager is not None
-            assert isinstance(run_manager, AsyncCallbackManagerForRetrieverRun)
-            if "throw_error" in kwargs:
-                raise ValueError("Test error")
-            return [
-                Document(page_content=f"Async query {query}", metadata=kwargs),
-            ]
-
     return FakeRetrieverV2()  # type: ignore[abstract]
 
 
-def test_fake_retriever_v2(fake_retriever_v2: BaseRetriever) -> None:
+@pytest.fixture
+def fake_erroring_retriever_v2() -> BaseRetriever:
+    return FakeRetrieverV2(throw_error=True)  # type: ignore[abstract]
+
+
+def test_fake_retriever_v2(
+    fake_retriever_v2: BaseRetriever, fake_erroring_retriever_v2: BaseRetriever
+) -> None:
     callbacks = FakeCallbackHandler()
     assert fake_retriever_v2._new_arg_supported is True
     results = fake_retriever_v2.get_relevant_documents("Foo", callbacks=[callbacks])
@@ -187,20 +190,17 @@ def test_fake_retriever_v2(fake_retriever_v2: BaseRetriever) -> None:
     assert callbacks.retriever_starts == 1
     assert callbacks.retriever_ends == 1
     assert callbacks.retriever_errors == 0
-    results2 = fake_retriever_v2.get_relevant_documents(
-        "Foo", callbacks=[callbacks], foo="bar"
-    )
-    assert results2[0].metadata == {"foo": "bar"}
+    fake_retriever_v2.get_relevant_documents("Foo", callbacks=[callbacks])
 
     with pytest.raises(ValueError, match="Test error"):
-        fake_retriever_v2.get_relevant_documents(
-            "Foo", callbacks=[callbacks], throw_error=True
-        )
+        fake_erroring_retriever_v2.get_relevant_documents("Foo", callbacks=[callbacks])
     assert callbacks.retriever_errors == 1
 
 
 @pytest.mark.asyncio
-async def test_fake_retriever_v2_async(fake_retriever_v2: BaseRetriever) -> None:
+async def test_fake_retriever_v2_async(
+    fake_retriever_v2: BaseRetriever, fake_erroring_retriever_v2: BaseRetriever
+) -> None:
     callbacks = FakeCallbackHandler()
     assert fake_retriever_v2._new_arg_supported is True
     results = await fake_retriever_v2.aget_relevant_documents(
@@ -210,11 +210,8 @@ async def test_fake_retriever_v2_async(fake_retriever_v2: BaseRetriever) -> None
     assert callbacks.retriever_starts == 1
     assert callbacks.retriever_ends == 1
     assert callbacks.retriever_errors == 0
-    results2 = await fake_retriever_v2.aget_relevant_documents(
-        "Foo", callbacks=[callbacks], foo="bar"
-    )
-    assert results2[0].metadata == {"foo": "bar"}
+    await fake_retriever_v2.aget_relevant_documents("Foo", callbacks=[callbacks])
     with pytest.raises(ValueError, match="Test error"):
-        await fake_retriever_v2.aget_relevant_documents(
-            "Foo", callbacks=[callbacks], throw_error=True
+        await fake_erroring_retriever_v2.aget_relevant_documents(
+            "Foo", callbacks=[callbacks]
         )