AzureSearch vectorstore does not work asyncronously

[thedavgar]

Checked other resources

• I added a very descriptive title to this issue.
• I searched the LangChain documentation with the integrated search.
• I used the GitHub search to find a similar question and didn't find it.
• I am sure that this is a bug in LangChain rather than my code.
• The bug is not resolved by updating to the latest stable version of LangChain (or the specific integration package).

Example Code

I am trying to use the Azure AI Search Vectorstore and retriever and the vectorstore and retriever (given from the vectorstore) work perfectly when doing retrieval of documents using the syncronous methods but gives an error when trying to run the async methods.

Creating the instances of embeddings and Azure Search

from azure.search.documents.indexes.models import (
    SearchField,
    SearchFieldDataType,
    SimpleField,
)
from langchain_openai import AzureOpenAIEmbeddings
from langchain_community.vectorstores import AzureSearch

fields = [
  SimpleField(
      name="content",
      type=SearchFieldDataType.String,
      key=True,
      filterable=True,
  ),
  SearchField(
      name="content_vector",
      type=SearchFieldDataType.Collection(SearchFieldDataType.Single),
      searchable=True,
      vector_search_dimensions=1536,
      vector_search_profile_name="myHnswProfile",
  ),
  SimpleField(
      name="document_name",
      type=SearchFieldDataType.String,
      key=True,
      filterable=True,
  )
]

encoder = AzureOpenAIEmbeddings(
    azure_endpoint=os.getenv("EMBEDDINGS_OPENAI_ENDPOINT"),
    deployment=os.getenv("EMBEDDINGS_DEPLOYMENT_NAME"),
    openai_api_version=os.getenv("OPENAI_API_VERSION"),
    openai_api_key=os.getenv("AZURE_OPENAI_API_KEY"),
)

vectorstore = AzureSearch(
    azure_search_endpoint=os.getenv("AI_SEARCH_ENDPOINT_SECRET"),
    azure_search_key=os.getenv("AI_SEARCH_API_KEY"),
    index_name=os.getenv("AI_SEARCH_INDEX_NAME_SECRET"),
    fields=fields,
    embedding_function=encoder,
)

retriever = vectorstore.as_retriever(search_type="hybrid", k=2)

Syncronous methods working and returning documents

vectorstore.vector_search("what is the capital of France")
retriever.invoke("what is the capital of France")

Asyncronous methods working and returning documents

await vectorstore.avector_search("what is the capital of France")
await retriever.ainvoke("what is the capital of France")

Error Message and Stack Trace (if applicable)

---

KeyError Traceback (most recent call last)
Cell In[15], line 1
----> 1 await vectorstore.avector_search("what is the capital of France")

File ~/.local/lib/python3.11/site-packages/langchain_community/vectorstores/azuresearch.py:695, in AzureSearch.avector_search(self, query, k, filters, **kwargs)
682 async def avector_search(
683 self, query: str, k: int = 4, *, filters: Optional[str] = None, **kwargs: Any
684 ) -> List[Document]:
685 """
686 Returns the most similar indexed documents to the query text.
687
(...)
693 List[Document]: A list of documents that are most similar to the query text.
694 """
--> 695 docs_and_scores = await self.avector_search_with_score(
696 query, k=k, filters=filters
697 )
698 return [doc for doc, _ in docs_and_scores]

File ~/.local/lib/python3.11/site-packages/langchain_community/vectorstores/azuresearch.py:742, in AzureSearch.avector_search_with_score(self, query, k, filters, **kwargs)
730 """Return docs most similar to query.
731
732 Args:
(...)
739 to the query and score for each
740 """
741 embedding = await self._aembed_query(query)
--> 742 docs, scores, _ = await self._asimple_search(
743 embedding, "", k, filters=filters, **kwargs
744 )
746 return list(zip(docs, scores))

File ~/.local/lib/python3.11/site-packages/langchain_community/vectorstores/azuresearch.py:1080, in AzureSearch._asimple_search(self, embedding, text_query, k, filters, **kwargs)
1066 async with self._async_client() as async_client:
1067 results = await async_client.search(
1068 search_text=text_query,
1069 vector_queries=[
(...)
1078 **kwargs,
1079 )
-> 1080 docs = [
1081 (
1082 _result_to_document(result),
1083 float(result["@search.score"]),
1084 result[FIELDS_CONTENT_VECTOR],
1085 )
1086 async for result in results
1087 ]
1088 if not docs:
1089 raise ValueError(f"No {docs=}")

File ~/.local/lib/python3.11/site-packages/langchain_community/vectorstores/azuresearch.py:1084, in (.0)
1066 async with self._async_client() as async_client:
1067 results = await async_client.search(
1068 search_text=text_query,
1069 vector_queries=[
(...)
1078 **kwargs,
1079 )
1080 docs = [
1081 (
1082 _result_to_document(result),
1083 float(result["@search.score"]),
-> 1084 result[FIELDS_CONTENT_VECTOR],
1085 )
1086 async for result in results
1087 ]
1088 if not docs:
1089 raise ValueError(f"No {docs=}")

KeyError: 'content_vector'

Description

The async methods for searching documents (at least) do not work and raise an error. The async client is not being used for async methods for retrieval possibly.

System Info

langchain==0.2.6
langchain-community==0.2.4
langchain-core==0.2.11
langchain-openai==0.1.8
langchain-text-splitters==0.2.1