AzureCosmosDBVectorSearch: 'vectorContent' error when executing .as_retriever() with no full traceback

[mnbenavides]

Checked other resources

• I added a very descriptive title to this question.
• I searched the LangChain documentation with the integrated search.
• I used the GitHub search to find a similar question and didn't find it.

Commit to Help

• I commit to help with one of those options 👆

Example Code

# Connection setup
client = MongoClient(cosmos_uri)
db = client["bd"]
collection = db["bd_prueba"]

Vector Index Configuration:

{
  "key": { "vectorContent": "cosmosSearch" },
  "cosmosSearchOptions": {
    "kind": "vector-diskann",
    "dimensions": 768,
    "similarity": "COS"
  }
}

Create e,beddings:
 embedder.embed_query(content)


Store database:

   mongo_doc = {
             "textContent": page_content,
             "vectorContent": embedding,
              "metadata": doc.metadata,
     }

  docs_to_insert.append(mongo_doc)
 collection.insert_many(docs_to_insert)


# Vector store setup
vector_store = AzureCosmosDBVectorSearch(
    collection=collection,
    index_name='vector_index',
    text_key='textContent',
    embedding_key='vectorContent',
    embedding=get_azure_embedding_function()
)

retriever = vector_store.as_retriever(
    search_type="mmr",
    search_kwargs={"k": 5, "fetch_k": 40, "lambda_mult": 0.7}
)

qa_chain = RetrievalQA.from_chain_type(
    llm=llm,
    retriever=retriever,
    return_source_documents=True,
    chain_type="map_reduce",
    chain_type_kwargs={
        "question_prompt": question_prompt,
        "combine_prompt": combine_prompt
    }
)

# Error is triggered here
result = await qa_chain.ainvoke("What is the triggering event for the ICA tax?")

Description

I'm using AzureCosmosDBVectorSearch with Azure Cosmos DB (MongoDB API, vCore). When I execute a retrieval query using .as_retriever(), I receive the following error message:
'vectorContent'
There’s no full traceback or exception context. The error occurs during the call to qa_chain.ainvoke(question)

The vector search should retrieve relevant documents using the "vector_index" created on the "bd_prueba" collection. An exception with the message 'vectorContent' is raised without a full stack trace.

System Info

Python:3.12.0

langchain==0.3.23
langchain-openai==0.3.3
langchain-core==0.3.51
langchain-community==0.3.21
langchain-chroma==0.2.1
langchain-mongodb==0.5.0
langchain-text-splitters==0.3.8

Platform: Windows