Got Multiple Values for Keyword Argument 'k' in as_retriever Function

[Panegyric0926]

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

from langchain_openai import AzureChatOpenAI
from langchain_openai  import AzureOpenAIEmbeddings
from langchain.vectorstores.azuresearch import AzureSearch
from langchain.chains import RetrievalQA

llm = AzureChatOpenAI(deployment_name=aoai_deployment_id, temperature=0)

embeddings = AzureOpenAIEmbeddings(deployment='text-embedding-3-large',
                                   model='text-embedding-3-large',
                                   azure_endpoint='https://cog-openai-od-ed.openai.azure.com',
                                   openai_api_type="azure",
                                   chunk_size=1)

acs = AzureSearch(azure_search_endpoint = endpoint,
                  azure_search_key = cog_search_api_key,
                  index_name = index_name,
                  embedding_function = embeddings.embed_query,
                  scoring_profiles = [sc],
                  default_scoring_profile = sc_name)

qa_chain = RetrievalQA.from_chain_type(
    llm,
    retriever = acs.as_retriever(search_type="similarity",search_kwargs={"filter":filters, "k":8}),
    # retriever = acs.as_retriever(search_kwargs={"filters":filters}),
    chain_type_kwargs={"prompt": template},
    return_source_documents=True)

Description

• I tried the above code, and got an error: TypeError: langchain_community.vectorstores.azuresearch.AzureSearch.vector_search() got multiple values for keyword argument 'k'.
• I deleted the "k" parameters in the code, as shown in the commented line, and the bug disappeared.
• But only 4 documents are returned, which is the default setting for the library. I need 8 documents from the retriever.

System Info

• requirement.txt:
  • azure-core
  • azure-identity
  • azure-search-documents
  • azure-storage-blob
  • langchain
  • langchain_community
  • langchain-openai
  • openai
  • tiktoken
• pip install --upgrade -r .\requirement.txt

[dosubot[bot]]

@Panegyric0926 Hello there! I'm here to help you with any bugs, questions, or contributions you might have. Let's tackle this together!

To resolve the TypeError: langchain_community.vectorstores.azuresearch.AzureSearch.vector_search() got multiple values for keyword argument 'k' error and ensure that 8 documents are returned, you need to set the k parameter directly in the as_retriever function without including it in the search_kwargs.

Here's the corrected code:

from langchain_openai import AzureChatOpenAI
from langchain_openai import AzureOpenAIEmbeddings
from langchain.vectorstores.azuresearch import AzureSearch
from langchain.chains import RetrievalQA

llm = AzureChatOpenAI(deployment_name=aoai_deployment_id, temperature=0)

embeddings = AzureOpenAIEmbeddings(deployment='text-embedding-3-large',
                                   model='text-embedding-3-large',
                                   azure_endpoint='https://cog-openai-od-ed.openai.azure.com',
                                   openai_api_type="azure",
                                   chunk_size=1)

acs = AzureSearch(azure_search_endpoint=endpoint,
                  azure_search_key=cog_search_api_key,
                  index_name=index_name,
                  embedding_function=embeddings.embed_query,
                  scoring_profiles=[sc],
                  default_scoring_profile=sc_name)

qa_chain = RetrievalQA.from_chain_type(
    llm,
    retriever=acs.as_retriever(search_type="similarity", k=8, search_kwargs={"filter": filters}),
    chain_type_kwargs={"prompt": template},
    return_source_documents=True
)

This modification ensures that the k parameter is set to 8 directly in the as_retriever function, avoiding the TypeError and returning the desired number of documents ^[1].

To continue talking to Dosu, mention @dosu.

[JuliaZamaitat]

Hi there, I have got the same issues, except it's also being an issue for the filter keyword:

TypeError: azure.search.documents._search_client.SearchClient.search() got multiple values for keyword argument 'filter'

The implementation is as you provided. If I pass it in directly without the search_kwargs, the filter does not get applied...

Any help/suggestions how to continue with this issue? Maybe the problem is the filter itself, is there any working example you could provide for it?

Thank you!

[PPraneesh]

Hey, did you solved this issue?

[ibrahoss]

I have a similar issue with the filter keyword, (azure.search.documents._search_client.SearchClient.search() got multiple values for keyword argument 'filter'.). Did anyone solve it?

[ibrahoss]

Ok, I fixed it by using "filters" instead of "filter" in the search_kwargs. filter causes a conflict, and the documentation was misleading by instructing to use filter, it needs to be updated.