Bedrock Knowledge Base Search Filter "in" broken

[manuwaik]

https://github.com/langchain-ai/langchain-aws/blame/d89fcf80c61174c21c88ef8d1cfbf3a55c7d84c5/libs/aws/langchain_aws/retrievers/bedrock.py#L24

I get error:

Parameter validation failed:
Unknown parameter in retrievalConfiguration.vectorSearchConfiguration.filter: "in_", must be one of: andAll, equals, greaterThan, greaterThanOrEquals, in, lessThan, lessThanOrEquals, listContains, notEquals, notIn, orAll, startsWith, stringContains

[isahers1]

Could you show a full code snippet that is causing the error so I can help debug?

[manuwaik]

# Retrieve chunks from relevant documents
retriever = AmazonKnowledgeBasesRetriever(
    knowledge_base_id="XXXXXXXX",
    retrieval_config={"vectorSearchConfiguration": 
                        {
                            "numberOfResults": 5,
                            "filter": {
                                "in": {
                                    "key": "tag",
                                    "value": ["document1", "document2"]
                                    }
                                }
                        }
                     },
)

docs = retriever.invoke("How many stars are in the universe?")

The metadata with key tag exists on the KB.
I have also tried using in_.

[karoaper]

The problem is here: https://github.com/langchain-ai/langchain-aws/blob/main/libs/aws/langchain_aws/retrievers/bedrock.py#L165. The call to model_dump should include "by_alias=True".

[renjiexu-amzn]

Oh, karoaper beated me on commenting the root cause. Would @isahers1 help merge my change in #233 and release a new minor version with the fix? Do you have an ETA on the release so that I could coordinate the work blocked by this one?

[3coins]

Fixed in v0.2.3.