rerank in hybrid search in qdrant

[railsstudent]

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

elif self.retrieval_mode == RetrievalMode.HYBRID:
            query_dense_embedding = self.embeddings.embed_query(query)
            query_sparse_embedding = self.sparse_embeddings.embed_query(query)
            results = self.client.query_points(
                prefetch=[
                    models.Prefetch(
                        using=self.vector_name,
                        query=query_dense_embedding,
                        filter=filter,
                        limit=k,
                        params=search_params,
                    ),
                    models.Prefetch(
                        using=self.sparse_vector_name,
                        query=models.SparseVector(
                            indices=query_sparse_embedding.indices,
                            values=query_sparse_embedding.values,
                        ),
                        filter=filter,
                        limit=k,
                        params=search_params,
                    ),
                ],
                query=hybrid_fusion or models.FusionQuery(fusion=models.Fusion.RRF),
                **query_options,
            ).points

        else:
            raise ValueError(f"Invalid retrieval mode. {self.retrieval_mode}.")
        return [
            (
                self._document_from_point(
                    result,
                    self.collection_name,
                    self.content_payload_key,
                    self.metadata_payload_key,
                ),
                result.score,
            )
            for result in results
        ]

Description

• This article shows how to perform a hybrid search in Langchain, and it is very useful. https://qdrant.tech/documentation/frameworks/langchain/?q=langchain#next-steps

• This article shows how to perform rerank in hybrid search in qdrant. https://qdrant.tech/documentation/search-precision/reranking-hybrid-search/?q=rerank#rerank. However, I cannot perform rerank after hybrid search in similarity_search_with_score

The document returns the documents and scores, but not the points. Therefore, I cannot use the qdrant client to query points using the late_interaction_embeddings and late_interaction_embedding_model to obtain the ranked points and documents.

System Info

System Information

OS: Linux
OS Version: #49~22.04.1-Ubuntu SMP PREEMPT_DYNAMIC Wed Nov 6 17:42:15 UTC 2
Python Version: 3.10.12 (main, Nov 6 2024, 20:22:13) [GCC 11.4.0]

Package Information

langchain_core: 0.3.29
langchain: 0.3.14
langchain_community: 0.3.14
langsmith: 0.2.10
langchain_ollama: 0.2.2
langchain_openai: 0.3.0
langchain_qdrant: 0.2.0
langchain_text_splitters: 0.3.5
langchain_unstructured: 0.1.6
langgraph_sdk: 0.1.51

Optional packages not installed

langserve

Other Dependencies

aiohttp: 3.11.11
async-timeout: 4.0.3
dataclasses-json: 0.6.7
fastembed: 0.5.0
httpx: 0.27.2
httpx-sse: 0.4.0
jsonpatch: 1.33
langsmith-pyo3: Installed. No version info available.
numpy: 1.26.4
ollama: 0.4.6
onnxruntime: 1.19.2
openai: 1.59.7
orjson: 3.10.14
packaging: 24.2
pydantic: 2.10.5
pydantic-settings: 2.7.1
PyYAML: 6.0.2
qdrant-client: 1.12.2
requests: 2.32.3
requests-toolbelt: 1.0.0
SQLAlchemy: 2.0.37
tenacity: 9.0.0
tiktoken: 0.8.0
typing-extensions: 4.12.2
unstructured-client: 0.29.0
unstructured[all-docs]: Installed. No version info available.
zstandard: Installed. No version info available.