Add ID field to similarity search retrieved documents in langchain-pinecone

[yahya-mouman]

Checked

• I searched existing ideas and did not find a similar one
• I added a very descriptive title
• I've clearly described the feature request and motivation for it

Feature request

Pinecone made the ID field a requirement to upsert documents so it makes more sense to add it to the similarity search result documents too.

Motivation

I was developing an integration for langchain when I noticed that the ID is never propagated forwards with the results.

Proposal (If applicable)

My proposal is simple. Add the ID field to the similarity_search_by_vector_with_score method result in the langchain-pinecone module

def similarity_search_by_vector_with_score(
        self,
        embedding: List[float],
        *,
        k: int = 4,
        filter: Optional[dict] = None,
        namespace: Optional[str] = None,
    ) -> List[Tuple[Document, float]]:
        """Return pinecone documents most similar to embedding, along with scores."""

        if namespace is None:
            namespace = self._namespace
        docs = []
        results = self._index.query(
            vector=embedding,
            top_k=k,
            include_metadata=True,
            namespace=namespace,
            filter=filter,
        )
        for res in results["matches"]:
            metadata = res["metadata"]
            id = res["id"]
            if self._text_key in metadata:
                text = metadata.pop(self._text_key)
                score = res["score"]
                docs.append((Document(page_content=text, metadata=metadata, id=id), score))
            else:
                logger.warning(
                    f"Found document with no `{self._text_key}` key. Skipping."
                )
        return docs

[yahya-mouman]

I made a PR for the change here link