[Bug] Sparse Index Query Not Supported with QueryOptions type

[harrisonlinowes]

Is this a new bug?
In other words: Is this an error, flaw, failure or fault? Please search Github issues and check our Community Forum to see if someone has already reported the bug you encountered.

If this is a request for help or troubleshooting code in your own Pinecone project, please join the Pinecone Community Forum.

• I believe this is a new bug
• I have searched the existing Github issues and Community Forum, and I could not find an existing post for this bug

Describe the bug
As of December of 2024 the recommended approach for hybrid search is to use cascading retrieval via a separate dense and sparse index (blog post). To implement this in typescript the SDK must support querying using only the sparseVector parameter.

Currently the query function is typed in a way that requires vector to be supplied. The QueryByVectorValues type must make vector parameter optional.

Error information
"my-sparse-index" must be a sparse index.

const index = pc.index<T>("my-sparse-index");
const res = await index.query({
  vector: [], // required to satisfy typing
  sparseVector: sparseVector.vectors[0],
  topK: 10,
});

Results in:
PineconeArgumentError: You must enter an array of `RecordValues` in order to query by vector values.

Steps to reproduce the issue locally
Include steps to reproduce the issue here. If you have sample code or a script that can be used to replicate this issue, please include that as well (including any dependent files to run the code).

See section above.

Environment

• OS and version: Mac
• Node version: v20.18.1
• Typescript SDK version: "@pinecone-database/pinecone": "^5.1.1",

Additional context
Current work around:

const index = pc.index<T>("my-sparse-index");
// @ts-ignore
const res = await index.query({
  sparseVector: sparseVector.vectors[0],
  topK: 10,
});