Pydantic v2 support / pydantic<2 dependency restriction

[MithunEndee]

Hi maintainers,

I noticed that VectorDBBench currently restricts the dependency to:

pydantic < 2

Since many libraries and environments are moving toward Pydantic v2, this version pin can sometimes create dependency conflicts when setting up VectorDBBench in newer environments.

As a small example from our side — we recently contributed a PR that was merged (thanks for that!). During integration we noticed that Endee was already using the latest Pydantic version on the client side, so to keep everything working smoothly with VectorDBBench we temporarily downgraded Pydantic. It worked for now, but we thought it might be helpful to raise this as a broader compatibility discussion.

Would it be possible to consider:

• Adding support for Pydantic v2, or
• Relaxing the version constraint if it’s feasible?

Totally understand that v2 introduces breaking changes, and this is just a friendly suggestion to help long-term ecosystem compatibility.

Thanks again for maintaining this project!

[liliu-z]

Hi @MithunEndee, thanks for raising this!

We agree that supporting Pydantic v2 would be beneficial for long-term compatibility. We've looked into it and it's definitely doable, but the migration is non-trivial — there are ~76 files using Pydantic v1 APIs (@validator, .dict(), SecretStr, etc.) that would need updating, plus the endee dependency currently requires Pydantic v1 as well.

We have this on our radar and will work on it when bandwidth allows. That said, contributions are very welcome! If you or anyone in the community would like to take a stab at it, we'd be happy to review and merge a PR.

For reference, the main changes needed are:

• @validator → @field_validator
• .dict() → .model_dump()
• Config inner class → model_config = ConfigDict(...)
• Update endee to a Pydantic v2-compatible version

Thanks again for the suggestion!

[liliu-z]

/keep-open

[XuanYang-cn]

@fer-marino migrate pydantic to v2 in this PR#674, but it is stale for a long time. So I'm going to do the migration myself soon.

[MithunEndee]

Thanks for picking this up!

The DB client side can also handle this via compatibility layers—if native support is added here, that’d be great. We’ve been working on making our client compatible with both pydantic v1 and v2 and will be updating it soon.