Ability to disable rate limiting for Nvidia API

[jamesbraza]

When using Nvidia DGX Cloud Lepton (added in #1218), we don't need 40 RPM limit

[dosubot]

Related Documentation

Checked 1 published document(s) in 1 knowledge base(s). No updates required.

^{How did I do? Any feedback?}  

[Copilot]

Pull request overview

This PR adds the ability to disable rate limiting for the Nvidia API when using Nvidia DGX Cloud Lepton. The key change is making the rate limit configurable by adding a new rate_limit parameter to the _call_nvidia_api function, which defaults to the existing 40 RPM limit but can be set to None to disable rate limiting.

• Added a configurable rate_limit parameter to _call_nvidia_api with default value maintaining backward compatibility
• Renamed the rate limit constant to NVIDIA_API_NEMOTRON_PARSE_RATE_LIMIT for clarity
• Added limits and numpy to the typing extra dependencies for proper type checking

Reviewed changes

Copilot reviewed 2 out of 3 changed files in this pull request and generated no comments.

File	Description
packages/paper-qa-nemotron/src/paperqa_nemotron/api.py	Added rate_limit parameter to _call_nvidia_api, renamed rate limit constant, and updated documentation
packages/paper-qa-nemotron/pyproject.toml	Added limits and numpy to typing dependencies for type checking
uv.lock	Updated lock file to reflect new typing dependencies

Comments suppressed due to low confidence (1)

packages/paper-qa-nemotron/src/paperqa_nemotron/api.py:203

• The overload signature is missing the rate_limit parameter that was added to the main function implementation. This creates an inconsistency between the overload type hints and the actual function signature. Add rate_limit: \"RateLimitItem | str | None\" = ... to this overload (and the other two overloads at lines 207 and 218) to maintain API consistency.

async def _call_nvidia_api(
    image: "np.ndarray",
    tool_name: Literal["markdown_bbox"],
    api_key: str | None = None,
    api_base: str = ...,
    model_name: str = ...,
    **completion_kwargs,
) -> list[NemotronParseMarkdownBBox]: ...

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.