[RFC]: Add validator and parser in `@stdlib/bench/*` for benchmark description

[Vinit-Pandit]

Description

This RFC proposes adding Validator and Parser for the benchmark description ( EX : # @stdlib/blas/ddot::stacks:size=1000,ndims=2,shape=(2,500) ) in @stdlib/bench/*

Until now, we have been using this descriptive format unconsciously. However, since these descriptors are crucial for visualizing the benchmarks, we need to formalize this approach by including validation and a parser that converts the description into JSON.

Related Issues

Related issues #6

Questions

No.

Other

No.

Checklist

• I have read and understood the Code of Conduct.
• Searched for existing issues and pull requests.
• The issue name begins with RFC:.

[Vinit-Pandit]

Few questions

1. Do we need to validate the package name (e.g., @stdlib/blas/ddot)? If so, what should we use for validation? We can’t use the namespace since they’re only for pre-registered packages.

2. Should this be implemented in@stdlib/bench/utils ?

[riyap03]

Hi everyone 👋
I'm Riya, a second-year CSE student from India passionate about open source and GSoC 2026 prep.

This RFC caught my interest — I’d love to contribute to the parser + validator idea for benchmark descriptions.

I had a few clarifying questions before I start:

Should the validator support only currently existing packages in @stdlib/, or can it be generic?

Should I place the parser/validator under @stdlib/bench/utils or create a new submodule?

Looking forward to your thoughts! 😊

– Riya
GitHub: @riyap03

[kgryte]

I think having a parser would be useful. As a first step, we could add a utility package to @stdlib/bench/utils/parse-description which takes a string as input and returns a JSON object with appropriate fields.

Prior to adding the utility, one should catalogue a wide variety of descriptions as used throughout the project so that we can be sure that the parser can adequately handle all of our use cases.

[riyap03]

I’ll start by going through @stdlib/bench/* and compiling a list of all the benchmark descriptions I find (including the tricky/edge cases). Once that’s ready, I can share the collection here so we can review patterns before jumping into the parser implementation.

Thanks for the direction! Excited to get my hands dirty with parse-description once we have the full picture 📚💙

– Riya