[Perf] Validate @config in pre-commit instead of dynamically

[lionelvillard]

This PR is about optimizing vLLM cold start by validating the code defining vLLM configuration at static time in precommit instead of at runtime. By doing so, vLLM cold start is reduced by at least 500ms per worker.

Before this PR:

After this PR:

@ProExpertProg @aarnphm @hmellor @mgoin

[github-actions]

👋 Hi! Thank you for contributing to the vLLM project.

💬 Join our developer Slack at https://slack.vllm.ai to discuss your PR in #pr-reviews, coordinate on features in #feat- channels, or join special interest groups in #sig- channels.

Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run fastcheck CI which starts running only a small and essential subset of CI tests to quickly catch errors. You can run other CI tests on top of those by going to your fastcheck build on Buildkite UI (linked in the PR checks section) and unblock them. If you do not have permission to unblock, ping simon-mo or khluu to add you in our Buildkite org.

Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging.

To run CI, PR reviewers can either: Add ready label to the PR or enable auto-merge.

🚀

[gemini-code-assist]

Summary of Changes

Hello @lionelvillard, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request refactors how configuration validation is performed by shifting it from runtime checks to a pre-commit static analysis. This change aims to improve code quality and consistency by catching configuration-related issues earlier in the development cycle, ensuring that all configuration fields are properly defined and documented.

Highlights

• Shifted Configuration Validation: I've moved the validation logic for @config decorated classes from dynamic runtime checks within vllm/config.py to a static pre-commit hook. This ensures that configuration fields adhere to required standards (default values, docstrings, correct Literal usage) before code is committed.
• New Static Analysis Tool: I've introduced a new Python script, tools/validate_config.py, which leverages Python's ast module to perform static analysis on configuration files. This script programmatically checks for the presence of default values and docstrings for fields in classes decorated with @config and @dataclass, and also validates the correct usage of Literal types.
• Pre-commit Hook Integration: A new validate-config hook has been added to .pre-commit-config.yaml. This hook executes the tools/validate_config.py script against vllm/config.py (and tests/test_config.py, though not shown in the patch) during the pre-commit stage, enforcing configuration standards early in the development workflow.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point in your pull request via creating an issue comment (i.e. comment on the pull request page) using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands.

Feature	Command	Description
Code Review	/gemini review	Performs a code review for the current pull request in its current state.
Pull Request Summary	/gemini summary	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in issue comments and review comments.
Help	/gemini help	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist is currently in preview and may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments to provide feedback.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

Footnotes

1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

[gemini-code-assist]

Code Review

This pull request refactors the validation of @config decorated classes by moving the logic from a runtime decorator to a pre-commit hook. The new validation script tools/validate_config.py is well-structured, but there are suggestions to improve its readability, style, and correctness. The tests for the removed runtime validation in tests/test_config.py should be removed to align with the changes in this PR.

[mgoin]

Looks reasonable to me, very nice! Can you maybe demonstrate speedup by timing the import of vllm.config?

[lionelvillard]

Can you maybe demonstrate speedup by timing the import of vllm.config?

I added 2 screenshots demonstrating the speedup

[mgoin]

Great work, LGTM

[tlrmchlsmth]

Does this file need a "Adapted from https://github.com/..." message? Not sure what the intent of linking this URL is.

[lionelvillard]

sure. The link is needed because of the MIT license:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

[lionelvillard]

/label startup-ux