Consolidate core types: move record.py, add structured output types, enforce strict typing

[nv-alicheng]

What does this PR do?

• Move record.py from async_utils/transport/ to core/ — it's a core data type, not transport-specific
• Replace raw str response output with TextModelOutput struct for type safety and downstream metric computation
• Add PromptData and ErrorData msgspec Structs with gc=False, array_like=True for hot-path performance
• Fix threading safety issues in HTTPEndpointClient, SampleEventHandler, EventRecorder, and shutdown_executor()
• Bump default worker init timeout from 30s → 60s

Type of change

• Bug fix
• New feature
• Documentation update
• Refactor/cleanup

Related issues

Testing

• Tests added/updated
• All tests pass locally
• Manual testing completed

Checklist

• Code follows project style
• Pre-commit hooks pass
• Documentation updated (if needed)

[github-actions]

MLCommons CLA bot All contributors have signed the MLCommons CLA ✍️ ✅

[gemini-code-assist]

Summary of Changes

Hello, 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 significantly refactors core data types within the inference endpoint system to improve type safety, performance, and maintainability. By introducing dedicated msgspec.Struct types for model outputs, prompt data, and error information, the system gains clearer data contracts and optimizes serialization/deserialization. The changes also involve relocating a core module and making minor adjustments to enhance threading safety and worker stability.

Highlights

• Core Type Consolidation: The record.py module was moved from async_utils/transport/ to core/, signifying its role as a fundamental data type.
• Structured Output Types: Introduced TextModelOutput, PromptData, and ErrorData as msgspec.Structs with gc=False and array_like=True for enhanced type safety and performance, replacing raw string or dictionary outputs.
• Strict Typing Enforcement: QueryResult and EventRecord now strictly use the new structured types for response_output and data fields, respectively.
• Threading Safety Improvements: Addressed potential threading issues in SingletonMixin by introducing a lock, and noted known limitations in SampleEventHandler and EventRecorder due to their planned deprecation.
• Increased Worker Initialization Timeout: The default worker initialization timeout was increased from 30s to 60s.

🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console.

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 by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

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 pull request 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 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. If you're interested in giving your feedback about your experience with Gemini Code Assist for GitHub and other Google products, sign up here.

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 introduces a significant and valuable refactoring by consolidating core data types. Moving record.py to core/ improves the project structure. The replacement of raw string outputs with the new TextModelOutput struct, along with the addition of PromptData and ErrorData, greatly enhances type safety and performance. The threading safety fix in SingletonMixin is also a welcome improvement. The changes are applied consistently across the codebase, including extensive updates to tests. I've found a couple of minor issues with incorrect comments in the SGLang adapter tests, which I've flagged for correction. Overall, this is a high-quality contribution that improves the maintainability and robustness of the system.