[Bugfix][Frontend] validate arg priority in frontend LLM class before add request

[junpuf]

Purpose

FIX #27595

Now:

    outputs = llm.generate(promtps, priority=[0, 0, 0])
              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/ec2-user/inference_playground/vllm/vllm/entrypoints/llm.py", line 432, in generate
    self._validate_and_add_requests(
  File "/home/ec2-user/inference_playground/vllm/vllm/entrypoints/llm.py", line 1591, in _validate_and_add_requests
    priority=priority[i] if priority else 0,
             ~~~~~~~~^^^
IndexError: list index out of range

After fix:

    outputs = llm.generate(promtps, priority=[0, 0, 0])
              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/ec2-user/inference_playground/vllm/vllm/entrypoints/llm.py", line 432, in generate
    self._validate_and_add_requests(
  File "/home/ec2-user/inference_playground/vllm/vllm/entrypoints/llm.py", line 1568, in _validate_and_add_requests
    raise ValueError("The lengths of prompts and priority must be the same.")
ValueError: The lengths of prompts and priority must be the same.

Test Plan

pytest -s -v tests/entrypoints/llm/test_generate.py

Test Result

Unit Test

Pass

Full CI Test:

• 6b286b7: An unrelated test failed test_chat_utils.py::test_resolve_content_format_fallbacks[deepseek-ai/deepseek-vl2-tiny-string which is fixed by [Bugfix][CI] Fix config resolving logic with remote models #27610
• 0073e9e Consumed test fix, but ubuntu was down
• 2f09f35 ubuntu seems to be recovered from outage, merge again to trigger CI, and got hit by second outage of the day
• fb2a80b ubuntu say its resolved, and failed again.

---

Essential Elements of an Effective PR Description Checklist

• [x ] The purpose of the PR, such as "Fix some issue (link existing issues this PR will resolve)".
• [x ] The test plan, such as providing test command.
• [x ] The test results, such as pasting the results comparison before and after, or e2e results
• (Optional) The necessary documentation update, such as updating supported_models.md and examples for a new model.
• (Optional) Release notes update. If your change is user facing, please update the release notes draft in the Google Doc.

[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 ask your reviewers to trigger select CI tests on top of fastcheck CI.

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.

If you have any questions, please reach out to us on Slack at https://slack.vllm.ai.

🚀

[gemini-code-assist]

Code Review

This pull request addresses a bug where the priority argument in the llm.generate function was not validated for length against the prompts argument, leading to an IndexError. The fix adds a check to ensure that the lengths of prompts and priority are the same, raising a ValueError if they differ. Additionally, a test case has been added to verify this validation.

[njhill]

Thanks @junpuf!

I guess we don't support priority here being a single int shared by all the requests?

[junpuf]

Thanks @junpuf!

I guess we don't support priority here being a single int shared by all the requests?

Looks like so, for now if priority is None then 0 will be used.

[DarkLight1337]

Can you merge with main to fix CI?

Sorry I didn't see your latest merge

[junpuf]

Can you merge with main to fix CI?

Branch is up to date with main, thanks for checking in.