[Xunzhuo]

This adds Rust frontend parity for OpenAI n > 1 on /v1/completions and /v1/chat/completions.

This PR was prepared with AI assistance.

Highlights:

• Removes the previous Only n=1 is supported compatibility rejection.
• Fan-outs n choices into child frontend requests, using child request IDs only when n > 1 so existing n=1 request-id behavior is preserved.
• Offsets the seed by choice index to match vLLM/Python parallel sampling behavior.
• Aggregates non-streaming choices and usage with prompt tokens counted once and completion tokens summed across choices.
• Merges streaming child streams into one SSE stream with indexed choices, one final usage chunk when requested, and one [DONE] sentinel.
• Adds route-level fan-out tests that verify child request IDs and seed offsets for chat and completions.

Test Plan

• Format Rust code.
• Check whitespace/conflict-marker hygiene.
• Verify n > 1 validation and route behavior for chat/completions, both streaming and non-streaming.
• Verify fan-out request IDs and seed offsets reach the mock engine for chat/completions.
• Run the full vllm-server test suite.

Test Result

• cargo fmt
• git diff --check
• cargo test -p vllm-server (178 passed, 0 failed)
• cargo test -p vllm-server non_stream_chat_completions_fans_out_choice_request_ids_and_seeds
• cargo test -p vllm-server non_stream_completions_fans_out_choice_request_ids_and_seeds
• cargo test -p vllm-server non_stream_chat_completions_support_n_greater_than_one
• cargo test -p vllm-server stream_chat_completions_support_n_greater_than_one
• cargo test -p vllm-server non_stream_completions_support_n_greater_than_one
• cargo test -p vllm-server stream_completions_support_n_greater_than_one

Not run: cargo nextest because cargo-nextest is not installed in this environment.

---

Essential Elements of an Effective PR Description Checklist

• The purpose of the PR.
• User-facing changes.
• How the implementation works at a high level.
• How the PR was tested.

[mergify]

This pull request has merge conflicts that must be resolved before it can be
merged. Please rebase the PR, @Xunzhuo.

https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/working-with-forks/syncing-a-fork

[mergify]

This pull request has merge conflicts that must be resolved before it can be
merged. Please rebase the PR, @Xunzhuo.

https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/working-with-forks/syncing-a-fork

[BugenZhao]

Thanks for the work. The overall implementation LGTM

[BugenZhao]

What about explicitly specify SelectAll<ChatCompletionEventStream> here so the semantics are more clear?

[Tamogh123]

Maybe renaming variable will make the variable's purpose clearer?

[github-actions]

This pull request has been automatically marked as stale because it has not had any activity within 90 days. It will be automatically closed if no further activity occurs within 30 days. Leave a comment if you feel this pull request should remain open. Thank you!

[github-actions]

This pull request has been automatically closed due to inactivity. Please feel free to reopen if you intend to continue working on it. Thank you!