Support embeddings endpoints

[gcalmettes]

🚀 Feature Description and Motivation

With the validation of the request body recently added, embeddings request cannot be validated under aibrix.
It would be nice to also support the /v1/embeddings endpoint in addition to the /v1/chat/completions and /v1/completions endpoints already supported.

Use Case

This would be useful in order to use AiBrix for embeddings-related workloads

Proposed Solution

I'll open a dedicated PR to add embeddings support

[Jeffwan]

This is important and feel free to open a PR. /cc @varungup90

[Jeffwan]

part of #846

[robin2008]

Will the PR also address the rerank endpoints?

[Jeffwan]

@robin2008 rerank needs a separate endpoint, we will add it soon. could you confirm which endpoint are you using now?

@varungup90 please check /rerank /v1/rerank and /v2/rerank interface in vllm/entrypoints/openai/api_server.py

[robin2008]

we are using vllm /v1/rerank endpoint provided by vllm-project/vllm#12376, thanks.

[varungup90]

This is completed.