[feat] Add Vertex AI compatible prediction route for /generate

[KCFindstr]

Motivation

The Google Cloud Vertex AI Online Prediction has some requirements on the prediction routes, request, and response formats. This PR enables SGLang docker containers to be served directly on Google Cloud Vertex AI without modification (unless advanced features such as multi-node serving are needed).

Modifications

I added a Vertex AI compatible route for the /generate API, but with Vertex-specific request and response formats. It dynamically mounts based on the AIP_PREDICT_ROUTE environment variable

Checklist

• Format your code according to the Code Formatting with Pre-Commit.
• Add unit tests as outlined in the Running Unit Tests.
• Update documentation / docstrings / example tutorials as needed, according to Writing Documentation.

  The new API route does not provide additional value to non-Vertex users. Will provide examples on Vertex AI.

• Provide throughput / latency benchmark results and accuracy evaluation results as needed, according to Benchmark and Profiling and Accuracy Results.

  N/A

• For reviewers: If you haven't made any contributions to this PR and are only assisting with merging the main branch, please remove yourself as a co-author when merging the PR.
• Please feel free to join our Slack channel at https://slack.sglang.ai to discuss your PR.