Fix v1_chat_generate_request to allow for None content

[Amadeus-Winarto]

Motivation

Related to #5452.
Allow content field to be None, as per OpenAI types: https://github.com/openai/openai-python/blob/05810dd4088b6dbfc4194d7b0bea03eec236c83a/src/openai/types/chat/chat_completion_assistant_message_param.py#L46C67-L47C1

This fixes issues when running agentic-workflow related use of SGLang, where the assistant's output may be fed back into /v1/chat/completions as is without modification. For example, in the following flow:

LLM attempts tool call i.e. content is None and tool_calls is a valid list
LLM's response is fed back into messages in ChatCompletionRequest without modification
It will currently fail on the request generation step since a non-string typed content is automatically assumed to be a list, even though content can be None, even though the OpenAI spec allows this. This change fixes this issue.

Modifications

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

[merrymercy]

please fix the lint

[Amadeus-Winarto]

please fix the lint

Oops thanks, it's fixed