WIP version of dynamic structural tags based on tools for gpt-oss

[alecsolder]

Run server with

CUDA_VISIBLE_DEVICES=0 HF_HUB_OFFLINE=1 vllm serve openai/gpt-oss-20b --tool-server=localhost:8081/browser,localhost:8081/python --structured_outputs_config='{"enable_in_reasoning": true, "reasoning_parser": "openai_gptoss"}'

Then you can do stuff like this

curl -X POST http://localhost:8000/v1/responses \
-H "Content-Type: application/json" \
-d '{
"model": "openai/gpt-oss-20b",
"input": "Search for vLLM performance.",
"tools": [{"type": "web_search_preview"}]
}'

curl -X POST http://localhost:8000/v1/responses \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/gpt-oss-20b",
    "input": "Look up the weather roughly in san francisco.",
    "enable_response_messages": true,
    "tools": [{
      "type": "function",
      "name": "get_weatherrrrr",
      "description": "Get current temperature for provided coordinates in celsius.",
      "parameters": {
        "type": "object",
        "properties": {
          "latitude": {"type": "number"},
          "longitude": {"type": "number"}
        },
        "required": ["latitude", "longitude"],
        "additionalProperties": false
      },
      "strict": true
    }]
  }'

And the chat format will be guided to ensure correct headers for all tools in the request specifically