[Feature]: Native OpenAI Chat Completions Proxy Endpoint (/v1/chat/completions) #208

Problem to solve

Add a native OpenAI-compatible Chat Completions endpoint (POST /v1/chat/completions) to translate standard OpenAI request payloads into the internal Responses API format (/v1/responses) and format all responses back to standard OpenAI Chat Completion objects (both streaming and non-streaming).

Proposed solution

Endpoint Specification
Endpoint: POST /v1/chat/completions
Content-Type: application/json
Headers: Authorization: Bearer <API_KEY>
Request Payload (ChatCompletionRequest)
Input Parameters:
`
{
"model": "string",
"messages": [
{
"role": "system | user | assistant | tool",
"content": "string | array",
"name": "string (optional)",
"tool_calls": "array (optional)"
}
],
"stream": false,
"temperature": 0.7,
"top_p": 1.0,
"max_tokens": 1024,
"stop": ["string"],
"frequency_penalty": 0.0,
"presence_penalty": 0.0,
"tools": ["array (optional)"],
"tool_choice": "string | object (optional)",
"response_format": "object (optional)"
}

`

Response Specification

1. Non-Streaming (stream: false)
   Response Status: 200 OK
   Content-Type: application/json
   Return Structure: Standard OpenAI ChatCompletionResponse object.

{ "id": "chatcmpl-1721641000", "object": "chat.completion", "created": 1721641000, "model": "deepseek-r1", "choices": [ { "index": 0, "message": { "role": "assistant", "content": "Response text here...", "reasoning_content": "Reasoning/thinking process (if model supports reasoning)", "tool_calls": [] }, "finish_reason": "stop" } ], "usage": { "prompt_tokens": 50, "completion_tokens": 150, "total_tokens": 200 } }

2. Streaming (stream: true)
   Response Status: 200 OK
   Content-Type: text/event-stream
   Headers:
   Cache-Control: no-cache
   Connection: keep-alive
   X-Accel-Buffering: no
   Return Structure: Server-Sent Events (SSE) yielding OpenAI delta chunks ending with data: [DONE].

`data: {"id":"chatcmpl-1721641000","object":"chat.completion.chunk","created":1721641000,"model":"deepseek-r1","choices":[{"index":0,"delta":{"role":"assistant","content":""},"finish_reason":null}]}

data: {"id":"chatcmpl-1721641000","object":"chat.completion.chunk","created":1721641000,"model":"deepseek-r1","choices":[{"index":0,"delta":{"content":"Hello"},"finish_reason":null}]}

data: {"id":"chatcmpl-1721641000","object":"chat.completion.chunk","created":1721641000,"model":"deepseek-r1","choices":[{"index":0,"delta":{},"finish_reason":"stop"}]}

data: [DONE]
`

Core Internal Functions Required
chat_completions(request: Request, body: ChatCompletionRequest, authorization: str | None)
Main router function directing traffic based on body.stream.
execute_non_streaming(body: ChatCompletionRequest) -> dict
Maps incoming messages/tools to upstream /v1/responses format, fires single async HTTP POST, decodes output/reasoning, and formats OpenAI response.
execute_streaming(body: ChatCompletionRequest) -> AsyncIterator[str]
Opens stream connection to upstream SSE, parses events (response.created, response.output_item.added, response.content_part.added, response.text.delta, response.done), formats and yields OpenAI-compliant SSE chunks (data: {...}\n\n).

Alternatives considered

No response

Additional context

No response