121 lines
3.9 KiB
Python
121 lines
3.9 KiB
Python
from __future__ import annotations
|
|
|
|
import json
|
|
from typing import AsyncGenerator, Optional
|
|
|
|
import httpx
|
|
from fastapi import APIRouter, Depends, Request, status
|
|
from fastapi.responses import StreamingResponse
|
|
from pydantic import BaseModel, Field
|
|
|
|
from app.auth.keycloak_dependencies import get_current_keycloak_username
|
|
from app.core.config import settings
|
|
|
|
router = APIRouter()
|
|
|
|
|
|
class CopilotChatStreamRequest(BaseModel):
|
|
message: str = Field(..., min_length=1, max_length=10000)
|
|
conversation_id: Optional[str] = Field(default=None, max_length=128)
|
|
|
|
|
|
def _sse_event(event: str, data: dict) -> str:
|
|
return f"event: {event}\ndata: {json.dumps(data, ensure_ascii=False)}\n\n"
|
|
|
|
|
|
@router.post(
|
|
"/chat/stream",
|
|
summary="Copilot 聊天流式响应",
|
|
description="向 Python Copilot sidecar 转发请求并通过 SSE 返回增量内容",
|
|
)
|
|
async def copilot_chat_stream(
|
|
payload: CopilotChatStreamRequest,
|
|
request: Request,
|
|
username: str = Depends(get_current_keycloak_username),
|
|
):
|
|
timeout = httpx.Timeout(
|
|
connect=10.0,
|
|
read=float(settings.COPILOT_STREAM_TIMEOUT_SECONDS),
|
|
write=10.0,
|
|
pool=10.0,
|
|
)
|
|
sidecar_url = settings.COPILOT_SIDECAR_URL.rstrip("/")
|
|
upstream_url = f"{sidecar_url}/chat/stream"
|
|
|
|
async def event_generator() -> AsyncGenerator[str, None]:
|
|
headers: dict[str, str] = {}
|
|
auth_header = request.headers.get("authorization")
|
|
project_id = request.headers.get("x-project-id")
|
|
if auth_header:
|
|
headers["authorization"] = auth_header
|
|
if project_id:
|
|
headers["x-project-id"] = project_id
|
|
|
|
body = {
|
|
"message": payload.message,
|
|
"conversationId": payload.conversation_id,
|
|
"userId": username,
|
|
}
|
|
|
|
try:
|
|
async with httpx.AsyncClient(timeout=timeout) as client:
|
|
async with client.stream(
|
|
"POST",
|
|
upstream_url,
|
|
json=body,
|
|
headers=headers,
|
|
) as response:
|
|
if response.status_code >= 400:
|
|
detail_text = await response.aread()
|
|
detail = detail_text.decode("utf-8", errors="replace")
|
|
yield _sse_event(
|
|
"error",
|
|
{
|
|
"message": "Copilot sidecar request failed",
|
|
"status": response.status_code,
|
|
"detail": detail,
|
|
},
|
|
)
|
|
return
|
|
|
|
async for line in response.aiter_lines():
|
|
if await request.is_disconnected():
|
|
return
|
|
yield f"{line}\n"
|
|
except httpx.ReadTimeout:
|
|
yield _sse_event(
|
|
"error",
|
|
{
|
|
"message": "Copilot stream timeout",
|
|
"status": status.HTTP_504_GATEWAY_TIMEOUT,
|
|
},
|
|
)
|
|
except httpx.ConnectError as exc:
|
|
yield _sse_event(
|
|
"error",
|
|
{
|
|
"message": "Copilot sidecar unavailable",
|
|
"status": status.HTTP_503_SERVICE_UNAVAILABLE,
|
|
"detail": str(exc),
|
|
},
|
|
)
|
|
except Exception as exc:
|
|
yield _sse_event(
|
|
"error",
|
|
{
|
|
"message": "Unexpected stream proxy error",
|
|
"status": status.HTTP_500_INTERNAL_SERVER_ERROR,
|
|
"detail": str(exc),
|
|
},
|
|
)
|
|
|
|
return StreamingResponse(
|
|
event_generator(),
|
|
media_type="text/event-stream",
|
|
headers={
|
|
"Cache-Control": "no-cache",
|
|
"Connection": "keep-alive",
|
|
"X-Accel-Buffering": "no",
|
|
},
|
|
)
|