modify: 流式输出

example/models/__init__.py

@@ -15,6 +15,7 @@ from .pydantic import ExampleResponse
 from .pydantic import HeartbeatMqttRequest
 from .pydantic import MultiplyGetRequest
 from .pydantic import MultiplyGetResponse
+from .pydantic import StreamChatPostRequest
 from .pydantic import SubtractGetRequest
 from .pydantic import SubtractGetResponse
 
@@ -31,6 +32,7 @@ __all__ = [
     "HeartbeatMqttRequest",
     "MultiplyGetRequest",
     "MultiplyGetResponse",
+    "StreamChatPostRequest",
     "SubtractGetRequest",
     "SubtractGetResponse",
 ]

example/models/pydantic.py

@@ -5,7 +5,7 @@
 
 from __future__ import annotations
 
-from typing import Any, Optional
+from typing import Optional
 from pydantic import BaseModel, Field
 
 class AddGetRequest(BaseModel):
@@ -67,6 +67,11 @@ class MultiplyGetResponse(BaseModel):
     operation: str = Field(default='multiply')
     timestamp: Optional[float] = Field(default=0.0)
 
+class StreamChatPostRequest(BaseModel):
+    """模拟 AI 流式对话（SSE，逐 token 输出） — 请求体"""
+    prompt: str = Field(...)
+    tokens: int = Field(...)
+
 class SubtractGetRequest(BaseModel):
     """Subtract two numbers — 请求体"""
     a: float = Field(...)
@@ -80,4 +85,4 @@ class SubtractGetResponse(BaseModel):
 
 
 # 所有响应模型的联合类型
-ExampleResponse = AddPostResponse | AddGetResponse | DivideGetResponse | SubtractGetResponse | CountGetResponse | MultiplyGetResponse
+ExampleResponse = MultiplyGetResponse | AddPostResponse | CountGetResponse | SubtractGetResponse | DivideGetResponse | AddGetResponse

example/router/__init__.py

@@ -5,6 +5,7 @@ from .subtract import subtract
 from .multiply import multiply
 from .divide import divide
 from .count import count_get
+from .stream_chat import stream_chat
 
 
 __all__ = [
@@ -15,5 +16,6 @@ __all__ = [
     "subtract",
     "multiply",
     "divide",
-    "count_get"
+    "count_get",
+    "stream_chat"
 ]

example/router/stream_chat.py

@@ -0,0 +1,34 @@
+"""
+流式端点示例 — @rpc_method 返回异步生成器，框架自动逐块多回复推送
+
+网关在 TOML 中声明 stream = true 时，该端点以 SSE（text/event-stream）输出:
+  POST /api/v1/calc/stream_chat  {"prompt": "你好", "tokens": 5}
+  响应: data: {"token": "..."} ... data: {"done": true}
+
+约定:
+  - 插件侧写法：函数返回生成器（yield），框架逐块 query.reply() 推送
+  - {"done": true} 是业务结束标记（插件自定义），不是协议哨兵——
+    流式协议以生成器自然结束为流终止信号
+  - 中途异常时框架自动发送 {"error": ...} 收尾块
+"""
+import asyncio
+import logging
+
+from chongming_worker import rpc_method
+
+logger = logging.getLogger("plugins.example.stream_chat")
+
+
+@rpc_method(
+    subject="example.stream_chat",
+    method="POST",
+    summary="模拟 AI 流式对话（SSE，逐 token 输出）",
+    stream=True,
+)
+async def stream_chat(prompt: str = "你好", tokens: int = 10):
+    """逐 token 模拟生成：把 return 改成 yield，框架即自动流式推送"""
+    logger.info("收到流式请求: prompt=%s, tokens=%d", prompt, tokens)
+    for i in range(tokens):
+        await asyncio.sleep(0.05)
+        yield {"token": f"{prompt}·{i + 1}", "index": i}
+    yield {"done": True}