Spaces:
Sleeping
Sleeping
:gem: [Feature] New ChatAPIApp: Enable fastapi for openai format api call
Browse files- apis/__init__.py +0 -0
- apis/chat_api.py +82 -0
apis/__init__.py
ADDED
File without changes
|
apis/chat_api.py
ADDED
@@ -0,0 +1,82 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import uvicorn
|
2 |
+
from fastapi import FastAPI
|
3 |
+
from pydantic import BaseModel, Field
|
4 |
+
from sse_starlette.sse import EventSourceResponse
|
5 |
+
from utils.logger import logger
|
6 |
+
from networks.message_streamer import MessageStreamer
|
7 |
+
from messagers.message_composer import MessageComposer
|
8 |
+
|
9 |
+
|
10 |
+
class ChatAPIApp:
|
11 |
+
def __init__(self):
|
12 |
+
self.app = FastAPI(
|
13 |
+
docs_url="/",
|
14 |
+
title="HuggingFace LLM API",
|
15 |
+
swagger_ui_parameters={"defaultModelsExpandDepth": -1},
|
16 |
+
version="1.0",
|
17 |
+
)
|
18 |
+
self.setup_routes()
|
19 |
+
|
20 |
+
def get_available_models(self):
|
21 |
+
self.available_models = [
|
22 |
+
{
|
23 |
+
"id": "mixtral-8x7b",
|
24 |
+
"description": "Mixtral-8x7B: https://huggingface.co/mistralai/Mixtral-8x7B-Instruct-v0.1",
|
25 |
+
},
|
26 |
+
]
|
27 |
+
return self.available_models
|
28 |
+
|
29 |
+
class ChatCompletionsPostItem(BaseModel):
|
30 |
+
model: str = Field(
|
31 |
+
default="mixtral-8x7b",
|
32 |
+
description="(str) `mixtral-8x7b`",
|
33 |
+
)
|
34 |
+
messages: list = Field(
|
35 |
+
default=[{"role": "user", "content": "Hello, who are you?"}],
|
36 |
+
description="(list) Messages",
|
37 |
+
)
|
38 |
+
temperature: float = Field(
|
39 |
+
default=0.01,
|
40 |
+
description="(float) Temperature",
|
41 |
+
)
|
42 |
+
max_tokens: int = Field(
|
43 |
+
default=32000,
|
44 |
+
description="(int) Max tokens",
|
45 |
+
)
|
46 |
+
stream: bool = Field(
|
47 |
+
default=True,
|
48 |
+
description="(bool) Stream",
|
49 |
+
)
|
50 |
+
|
51 |
+
def chat_completions(self, item: ChatCompletionsPostItem):
|
52 |
+
streamer = MessageStreamer(model=item.model)
|
53 |
+
composer = MessageComposer(model=item.model)
|
54 |
+
composer.merge(messages=item.messages)
|
55 |
+
return EventSourceResponse(
|
56 |
+
streamer.chat(
|
57 |
+
prompt=composer.merged_str,
|
58 |
+
temperature=item.temperature,
|
59 |
+
max_new_tokens=item.max_tokens,
|
60 |
+
stream=item.stream,
|
61 |
+
yield_output=True,
|
62 |
+
),
|
63 |
+
media_type="text/event-stream",
|
64 |
+
)
|
65 |
+
|
66 |
+
def setup_routes(self):
|
67 |
+
for prefix in ["", "/v1"]:
|
68 |
+
self.app.get(
|
69 |
+
prefix + "/models",
|
70 |
+
summary="Get available models",
|
71 |
+
)(self.get_available_models)
|
72 |
+
|
73 |
+
self.app.post(
|
74 |
+
prefix + "/chat/completions",
|
75 |
+
summary="Chat completions in conversation session",
|
76 |
+
)(self.chat_completions)
|
77 |
+
|
78 |
+
|
79 |
+
app = ChatAPIApp().app
|
80 |
+
|
81 |
+
if __name__ == "__main__":
|
82 |
+
uvicorn.run("__main__:app", host="0.0.0.0", port=23333, reload=True)
|