Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -1,5 +1,8 @@
|
|
| 1 |
import os
|
|
|
|
| 2 |
import gradio as gr
|
|
|
|
|
|
|
| 3 |
from openai import OpenAI
|
| 4 |
|
| 5 |
# ---------------------------
|
|
@@ -27,7 +30,23 @@ SYSTEM_PROMPT = (
|
|
| 27 |
MODEL_ID = "Qwen/Qwen3-Next-80B-A3B-Instruct:novita"
|
| 28 |
|
| 29 |
# ---------------------------
|
| 30 |
-
# 3.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 31 |
# ---------------------------
|
| 32 |
|
| 33 |
def chat_fn(message, history):
|
|
@@ -52,7 +71,7 @@ def chat_fn(message, history):
|
|
| 52 |
messages=messages,
|
| 53 |
)
|
| 54 |
|
| 55 |
-
reply = completion.choices[0].message
|
| 56 |
|
| 57 |
except Exception as e:
|
| 58 |
reply = f"Error: {str(e)}"
|
|
@@ -60,7 +79,37 @@ def chat_fn(message, history):
|
|
| 60 |
return reply
|
| 61 |
|
| 62 |
# ---------------------------
|
| 63 |
-
#
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 64 |
# ---------------------------
|
| 65 |
|
| 66 |
with gr.Blocks(title="YAH Assistant") as demo:
|
|
@@ -68,6 +117,17 @@ with gr.Blocks(title="YAH Assistant") as demo:
|
|
| 68 |
"""
|
| 69 |
## YAH Assistant
|
| 70 |
Large-model chat interface powered by Hugging Face Router.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 71 |
"""
|
| 72 |
)
|
| 73 |
|
|
@@ -81,5 +141,13 @@ with gr.Blocks(title="YAH Assistant") as demo:
|
|
| 81 |
|
| 82 |
msg.submit(respond, [msg, chatbot], [msg, chatbot])
|
| 83 |
|
| 84 |
-
#
|
| 85 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
import os
|
| 2 |
+
import json
|
| 3 |
import gradio as gr
|
| 4 |
+
from fastapi import FastAPI, HTTPException
|
| 5 |
+
from pydantic import BaseModel
|
| 6 |
from openai import OpenAI
|
| 7 |
|
| 8 |
# ---------------------------
|
|
|
|
| 30 |
MODEL_ID = "Qwen/Qwen3-Next-80B-A3B-Instruct:novita"
|
| 31 |
|
| 32 |
# ---------------------------
|
| 33 |
+
# 3. Pydantic models for API
|
| 34 |
+
# ---------------------------
|
| 35 |
+
|
| 36 |
+
class ChatMessage(BaseModel):
|
| 37 |
+
role: str
|
| 38 |
+
content: str
|
| 39 |
+
|
| 40 |
+
class ChatRequest(BaseModel):
|
| 41 |
+
message: str
|
| 42 |
+
history: list[tuple[str, str]] = []
|
| 43 |
+
|
| 44 |
+
class ChatResponse(BaseModel):
|
| 45 |
+
response: str
|
| 46 |
+
error: str = None
|
| 47 |
+
|
| 48 |
+
# ---------------------------
|
| 49 |
+
# 4. Core chat function
|
| 50 |
# ---------------------------
|
| 51 |
|
| 52 |
def chat_fn(message, history):
|
|
|
|
| 71 |
messages=messages,
|
| 72 |
)
|
| 73 |
|
| 74 |
+
reply = completion.choices[0].message.content
|
| 75 |
|
| 76 |
except Exception as e:
|
| 77 |
reply = f"Error: {str(e)}"
|
|
|
|
| 79 |
return reply
|
| 80 |
|
| 81 |
# ---------------------------
|
| 82 |
+
# 5. API Endpoint Function
|
| 83 |
+
# ---------------------------
|
| 84 |
+
|
| 85 |
+
def chat_api_endpoint(chat_request: ChatRequest) -> ChatResponse:
|
| 86 |
+
"""
|
| 87 |
+
API endpoint for chat interactions
|
| 88 |
+
"""
|
| 89 |
+
try:
|
| 90 |
+
reply = chat_fn(chat_request.message, chat_request.history)
|
| 91 |
+
return ChatResponse(response=reply)
|
| 92 |
+
except Exception as e:
|
| 93 |
+
return ChatResponse(response="", error=str(e))
|
| 94 |
+
|
| 95 |
+
# ---------------------------
|
| 96 |
+
# 6. FastAPI App Setup
|
| 97 |
+
# ---------------------------
|
| 98 |
+
|
| 99 |
+
app = FastAPI(title="YAH Assistant API")
|
| 100 |
+
|
| 101 |
+
# Add the API endpoint
|
| 102 |
+
@app.post("/chat/")
|
| 103 |
+
async def chat(request: ChatRequest):
|
| 104 |
+
return chat_api_endpoint(request)
|
| 105 |
+
|
| 106 |
+
# Health check endpoint
|
| 107 |
+
@app.get("/health")
|
| 108 |
+
async def health_check():
|
| 109 |
+
return {"status": "healthy", "model": MODEL_ID}
|
| 110 |
+
|
| 111 |
+
# ---------------------------
|
| 112 |
+
# 7. Gradio UI (keeps the original interface)
|
| 113 |
# ---------------------------
|
| 114 |
|
| 115 |
with gr.Blocks(title="YAH Assistant") as demo:
|
|
|
|
| 117 |
"""
|
| 118 |
## YAH Assistant
|
| 119 |
Large-model chat interface powered by Hugging Face Router.
|
| 120 |
+
|
| 121 |
+
### API Usage
|
| 122 |
+
You can also interact with this assistant via API:
|
| 123 |
+
- **Endpoint:** `POST /chat/`
|
| 124 |
+
- **Body:**
|
| 125 |
+
```json
|
| 126 |
+
{
|
| 127 |
+
"message": "Your message here",
|
| 128 |
+
"history": [["Hello", "Hi there!"], ["How are you?", "I'm good!"]]
|
| 129 |
+
}
|
| 130 |
+
```
|
| 131 |
"""
|
| 132 |
)
|
| 133 |
|
|
|
|
| 141 |
|
| 142 |
msg.submit(respond, [msg, chatbot], [msg, chatbot])
|
| 143 |
|
| 144 |
+
# ---------------------------
|
| 145 |
+
# 8. Mount Gradio app to FastAPI
|
| 146 |
+
# ---------------------------
|
| 147 |
+
|
| 148 |
+
app = gr.mount_gradio_app(app, demo, path="/")
|
| 149 |
+
|
| 150 |
+
# Launch instructions
|
| 151 |
+
if __name__ == "__main__":
|
| 152 |
+
import uvicorn
|
| 153 |
+
uvicorn.run(app, host="0.0.0.0", port=7860)
|