app.py

import gradio as gr
from fastapi import FastAPI, HTTPException
from fastapi.middleware.cors import CORSMiddleware
from pydantic import BaseModel
from typing import List, Tuple
from huggingface_hub import InferenceClient
import os
from dotenv import load_dotenv

load_dotenv()

app = FastAPI()

# Add CORS middleware
app.add_middleware(
    CORSMiddleware,
    allow_origins=["*"],
    allow_credentials=True,
    allow_methods=["*"],
    allow_headers=["*"],
)

client = InferenceClient("HuggingFaceH4/zephyr-7b-beta", token=os.getenv("HF_TOKEN"))

class ChatRequest(BaseModel):
    message: str
    history: List[Tuple[str, str]]
    system_message: str
    max_tokens: int
    temperature: float
    top_p: float

def respond(
    message,
    history: list[tuple[str, str]],
    max_tokens,
    temperature,
    top_p,
    system_message: str = """You are a chatbot serving a user a text based adventure. When the user says 'start adventure', you will write a short (((70 word))) adventure story with 2 to 4 choices for the user to take at the end. Progress the story based on their choices. Number the choices as 1,2,3 and 4 etc. Don't take the choice yourself. Wait for the user to respond.""",
):
    messages = [{"role": "system", "content": system_message}]

    for val in history:
        if val[0]:
            messages.append({"role": "user", "content": val[0]})
        if val[1]:
            messages.append({"role": "assistant", "content": val[1]})

    messages.append({"role": "user", "content": message})

    response = ""

    for message in client.chat_completion(
        messages,
        max_tokens=max_tokens,
        stream=True,
        temperature=temperature,
        top_p=top_p,
    ):
        token = message.choices[0].delta.content

        response += token
        yield response

@app.post("/chat")
async def chat_endpoint(request: ChatRequest):
    try:
        response = respond(
            request.message,
            request.history,
            request.max_tokens,
            request.temperature,
            request.top_p,
            request.system_message,
        )
        return {"response": list(response)}
    except Exception as e:
        raise HTTPException(status_code=500, detail=str(e))

# Gradio interface
demo = gr.ChatInterface(
    respond,
    additional_inputs=[
        gr.Slider(minimum=1, maximum=2048, value=250, step=1, label="Max new tokens"),
        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
        gr.Slider(
            minimum=0.1,
            maximum=1.0,
            value=0.95,
            step=0.05,
            label="Top-p (nucleus sampling)",
        ),
    ],
)

# Mount the Gradio app
app = gr.mount_gradio_app(app, demo, path="/")

if __name__ == "__main__":
    import uvicorn
    uvicorn.run(app, host="0.0.0.0", port=7860)