app.py

import uvicorn
from fastapi import FastAPI, UploadFile, File
import joblib
import numpy as np
import pandas as pd
import gradio as gr
import xgboost as xgb

# Create the app object
app = FastAPI()

# Load the saved model
model = joblib.load("model.pkl")

# Define the routes
@app.get('/')
def index():
    return {'message': 'Hello, Everyone!'}

@app.get('/Welcome/{name}')
def get_name(name: str):
    return {'Welcome to my ML model': name}

@app.post('/predict_csv')
def smoker_status(file: UploadFile = File(...)):
    # Read the CSV file
    df = pd.read_csv(file.name)
    df.drop(['relaxation', 'LDL', 'waist(cm)'], axis=1, inplace=True)

    # Make predictions for each row
    predictions = []
    for _, row in df.iterrows():
        input_data = np.array([row.values])
        prediction = model.predict(input_data)
        if prediction == 0:
            ispu = "NOT SMOKER"
        else:
            ispu = "SMOKER"
        predictions.append(ispu)

    return {"predictions": predictions}

# Define a Gradio interface for your model using gr.inputs.File
iface = gr.Interface(
    fn=smoker_status,
    inputs=gr.inputs.File(),
    outputs=gr.outputs.Textbox(),
    live=True,
    title="CSV File Predictions",
    description="Upload a CSV file to make predictions on its data.",
)

if __name__ == '__main__':
    # Start Gradio alongside FastAPI without attempting to share
    iface.launch()

    # Start FastAPI
    uvicorn.run("app:app", host='0.0.0.0', port=8080)