app.py

import pandas as pd
import gradio as gr
from sklearn.ensemble import GradientBoostingRegressor
import pickle

# Load the saved model
gb_model = pickle.load(open('gb_model.pkl', 'rb'))

# Define the prediction function
def predict_leachate_volume(EC_rock, Ph_rock, Corg_rock, Ca_rock, K_rock, Mg_rock, Na_rock, SAR_rock, SiO2_rock):
    input_data = [[EC_rock, Ph_rock, Corg_rock, Ca_rock, K_rock, Mg_rock, Na_rock, SAR_rock, SiO2_rock]]
    prediction = gb_model.predict(input_data)
    return prediction[0]

# Create Gradio inputs
inputs = [
    gr.Number(label="EC_rock", value=0),
    gr.Number(label="Ph_rock", value=10),
    gr.Number(label="Corg_rock (%)", value=10),
    gr.Number(label="Ca_rock", value=60),
    gr.Number(label="K_rock", value=40),
    gr.Number(label="Mg_rock", value=6999),
    gr.Number(label="Na_rock", value=2999),
    gr.Number(label="SAR_rock", value=1999),
    gr.Number(label="SiO2_rock", value=56)
]

# Create Gradio output
outputs = gr.Textbox(label="Predicted Leachate Volume")

# Set up the Gradio interface
interface = gr.Interface(fn=predict_leachate_volume, inputs=inputs, outputs=outputs, live=True, 
                         title="Leachate Volume Prediction", description="Enter the rock characteristics to predict the leachate volume.")

# Launch the interface
interface.launch(share=True)