app.py

import gradio as gr
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
from sklearn.ensemble import RandomForestClassifier
from sklearn.impute import SimpleImputer
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score
import joblib
from datetime import datetime

# Load and prepare data
df = pd.read_csv("water_potability (1).csv")
imputer = SimpleImputer(strategy='mean')
df_imputed = pd.DataFrame(imputer.fit_transform(df), columns=df.columns)

X = df_imputed.drop('Potability', axis=1)
y = df_imputed['Potability']
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

model = RandomForestClassifier(random_state=42)
model.fit(X_train, y_train)
y_pred = model.predict(X_test)

# Metrics
accuracy = accuracy_score(y_test, y_pred)
precision = precision_score(y_test, y_pred)
recall = recall_score(y_test, y_pred)
f1 = f1_score(y_test, y_pred)

feature_importance = pd.Series(model.feature_importances_, index=X.columns).sort_values(ascending=False)

def predict(ph, hardness, solids, chloramines, sulfate,
            conductivity, organic_carbon, trihalomethanes, turbidity):
    
    input_data = pd.DataFrame([[
        ph, hardness, solids, chloramines, sulfate,
        conductivity, organic_carbon, trihalomethanes, turbidity
    ]], columns=X.columns)

    prediction = model.predict(input_data)[0]
    label = "✅ Safe to Drink" if prediction == 1 else "❌ Not Safe to Drink"
    proba = model.predict_proba(input_data)[0][prediction]

    # --- Visuals ---
    # Feature Importance
    fig1, ax1 = plt.subplots(figsize=(6, 5))
    feature_importance.plot(kind='barh', ax=ax1)
    ax1.set_title("Feature Importance")
    ax1.set_xlabel("Importance")
    ax1.invert_yaxis()

    # Class Distribution
    fig2, ax2 = plt.subplots(figsize=(4, 4))
    sns.countplot(x='Potability', data=df_imputed, ax=ax2)
    ax2.set_title("Potability Class Distribution")

    # Heatmap
    fig3, ax3 = plt.subplots(figsize=(6, 5))
    sns.heatmap(df_imputed.corr(), cmap='coolwarm', annot=False, ax=ax3)
    ax3.set_title("Feature Correlation Heatmap")

    # Metrics Text
    metrics_info = (
        f"📊 Model Performance on Test Set:\n\n"
        f"- Accuracy : {accuracy:.2f}\n"
        f"- Precision: {precision:.2f}\n"
        f"- Recall   : {recall:.2f}\n"
        f"- F1 Score : {f1:.2f}\n\n"
        f"Prediction Confidence: {proba:.2f}"
    )

    return f"{label}\n\n{metrics_info}", fig1, fig2, fig3

# Gradio Interface
with gr.Blocks() as demo:
    gr.Markdown("# 💧 Trustworthy Water Quality Predictor")
    gr.Markdown("Uses Random Forest with data imputation, performance metrics, and feature insights.")

    with gr.Row():
        with gr.Column():
            ph = gr.Slider(0, 14, step=0.1, label="pH")
            hardness = gr.Slider(50, 300, step=1, label="Hardness")
            solids = gr.Slider(3000, 50000, step=100, label="Solids")
            chloramines = gr.Slider(0, 15, step=0.1, label="Chloramines")
            sulfate = gr.Slider(100, 500, step=1, label="Sulfate")
            conductivity = gr.Slider(100, 800, step=1, label="Conductivity")
            organic_carbon = gr.Slider(2, 30, step=0.1, label="Organic Carbon")
            trihalomethanes = gr.Slider(0, 120, step=1, label="Trihalomethanes")
            turbidity = gr.Slider(0, 7, step=0.1, label="Turbidity")
            submit = gr.Button("Predict")

        with gr.Column():
            result = gr.Textbox(label="Prediction + Metrics")
            fig1 = gr.Plot(label="Feature Importance")
            fig2 = gr.Plot(label="Potability Class Distribution")
            fig3 = gr.Plot(label="Correlation Heatmap")

    submit.click(
        fn=predict,
        inputs=[ph, hardness, solids, chloramines, sulfate,
                conductivity, organic_carbon, trihalomethanes, turbidity],
        outputs=[result, fig1, fig2, fig3]
    )

demo.launch()