app.py

import gradio as gr
import pickle
import pandas as pd
from sklearn.compose import ColumnTransformer
from sklearn.pipeline import Pipeline
import numpy as np
import pandas as pd
from sklearn.preprocessing import StandardScaler
from sklearn.preprocessing import OneHotEncoder

# Load the model and preprocessor
model = pickle.load(open('model.pkl', 'rb'))
preprocessor = pickle.load(open('preprocess_pipeline.pkl', 'rb'))



def call_gradio(Sex, Equipment, Age, BodyweightKg, BestSquatKg, Bestbenchkg):
    # Convert inputs to appropriate data types
    Age = float(Age)
    BodyweightKg = float(BodyweightKg)
    BestSquatKg = float(BestSquatKg)
    Bestbenchkg = float(Bestbenchkg)
    
    # Create a DataFrame with the input data
    df_x = pd.DataFrame({
        'Sex': [Sex],
        'Equipment': [Equipment],
        'Age': [Age],
        'BodyweightKg': [BodyweightKg],
        'BestSquatKg': [BestSquatKg],
        'Bestbench(kg)': [Bestbenchkg],
    })
    
    # Define the categorical and numerical feature lists
    categorical_features = ['Sex', 'Equipment']
    numerical_features = df_x.drop(categorical_features, axis=1).columns.tolist()
    
    # Ensure that the preprocessor is fitted
    if not hasattr(preprocessor, 'transformers_'):
        raise RuntimeError("The preprocessor is not fitted yet. Fit the preprocessor before calling this function.")
    
    
    # Transform the data using the preprocessor
    X_processed = preprocessor.transform(df_x)
    
    # Access the OneHotEncoder directly to get feature names
    ohe = preprocessor.named_transformers_['cat'].named_steps['onehot']
    cat_feature_names = ohe.get_feature_names_out(categorical_features)
    
    # Combine numerical features and one-hot encoded feature names
    all_feature_names = numerical_features + list(cat_feature_names)
    
    # Create a DataFrame with processed features
    X_processed_df = pd.DataFrame(X_processed, columns=all_feature_names)
    
    # Predict using the model
    y_pred = model.predict(X_processed_df)
    
    max_kg = int(y_pred[0])
    return max_kg

# Define Gradio inputs and outputs
sex_dropdown = gr.Dropdown(choices=['M', 'F'], label="Sex", info="Select Male or Female")
equipment_dropdown = gr.Dropdown(choices=['Raw', 'Wraps', 'Single-ply', 'Multi-ply'], label="Equipment", info="Select the equipment used for the competition.")
age_textbox = gr.Textbox(lines=1, label="Age", info="Enter your Age")
bodyweight_kg_textbox = gr.Textbox(lines=1, label="BodyweightKg", info="Enter your Bodyweight in Kg")
best_squat_kg_textbox = gr.Textbox(lines=1, label="BestSquatKg", info="Enter your Best Squat in Kg")
best_bench_kg_textbox = gr.Textbox(lines=1, label="BestbenchKg", info="Enter your Best Bench in Kg")

# Custom description with image and footer
description = """
<div style='text-align: center;'>
    <h1 style='font-size: 50px;'>PowerLift Muscle Map</h1> <!-- Title with increased font size -->
    <p>Use this model to estimate your best Deadlift (kg) based on your selected features. Input your details and see the predicted weight (kg) you could lift.</p>
    <p><strong>Output:</strong> Estimated Best Deadlift (kg)</p>
    <br>
    <p style='font-size: 10px; color: #555;'>❤️ PDS</p>
</div>
"""

# Create and launch Gradio interface
iface = gr.Interface(
    fn=call_gradio,
    inputs=[sex_dropdown, equipment_dropdown, age_textbox, bodyweight_kg_textbox, best_squat_kg_textbox, best_bench_kg_textbox],
    outputs="number",
    description=description,
)

iface.launch()