app.py

import gradio as gr

import matplotlib.pyplot as plt
import numpy as np
from sklearn import datasets, linear_model
from sklearn.metrics import mean_squared_error, r2_score

import matplotlib
matplotlib.use('agg')

FIGSIZE = (10,10)

feature_names = ["Age", "Body-Mass Index (BMI)", "Blood Pressure", 
                "Total serum Cholesterol", "Low-Density Lipoproteins (LDL)", 
                "High-Density Lipoproteins (HDL)", "Total cholesterol / HDL", 
                "log(Serum Triglycerides Level) (possibly)","Blood Sugar Level"]

def create_dataset(feature_id=2):
    # Load the diabetes dataset
    diabetes_X, diabetes_y = datasets.load_diabetes(return_X_y=True)

    # Use only one feature
    diabetes_X = diabetes_X[:, np.newaxis, feature_id]

    # Split the data into training/testing sets
    diabetes_X_train = diabetes_X[:-20]
    diabetes_X_test = diabetes_X[-20:]

    # Split the targets into training/testing sets
    diabetes_y_train = diabetes_y[:-20]
    diabetes_y_test = diabetes_y[-20:]

    return diabetes_X_train, diabetes_X_test, diabetes_y_train, diabetes_y_test

def train_model(input_data):

    # We removed the sex variable 
    if input_data == 'age':
        feature_id = 0
    else:
        feature_id = feature_names.index(input_data) + 1

    diabetes_X_train, diabetes_X_test, diabetes_y_train, diabetes_y_test = create_dataset(feature_id)

    
    # Create linear regression object
    regr = linear_model.LinearRegression()

    # Train the model using the training sets
    regr.fit(diabetes_X_train, diabetes_y_train)

    # Make predictions using the testing set
    diabetes_y_pred = regr.predict(diabetes_X_test)

    mse = mean_squared_error(diabetes_y_test, diabetes_y_pred)
    r2 = r2_score(diabetes_y_test, diabetes_y_pred)

    # Plot outputs
    fig = plt.figure(figsize=FIGSIZE)
    
    # plt.title(input_data)
    plt.scatter(diabetes_X_test, diabetes_y_test, color="black")
    plt.plot(diabetes_X_test, diabetes_y_pred, color="blue", linewidth=3)

    plt.xlabel(input_data, fontsize=18)
    plt.ylabel("Disease progression", fontsize=18)

    plt.xticks(())
    plt.yticks(())

    return fig, regr.coef_, mse, r2

title = "Linear Regression Example 📈"
description = """The example shows how linear regression attempts to draw a straight line that will best minimize the residual sum of squares between the observed responses in the dataset. 

                The diabetes dataset contains baseline variables (features), age, sex, body mass index, average blood pressure, and six blood serum measurements that were obtained for 442 diabetes patients.
                The predictive variable is a quantitative measure of the disease progression one year after the baseline.

                When selecting a feature from the drop-down menu, a linear regression model is trained for the specific feature and the predictive variable.
                The figure shows a scatter plot of the test set as well as the linear model (line).
                The mean square error and R2 scores are calculated using the test set and they are printed, along with the regression coefficiet of the model. 
                """
with gr.Blocks() as demo:
    gr.Markdown(f"## {title}")
    gr.Markdown(description)

    with gr.Column():
    
        with gr.Row():
            plot = gr.Plot()
            with gr.Column():
                input_data = gr.Dropdown(choices=feature_names, label="Feature", value="Body-Mass Index")
                coef = gr.Textbox(label="Coefficients")
                mse = gr.Textbox(label="Mean Squared Error (MSE)")
                r2 = gr.Textbox(label="R2 score")

        input_data.change(fn=train_model, inputs=[input_data], outputs=[plot, coef, mse, r2], queue=False)

    
demo.launch(enable_queue=True)