app.py

# pip install gradio
# pip install pandas
# pip install scikit-learn
# pip install matplotlib

import warnings 
warnings.filterwarnings('ignore')

from sklearn.mixture import GaussianMixture
from sklearn.metrics import silhouette_score, silhouette_samples
from sklearn.preprocessing import StandardScaler
import matplotlib.pyplot as plt
import pathlib
import pandas as pd
import gradio as gr
import numpy as np

plt.switch_backend('agg')

def gmm_cluster(input_file, first_col, second_col, n_clusters):
    data = pd.read_csv(pathlib.Path(input_file.name))

    features = [first_col, second_col]

    X = data[features]
    
    fig1 = plt.figure()

    plt.scatter(X[first_col],X[second_col])

    z = StandardScaler()

    X[features] = z.fit_transform(X)

    fig2 = plt.figure()

    plt.scatter(X[first_col],X[second_col])

    gm = GaussianMixture(n_components=int(n_clusters))

    EM = gm.fit(X) 

    cluster = gm.predict(X)

    labels = gm.predict(X[features])
    
    frame = pd.DataFrame(X)
    frame['cluster'] = labels
    
    frame.columns = [first_col, second_col, 'cluster']
    
    fig3 = plt.figure()
    score = silhouette_score(X, cluster)
    
    for k in range(0,int(n_clusters)):
        color = plt.cm.nipy_spectral(float(k) / n_clusters)
        
        data = frame[frame["cluster"]==k]
        plt.scatter(data[first_col],data[second_col],color=color)
        
    fig4 = plt.figure(figsize = (6,6))
    plt.gca().set_xlim([-0.1,1])
    plt.gca().set_ylim([0, len(X) + (n_clusters + 1) * 10])

    # gm = GaussianMixture(n_components=n_clusters)
    # labels = gm.fit_predict(X)

    y_lower = 10

    sample_silhouette_values = silhouette_samples(X, labels)

    for i in range(0,int(n_clusters)):
        ith_cluster_silhouette_values = \
        sample_silhouette_values[labels == i]
        ith_cluster_silhouette_values.sort()
        size_cluster_i = ith_cluster_silhouette_values.shape[0]
        y_upper = y_lower + size_cluster_i
        color = plt.cm.nipy_spectral(float(i) / n_clusters)
        plt.gca().fill_betweenx(np.arange(y_lower, y_upper),
                              0, ith_cluster_silhouette_values,
                              facecolor=color, edgecolor=color, alpha=0.7)
        plt.gca().text(-0.05, y_lower + 0.5 * size_cluster_i, str(i))
        y_lower = y_upper + 10

    plt.gca().axvline(x=silhouette_score(X, labels), color="red", linestyle="--", label = "Avg silhouette score")
    plt.title("The silhouette plot")
    plt.xlabel("The silhouette coefficient values")
    plt.ylabel("Cluster label")
    plt.legend()
    
    return fig1, fig3, score, fig4

def show_dataset(input_file):
    new_data = pd.read_csv(pathlib.Path(input_file.name))    
    data_set = new_data.head(9)
    
    return data_set

title = 'Gaussian Mixture Model'

with gr.Blocks(title=title) as demo:
    gr.HTML(f"{title}")
    with gr.Row():
        with gr.Column():
            inp_file = gr.File(label="Input")
            inp_col_1 = gr.Text(label='Feature 1')
            inp_col_2 = gr.Text(label='Feature 2')
            inp_num_clusters = gr.Slider(
            minimum=2,
            maximum=7,
            value=0,
            step=1,
            label='Number of clusters'
            )
            out3 = gr.Textbox(label='Silhouette score')
        output_file = gr.Dataframe(label='Dataset')
    with gr.Row():
        out1 = gr.Plot()
        out2 = gr.Plot()
        out4 = gr.Plot()

    inp_file.change(fn=show_dataset, inputs=[inp_file], outputs=[output_file])
    inp_num_clusters.change(fn=gmm_cluster, inputs=[inp_file, inp_col_1, inp_col_2, inp_num_clusters], outputs=[out1, out2, out3, out4])
demo.launch()