import numpy as np
import matplotlib.pyplot as plt

from sklearn.cluster import AgglomerativeClustering
from sklearn.neighbors import kneighbors_graph

import gradio as gr


np.random.seed(42)

def agglomorative_cluster(n_samples: int, n_neighbours: int, n_clusters: int, linkage: str, connectivity: bool) -> "plt.Figure":

    t = 1.5 * np.pi * (1 + 3 * np.random.rand(1, n_samples))
    x = t * np.cos(t)
    y = t * np.sin(t)

    X = np.concatenate((x, y))
    X += 0.7 * np.random.randn(2, n_samples)
    X = X.T

    knn_graph = kneighbors_graph(X, n_neighbors=n_neighbours, include_self=False)
    connectivity = knn_graph if not connectivity else None

    fig, ax = plt.subplots(1, 1, figsize=(24, 15))
    model = AgglomerativeClustering(linkage=linkage, connectivity=connectivity, n_clusters=int(n_clusters))
    model.fit(X)
    ax.scatter(X[:, 0], X[:, 1], c=model.labels_, cmap=plt.cm.nipy_spectral)
    ax.axis("equal")
    ax.axis("off")

    return fig





demo = gr.Interface(
    fn = agglomorative_cluster,
    inputs = [gr.Slider(0, 100_000, label="n_samples", info="the number of samples in the data.", step=1), 
              gr.Slider(0, 30, label="n_neighbours", info="the number of neighbours in the data", step=1), 
              gr.Dropdown([3, 30], label="n_clusters", info="the number of clusters in the data"), 
              gr.Dropdown(['average', 'complete', 'ward', 'single'], label="linkage", info="the different types of aggolomorative clustering techniques"),
              gr.Checkbox(True, label="connectivity", info="whether to impose a connectivity into the graph")],
    
    outputs = [gr.Plot(label="Plot")]
)

demo.launch()