import numpy as np
import matplotlib.pyplot as plt

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

import gradio as gr

plt.switch_backend("agg")

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



with gr.Blocks() as demo:
    gr.Markdown(""" 
    # Agglomorative Clustering with and without Structure

    This space is an implementation of the scikit-learn document [Agglomorative clustering with and without structure](https://scikit-learn.org/stable/auto_examples/cluster/plot_agglomerative_clustering.html#sphx-glr-auto-examples-cluster-plot-agglomerative-clustering-py)
    This space shows the effects of imposing **connectivity graph** to capture local structure in the data. 
    You can uncheck the checkbox `connectivity` to see the effect on data clustering without **connectivity graph**. There are other parameters in this space 
    which you can play with such as `n_samples` (the number of data samples), `n_neighbours` (the number of neighbours), `n_clusters` (the number of clusters) and
    what type of linkage to use for Agglomorative clustering `linkage`. 

    Have fun playing with the tool 🤗
    """)

    n_samples = gr.Slider(0, 20_000, label="n_samples", info="the number of samples in the data.", step=1)
    n_neighbours = gr.Slider(0, 30, label="n_neighbours", info="the number of neighbours in the data", step=1)
    n_clusters = gr.Slider(3, 30, label="n_clusters", info="the number of clusters in the data", step=2)
    linkage = gr.Dropdown(['average', 'complete', 'ward', 'single'], label="linkage", info="the different types of aggolomorative clustering techniques")
    connectivity = gr.Checkbox(True, label="connectivity", info="whether to impose a connectivity into the graph")
    output = gr.Plot(label="Plot") 

    plot_btn = gr.Button("Plot")
    plot_btn.click(fn=agglomorative_cluster, inputs=[n_samples, n_neighbours, n_clusters, linkage, connectivity],
                   outputs=output, api_name="plotcluster")

demo.launch()