test_pca.py

import pandas as pd

from sequana.viz.pca import PCA

from . import test_dir


def test_pca():

    filename = f"{test_dir}/data/misc/test_pca.csv"
    df = pd.read_csv(filename)
    df = df.set_index("Id")

    p = PCA(df)
    p = PCA(df, colors={"A1": "r", "A2": "r", "A3": "r", "B1": "b", "B2": "b", "B3": "b"})

    # default
    p.plot()

    # test some optins
    p.plot(transform="log", switch_y=True, switch_x=True, max_features=500)
    p.plot(transform="anscombe", switch_y=True, switch_x=True, max_features=500)

    # set max_features to large values to test option
    p.plot_pca_vs_max_features(step=10000)
    p.plot_pca_vs_max_features(n_components=3)

    # 3 components
    p.plot(n_components=3, switch_z=True)