mmselfsup/tests/test_utils/test_clustering.py

# Copyright (c) OpenMMLab. All rights reserved.
from unittest.mock import patch

import numpy as np
import pytest

from mmselfsup.utils.clustering import PIC, Kmeans


@pytest.fixture
def mock_faiss_in_clutering():
    with patch('mmselfsup.utils.clustering.faiss') as faiss:
        yield faiss


@pytest.fixture
def mock_faiss(mock_faiss_in_clutering):
    mock_PCAmatrix = mock_faiss_in_clutering.PCAMatrix.return_value
    mock_GpuIndexFlatL2 = mock_faiss_in_clutering.GpuIndexFlatL2.return_value

    mock_PCAmatrix.apply_py.return_value = np.random.rand(10, 8)
    mock_GpuIndexFlatL2.search.return_value = (
        np.random.rand(1000, 6),
        np.random.rand(1000, 6),
    )


@pytest.mark.parametrize('verbose', [True, False])
def test_kmeans(mock_faiss, verbose):
    fake_input = np.random.rand(10, 8).astype(np.float32)
    pca_dim = 2

    kmeans = Kmeans(2, pca_dim)
    loss = kmeans.cluster(fake_input, verbose=verbose)
    assert loss is not None

    with pytest.raises(AssertionError):
        loss = kmeans.cluster(np.random.rand(10, 8), verbose=verbose)


def test_pic(mock_faiss):
    fake_input = np.random.rand(1000, 16).astype(np.float32)
    pic = PIC(pca_dim=8)
    res = pic.cluster(fake_input)
    assert res == 0