m@30: # -*- coding: utf-8 -*-
m@30: """
m@30: Created on Fri Sep  1 19:11:52 2017
m@30: 
m@30: @author: mariapanteli
m@30: """
m@30: 
m@30: import pytest
m@30: 
m@30: import numpy as np
m@30: import pandas as pd
m@30: import pickle
m@30: import os
m@30: 
m@30: import scripts.utils as utils
m@30: 
m@30: 
m@30: def test_get_outliers():
m@30:     np.random.seed(1)
m@30:     X = np.random.randn(100, 3)
m@30:     # create outliers by shifting the entries of the last 5 samples
m@30:     X[-5:, :] = X[-5:, :] + 10
m@30:     Y = np.concatenate([np.repeat('a', 95), np.repeat('b', 5)])
m@30:     threshold, y_pred, MD = utils.get_outliers(X)
m@30:     # expect that items from country 'b' are detected as outliers
m@30:     assert np.array_equal(y_pred[-5:], np.ones(5))
m@30: 
m@30: 
m@30: def test_get_outliers():
m@30:     np.random.seed(1)
m@30:     X = np.random.randn(100, 3)
m@30:     # create outliers by shifting the entries of the last 5 samples
m@30:     X[-5:, :] = X[-5:, :] + 10
m@30:     Y = np.concatenate([np.repeat('a', 95), np.repeat('b', 5)])
m@30:     threshold, y_pred, MD = utils.get_outliers_Mahal(X)
m@30:     # expect that items from country 'b' are detected as outliers
m@30:     assert np.array_equal(y_pred[-5:], np.ones(5))
m@30: 
m@30: 
m@30: def test_pca_data():
m@30:     np.random.seed(1)
m@30:     X = np.random.randn(100, 3)
m@30:     X[-5:, :] = X[-5:, :] + 10
m@30:     X_pca, n_pc = utils.pca_data(X, min_variance=0.8)
m@30:     assert n_pc < X.shape[1]
m@30: 
m@30: 
m@30: def test_get_local_outliers_from_neighbors_dict():
m@30:     np.random.seed(1)
m@30:     X = np.random.randn(100, 3)
m@30:     n_outliers = 3
m@30:     X[-n_outliers:, :] = X[-n_outliers:, :] + 10
m@30:     Y = np.concatenate([np.repeat('a', 20), np.repeat('b', 20), np.repeat('c', 20), 
m@30:                         np.repeat('k', 20), np.repeat('l', 20)])
m@30:     w_dict = {'a': ['b', 'c'], 'b': ['a', 'c'], 'c': ['b', 'a'], 'k': ['l'], 'l':['k']}
m@30:     spatial_outliers = utils.get_local_outliers_from_neighbors_dict(X, Y, w_dict)
m@30:     # last n samples of 'l' country must be outliers
m@30:     assert np.array_equal(spatial_outliers[-1][3][-n_outliers:], np.ones(n_outliers))
m@30: 
m@30: 
m@30: def test_best_n_clusters_silhouette():
m@30:     np.random.seed(1)
m@30:     X = np.random.randn(100, 3)
m@30:     X[:30, :] = X[:30, :] + 10
m@30:     X[-30:, :] = X[-30:, :] + 20
m@30:     bestncl, _ = utils.best_n_clusters_silhouette(X, max_ncl=10)
m@30:     assert bestncl == 3
m@30: