def test_refine_cluster(self):
me = Analyze()
# only close and distal
me.self_low = False
me.silhouette = 0.5
me.df = pd.DataFrame(np.array(
[self.dist_gamma, self.dist_lognorm[:800]]).T,
columns=['close', 'distal'])
me.df['hgt'] = (me.df['close'] < 2) & (me.df['distal'] > 2)
me.refine_cluster(me.calc_cluster_props())
self.assertEqual(me.df[me.df['hgt']].shape[0], 11)
# all three groups
me.self_low = True
me.df = pd.DataFrame(np.array(
[self.dist_norm1[:800], self.dist_gamma,
self.dist_lognorm[:800]]).T,
columns=['self', 'close', 'distal'])
me.df['hgt'] = (me.df['close'] < 2) & (me.df['distal'] > 2)
me.refine_cluster(me.calc_cluster_props())
self.assertEqual(me.df[me.df['hgt']].shape[0], 4)
def test_calc_cluster_props(self):
me = Analyze()
me.self_low = False
me.df = pd.DataFrame(np.array(
[self.dist_gamma, self.dist_lognorm[:800]]).T,
columns=['close', 'distal'])
me.df['hgt'] = (me.df['close'] < 2) & (me.df['distal'] > 2)
obs = me.calc_cluster_props()
self.assertAlmostEqual(obs[0], 1.094658052928843)
self.assertAlmostEqual(obs[1], 4.30076698399293)
obs = me.df['silh'].describe()
self.assertAlmostEqual(obs['mean'], 0.312495082044277)
self.assertAlmostEqual(obs['std'], 0.21945541659155993)
self.assertEqual(me.df.query('hgt & silh < 0.5').shape[0], 35)