def test_krebs_dataset():
krebs_data_block1, krebs_matches_block1 = load_krebsregister(1)
krebs_data_block10, krebs_matches_block10 = load_krebsregister(10)
assert len(krebs_data_block1) > 0
assert len(krebs_data_block10) > 0
# load not existing block
with pytest.raises(ValueError):
load_krebsregister(11)
# missing values
krebs_block10, matches = load_krebsregister(10, missing_values=0)
assert krebs_block10.isnull().sum().sum() == 0
def test_krebs_dataset_environ(tmpdir):
path = Path(str(tmpdir)).expanduser()
environ['RL_DATA'] = str(path)
krebs_data, krebs_matches = load_krebsregister()
for i in range(1, 11):
assert Path(path, "krebsregister", "block_{}.zip".format(i)).is_file()
def test_krebs_dataset(self):
krebs_data, krebs_matches = load_krebsregister()
krebs_data_block1, krebs_matches_block1 = load_krebsregister(1)
krebs_data_block10, krebs_matches_block10 = load_krebsregister(10)
# count the number of recordss
self.assertEqual(type(krebs_data), pandas.DataFrame)
self.assertEqual(type(krebs_matches), pandas.MultiIndex)
self.assertEqual(len(krebs_data), 5749132)
self.assertEqual(len(krebs_matches), 20931)
self.assertGreater(len(krebs_data_block1), 0)
self.assertGreater(len(krebs_data_block10), 0)
# load not existing block
self.assertRaises(ValueError, load_krebsregister, 11)
# missing values
krebs_data_block10, krebs_matches = load_krebsregister(10, missing_values=0)
self.assertEqual(krebs_data_block10.isnull().sum().sum(), 0)
def test_krebs_dataset_download():
# remove downloaded datasets
clear_data_home()
krebs_data, krebs_matches = load_krebsregister()
for i in range(1, 11):
assert Path(get_data_home(), "krebsregister",
"block_{}.zip".format(i)).is_file()
# count the number of recordss
assert type(krebs_data), pandas.DataFrame
assert type(krebs_matches), pandas.MultiIndex
assert len(krebs_data) == 5749132
assert len(krebs_matches) == 20931
def test_krebs_dataset(self):
krebs_data, krebs_matches = load_krebsregister()
krebs_data_block1, krebs_matches_block1 = load_krebsregister(1)
krebs_data_block10, krebs_matches_block10 = load_krebsregister(10)
# count the number of recordss
assert type(krebs_data), pandas.DataFrame
assert type(krebs_matches), pandas.MultiIndex
assert len(krebs_data) == 5749132
assert len(krebs_matches) == 20931
assert len(krebs_data_block1) > 0
assert len(krebs_data_block10) > 0
# load not existing block
with pytest.raises(ValueError):
load_krebsregister(11)
# missing values
krebs_block10, matches = load_krebsregister(10, missing_values=0)
assert krebs_block10.isnull().sum().sum() == 0
def test_krebs_shuffle():
# missing values
krebs_block10, matches = load_krebsregister(10, shuffle=False)
def test_krebs_missings():
# missing values
krebs_block10, matches = load_krebsregister(10, missing_values=0)
assert krebs_block10.isnull().sum().sum() == 0
def test_krebs_shuffle(self):
# missing values
krebs_data_block10, krebs_matches = load_krebsregister(10, shuffle=False)
def test_krebs_missings(self):
# missing values
krebs_data_block10, krebs_matches = load_krebsregister(10, missing_values=0)
self.assertEqual(krebs_data_block10.isnull().sum().sum(), 0)
import recordlinkage as rl from recordlinkage.datasets import load_krebsregister krebs_X, krebs_true_links = load_krebsregister(missing_values=0) print(krebs_true_links) # Train the classifier ecm = rl.ECMClassifier(binarize=0.8) result_ecm = ecm.fit_predict(krebs_X) len(result_ecm) print(rl.confusion_matrix(krebs_true_links, result_ecm, len(krebs_X))) # The F-score for this classification is print(rl.fscore(krebs_true_links, result_ecm)) print(ecm.log_weights)
