def test_cluster_ncluster():
'''
We test if 3 sets of data are clustered into 3 sets.
n_clusters is specified.
'''
CLASSIFIERS = ('cluster.ward', 'cluster.spectral')
G1 = 'First set'
G2 = 'Second group'
G3 = 'Third cluster'
data = enumerate((G1, G1, G1, G2, G2, G2, G3, G3, G3))
data = list(data)
for classifier in CLASSIFIERS:
baf = Chain().load('data.simple',data) \
.process('vectorize.sklearn') \
.process(classifier, n_clusters=3)
result = baf.get_chain('result')
assert result[0] == result[1] == result[2]
assert result[3] == result[4] == result[5]
assert result[6] == result[7] == result[8]
assert len(set(result)) == 3
def test_core_chain():
'''
Create a processor and test if
.get on processor will propagate the call
'''
data = (
(0, '0'),
(1, '1'),
)
test_subject = Chain()
with pytest.raises(KeyError) as excinfo:
assert test_subject.get('test') == 'ok'
test_subject.data['test'] = 'ok'
assert test_subject.get('test') == 'ok'
test_subject.load('data.simple', data)
# get data from the main class through the chain
assert test_subject.chain[-1].get('base_data').data == data
assert test_subject.get_chain('base_data').data == data
# get data from the chain
test_subject.chain[-1]._data['last_data'] = 'last_data'
assert test_subject.get_chain('last_data') == 'last_data'
def test_data_simple():
data = ((0, 'data test 1'), (1, 'Data test 2'), (2, 'other data test 3.'))
test_subject = Chain().load('data.simple', data)
data = test_subject.data['base_data']
assert test_subject.data['base_data'] == test_subject.data['main_data']
assert data.get_data() == [
'data test 1', 'Data test 2', 'other data test 3.'
]
assert data.get_uids() == [0, 1, 2]
def test_core_data():
'''
Load dummy data and verify it.
'''
data = (
(0, '0'),
(1, '1'),
)
test_subject = Chain()
test_subject.load('data.simple', data)
assert test_subject.get('base_data').data == data
def test_core_data_chain():
'''
When overriding data the last data should be used.
'''
data = (
(0, '0'),
(1, '1'),
)
data_new = ((0, "new"), (1, "new"))
test_subject = Chain()
test_subject.load('data.simple', data)
assert test_subject.get_chain('data_source').data == data
test_subject.load('data.simple', data_new)
assert test_subject.get_chain('data_source').data == data_new
def test_cluster():
data = ((0, 'data test 1'), (1, 'Data test 2'), (2, 'other data test 3.'))
test_subject = Chain().load(
'data.simple',
data).process('vectorize.sklearn').process('cluster.ward')
assert isinstance(test_subject.chain[0], SimpleDataSource)
assert isinstance(test_subject.chain[1], CountVectorizer)
assert isinstance(test_subject.chain[2], WardClusterizer)
result = test_subject.get_chain('result')
assert len(result) == len(data)
assert result.tolist() == test_subject.chain[-1].get('result').tolist()
assert result.tolist() == test_subject.chain[-1]._data['result'].tolist()
def fixture():
'''
Creates a processor chain with test data.
'''
data = ((0, 'data test 1'), (1, 'Data test 2'), (2, 'other data test 3.'))
return Chain().load('data.simple', data)
