def test_similarity_threshold_real(testfiles):
annotation = first_entry(testfiles, 'annotation')
probes = first_entry(testfiles, 'probes')
microarray = first_entry(testfiles, 'microarray')
out1 = samples_.similarity_threshold(microarray, annotation, probes)
out2 = samples_.similarity_threshold(microarray,
annotation,
probes,
threshold=np.inf)
assert out1.shape[0] < out2.shape[0]
def test_collapse_probes_donors(testfiles, donor_probes):
out = probes_.collapse_probes(flatten_dict(testfiles, 'microarray'),
flatten_dict(testfiles, 'annotation'),
first_entry(testfiles, 'probes'),
method='max_intensity',
donor_probes=donor_probes)
out = list(out.values())
assert len(out) == 2 # number of donors
assert np.all([len(exp) == n_samp for exp, n_samp in zip(out, [363, 470])])
assert np.all([len(exp.columns) == 29131 for exp in out])
def test_collapse_probes(testfiles, method):
# we've aleady tested the underlying methods so here we just want to do
# some smoke tests to make sure the function returns what we expected
# regardless of the provided method
out = probes_.collapse_probes(flatten_dict(testfiles, 'microarray'),
flatten_dict(testfiles, 'annotation'),
first_entry(testfiles, 'probes'),
method=method)
out = list(out.values())
assert len(out) == 2 # number of donors
assert np.all([len(exp) == n_samp for exp, n_samp in zip(out, [363, 470])])
assert np.all([len(exp.columns) == 29131 for exp in out])
def test_reannotate_probes(testfiles):
# set up a few useful variables
probe_file = first_entry(testfiles, 'probes')
# should work with either a filename _or_ a dataframe
reannot = probes_.reannotate_probes(probe_file)
probe_df = abagen.io.read_probes(probe_file)
pd.testing.assert_frame_equal(reannot, probes_.reannotate_probes(probe_df))
# expected output
cols = ['probe_name', 'gene_symbol', 'entrez_id']
assert np.all(reannot.columns == cols)
assert reannot.index.name == 'probe_id'
assert reannot.shape == (45821, 3)
def test_filter_probes(testfiles, threshold, expected_length):
# set up a few useful variables
pacall = flatten_dict(testfiles, 'pacall')
probe_file = first_entry(testfiles, 'probes')
samples = flatten_dict(testfiles, 'annotation')
probe_df = abagen.io.read_probes(probe_file)
# should work with either a filename _or_ a dataframe
filtered = probes_.filter_probes(pacall,
samples,
probe_file,
threshold=threshold)
pd.testing.assert_frame_equal(
filtered,
probes_.filter_probes(pacall, samples, probe_df, threshold=threshold))
# provided threshold returns expected output
cols = [
'probe_name', 'gene_id', 'gene_symbol', 'gene_name', 'entrez_id',
'chromosome'
]
assert np.all(filtered.columns == cols)
assert filtered.index.name == 'probe_id'
assert len(filtered) == expected_length
def test_first_entry(indict, subkey, out):
assert utils.first_entry(indict, subkey) == out
def test_collapse_probes_errors(testfiles):
with pytest.raises(ValueError):
probes_.collapse_probes(flatten_dict(testfiles, 'microarray'),
flatten_dict(testfiles, 'annotation'),
first_entry(testfiles, 'probes'),
method='notamethod')
def test_label_samples_real(testfiles, atlas):
out = samples_.label_samples(first_entry(testfiles, 'annotation'),
atlas['image'])
assert isinstance(out, pd.DataFrame)
assert out.index.name == 'sample_id'
assert out.columns == ['label']