def test_empty_metadata_values(self):
# Trusting that the code is sane enough to not invent a distinction
# between feature and sample metadata where there is none
sample_mc = qiime2.MetadataCategory(
pd.Series(['a_new', 'a_new', None], index=['a', 'b', 'c']))
sample_ids = sample_mc.to_series().index
data = np.array([[1, 2, 3], [30, 20, 10]])
table = biom.Table(data,
sample_ids=sample_ids,
observation_ids=['x', 'y'])
with self.assertRaisesRegex(ValueError, 'missing.*value.*{\'c\'}'):
group(table, axis='sample', metadata=sample_mc, mode='sum')
nan_mc = qiime2.MetadataCategory(
pd.Series(['a_new', float('nan'), 'a_new'], index=['a', 'b', 'c']))
with self.assertRaisesRegex(ValueError, 'missing.*value.*{\'b\'}'):
group(table, axis='sample', metadata=nan_mc, mode='sum')
empty_str = qiime2.MetadataCategory(
pd.Series(['', 'y_new'], index=['x', 'y']))
with self.assertRaisesRegex(ValueError, 'missing.*value.*{\'x\'}'):
group(table,
axis='feature',
metadata=empty_str,
mode='median-ceiling')
def _shared_setup(self):
sample_mc = qiime2.MetadataCategory(
pd.Series([
'treatment', 'treatment', 'control', 'other', 'control',
'other', 'other'
],
index=['a', 'b', 'c', 'd', 'e', 'f', 'g']))
feature_mc = qiime2.MetadataCategory(
pd.Series(['g0', 'g1', 'g1', 'g1', 'g0'],
index=['v', 'w', 'x', 'y', 'z']))
data = np.array([
# t t c o c o o
# a b c d e f g
[0, 0, 0, 0, 1, 0, 2], # v g0
[10, 10, 10, 10, 10, 100, 1], # w g1
[12, 3, 14, 0, 0, 3, 34], # x g1
[1, 1, 1, 1, 1, 1, 1], # y g1
[0, 1, 11, 111, 1111, 20, 20]
]) # z g0
table = biom.Table(data,
sample_ids=sample_mc.to_series().index,
observation_ids=feature_mc.to_series().index)
return sample_mc, feature_mc, table
def test_numeric(self):
data = np.array([[1, 2, 3], [30, 20, 10]])
table = biom.Table(data,
sample_ids=['a', 'b', 'c'],
observation_ids=['x', 'y'])
# ints
sample_mc = qiime2.MetadataCategory(
pd.Series(['1', '2', '3'], index=['a', 'b', 'c']))
with self.assertRaisesRegex(ValueError, 'numeric'):
group(table, axis='sample', metadata=sample_mc, mode='sum')
# floats
sample_mc = qiime2.MetadataCategory(
pd.Series(['1.1', '2.2', '3.3333'], index=['a', 'b', 'c']))
with self.assertRaisesRegex(ValueError, 'numeric'):
group(table, axis='sample', metadata=sample_mc, mode='sum')
# mixed
sample_mc = qiime2.MetadataCategory(
pd.Series(['0', '42', '4.2'], index=['a', 'b', 'c']))
with self.assertRaisesRegex(ValueError, 'numeric'):
group(table, axis='sample', metadata=sample_mc, mode='sum')
def setUp(self):
barcodes = [('@s1/2 abc/2', 'AAAA', '+', 'YYYY'),
('@s2/2 abc/2', 'TTAA', '+', 'PPPP'),
('@s3/2 abc/2', 'AACC', '+', 'PPPP'),
('@s4/2 abc/2', 'TTAA', '+', 'PPPP'),
('@s5/2 abc/2', 'AACC', '+', 'PPPP'),
('@s6/2 abc/2', 'AAAA', '+', 'PPPP'),
('@s7/2 abc/2', 'CGGC', '+', 'PPPP'),
('@s8/2 abc/2', 'GGAA', '+', 'PPPP'),
('@s9/2 abc/2', 'CGGC', '+', 'PPPP'),
('@s10/2 abc/2', 'CGGC', '+', 'PPPP'),
('@s11/2 abc/2', 'GGAA', '+', 'PPPP')]
self.sequences = [('@s1/1 abc/1', 'GGG', '+', 'YYY'),
('@s2/1 abc/1', 'CCC', '+', 'PPP'),
('@s3/1 abc/1', 'AAA', '+', 'PPP'),
('@s4/1 abc/1', 'TTT', '+', 'PPP'),
('@s5/1 abc/1', 'ATA', '+', 'PPP'),
('@s6/1 abc/1', 'TAT', '+', 'PPP'),
('@s7/1 abc/1', 'CGC', '+', 'PPP'),
('@s8/1 abc/1', 'GCG', '+', 'PPP'),
('@s9/1 abc/1', 'ACG', '+', 'PPP'),
('@s10/1 abc/1', 'GCA', '+', 'PPP'),
('@s11/1 abc/1', 'TGA', '+', 'PPP')]
self.bsi = BarcodeSequenceFastqIterator(barcodes, self.sequences)
barcode_map = pd.Series(['AAAA', 'AACC', 'TTAA', 'GGAA', 'CGGC'],
index=['sample1', 'sample2', 'sample3',
'sample4', 'sample5'])
self.barcode_map = qiime2.MetadataCategory(barcode_map)
def test_single_sample(self):
bsi = BarcodeSequenceFastqIterator(self.barcodes[:1],
self.sequences[:1])
barcode_map = pd.Series(['AAAA'], index=['sample1'])
barcode_map = qiime2.MetadataCategory(barcode_map)
demux_data = emp_single(bsi, barcode_map)
# test that an index.html file is created and that it has size > 0
# TODO: Remove _PlotQualView wrapper
with tempfile.TemporaryDirectory() as output_dir:
result = summarize(output_dir, _PlotQualView(demux_data,
paired=False), n=1)
self.assertTrue(result is None)
index_fp = os.path.join(output_dir, 'overview.html')
self.assertTrue(os.path.exists(index_fp))
self.assertTrue(os.path.getsize(index_fp) > 0)
csv_fp = os.path.join(output_dir, 'per-sample-fastq-counts.csv')
self.assertTrue(os.path.exists(csv_fp))
self.assertTrue(os.path.getsize(csv_fp) > 0)
pdf_fp = os.path.join(output_dir, 'demultiplex-summary.pdf')
self.assertFalse(os.path.exists(pdf_fp))
png_fp = os.path.join(output_dir, 'demultiplex-summary.png')
self.assertFalse(os.path.exists(png_fp))
with open(index_fp, 'r') as fh:
html = fh.read()
self.assertIn('<td>Minimum:</td><td>1</td>', html)
self.assertIn('<td>Maximum:</td><td>1</td>', html)
def test_paired_end(self):
barcodes = self.barcodes[:3]
forward = self.sequences[:3]
reverse = [('@s1/1 abc/1', 'CCC', '+', 'YYY'),
('@s2/1 abc/1', 'GGG', '+', 'PPP'),
('@s3/1 abc/1', 'TTT', '+', 'PPP')]
bpsi = BarcodePairedSequenceFastqIterator(barcodes, forward, reverse)
barcode_map = pd.Series(['AAAA', 'AACC', 'TTAA'],
index=['sample1', 'sample2', 'sample3'])
barcode_map = qiime2.MetadataCategory(barcode_map)
demux_data = emp_paired(bpsi, barcode_map)
with tempfile.TemporaryDirectory() as output_dir:
result = summarize(output_dir, _PlotQualView(demux_data,
paired=True), n=2)
self.assertTrue(result is None)
plot_fp = os.path.join(output_dir, 'quality-plot.html')
with open(plot_fp, 'r') as fh:
html = fh.read()
self.assertIn('<h5 class="text-center">Forward Reads</h5>',
html)
self.assertIn('<h5 class="text-center">Reverse Reads</h5>',
html)
def test_permanova_pairwise(self):
dm = skbio.DistanceMatrix(
[[0.00, 0.25, 0.25], [0.25, 0.00, 0.00], [0.25, 0.00, 0.00]],
ids=['sample1', 'sample2', 'sample3'])
md = qiime2.MetadataCategory(
pd.Series(['a', 'b', 'b'],
name='a or b',
index=['sample1', 'sample2', 'sample3']))
with tempfile.TemporaryDirectory() as output_dir:
beta_group_significance(output_dir, dm, md, pairwise=True)
index_fp = os.path.join(output_dir, 'index.html')
self.assertTrue(os.path.exists(index_fp))
# all expected boxplots are generated
self.assertTrue(
os.path.exists(os.path.join(output_dir, 'a-boxplots.pdf')))
self.assertTrue(
os.path.exists(os.path.join(output_dir, 'a-boxplots.png')))
self.assertTrue(
os.path.exists(os.path.join(output_dir, 'b-boxplots.pdf')))
self.assertTrue(
os.path.exists(os.path.join(output_dir, 'b-boxplots.png')))
# no extra boxplots are generated
self.assertEqual(len(glob.glob('%s/*-boxplots.pdf' % output_dir)),
2)
self.assertEqual(len(glob.glob('%s/*-boxplots.png' % output_dir)),
2)
self.assertTrue('PERMANOVA results' in open(index_fp).read())
self.assertTrue('Pairwise permanova' in open(index_fp).read())
self.assertFalse('Warning' in open(index_fp).read())
def test_with_metadata(self):
md = qiime2.MetadataCategory(
pd.Series(['milo', 'summer', 'russ'],
name='pet',
index=['S1', 'S2', 'S3']))
heatmap(self.output_dir, self.table, metadata=md)
self.assertBasicVizValidity(self.output_dir)
def test_one_sample(self):
md = qiime2.MetadataCategory(
pd.Series([1.5], name='number', index=['sample1']))
exp = skbio.DistanceMatrix([[0.0]], ids=['sample1'])
obs = distance_matrix(md)
self.assertEqual(exp, obs)
def test_missing_values(self):
md = qiime2.MetadataCategory(
pd.Series([1.0, 2.0, np.nan, 4.0],
name='number',
index=['sample1', 'sample2', 'sample3', 'sample4']))
with self.assertRaisesRegex(ValueError, 'missing values'):
distance_matrix(md)
def test_identity_groups(self):
# These map to the same values as before
sample_mc = qiime2.MetadataCategory(
pd.Series(['a', 'b', 'c'], index=['a', 'b', 'c']))
feature_mc = qiime2.MetadataCategory(
pd.Series(['x', 'y'], index=['x', 'y']))
table = biom.Table(np.array([[1, 2, 3], [30, 20, 10]]),
sample_ids=sample_mc.to_series().index,
observation_ids=feature_mc.to_series().index)
# Sample x Sum
result = group(table, axis='sample', metadata=sample_mc, mode='sum')
self.assertEqual(table, result)
# Sample x Mean
result = group(table,
axis='sample',
metadata=sample_mc,
mode='mean-ceiling')
self.assertEqual(table, result)
# Sample x Median
result = group(table,
axis='sample',
metadata=sample_mc,
mode='median-ceiling')
self.assertEqual(table, result)
# Feature x Sum
result = group(table, axis='feature', metadata=feature_mc, mode='sum')
self.assertEqual(table, result)
# Feature x Mean
result = group(table,
axis='feature',
metadata=feature_mc,
mode='mean-ceiling')
self.assertEqual(table, result)
# Feature x Median
result = group(table,
axis='feature',
metadata=feature_mc,
mode='median-ceiling')
self.assertEqual(table, result)
def test_no_sample_cluster(self):
md = qiime2.MetadataCategory(
pd.Series(['milo', 'summer', 'russ'],
name='pet',
index=['S1', 'S2', 'S3']))
heatmap(self.output_dir, self.table, metadata=md, cluster='features')
self.assertBasicVizValidity(self.output_dir)
def test_non_numeric_category(self):
md = qiime2.MetadataCategory(
pd.Series(['x1', 'x2', '3', '4'],
name='number',
index=['sample1', 'sample2', 'sample3', 'sample4']))
with self.assertRaisesRegex(ValueError,
'non-numeric values.*\n\n.*x1'):
distance_matrix(md)
def test_bad_method(self):
alpha_div = pd.Series([2.0, 4.0, 6.0], name='alpha-div',
index=['sample1', 'sample2', 'sample3'])
md = qiime2.MetadataCategory(
pd.Series(['1.0', '2.0', '3.0'], name='value',
index=['sample1', 'sample2', 'sample3']))
with tempfile.TemporaryDirectory() as output_dir:
with self.assertRaises(ValueError):
alpha_correlation(output_dir, alpha_div, md, method='bad!')
def test_error_on_missing_metadata(self):
dm = skbio.DistanceMatrix(
[[0.00, 0.25, 0.25], [0.25, 0.00, 0.00], [0.25, 0.00, 0.00]],
ids=['sample1', 'sample2', 'sample3'])
md = qiime2.MetadataCategory(
pd.Series([1, 2], name='number', index=['sample1', 'sample2']))
with tempfile.TemporaryDirectory() as output_dir:
with self.assertRaisesRegex(ValueError, 'no data: sample3'):
beta_correlation(output_dir, dm, md)
def test_float_category(self):
md = qiime2.MetadataCategory(
pd.Series([1.5, 2.0, 3.0],
name='number',
index=['sample1', 'sample2', 'sample3']))
exp = skbio.DistanceMatrix(
[[0.0, 0.5, 1.5], [0.5, 0.0, 1.0], [1.5, 1.0, 0.0]],
ids=['sample1', 'sample2', 'sample3'])
obs = distance_matrix(md)
self.assertEqual(exp, obs)
def test_empty_table(self):
mc = qiime2.MetadataCategory(
pd.Series(['a_new', 'b_new'], index=['a', 'b']))
table = biom.Table(np.array([[]]), sample_ids=[], observation_ids=[])
with self.assertRaisesRegex(ValueError, 'empty table'):
group(table, axis='sample', metadata=mc, mode='sum')
with self.assertRaisesRegex(ValueError, 'empty table'):
group(table, axis='feature', metadata=mc, mode='sum')
def test_int_category(self):
md = qiime2.MetadataCategory(
pd.Series([1, 2, 3],
name='number',
index=['sample1', 'sample2', 'sample3']))
exp = skbio.DistanceMatrix([[0, 1, 2], [1, 0, 1], [2, 1, 0]],
ids=['sample1', 'sample2', 'sample3'])
obs = distance_matrix(md)
self.assertEqual(exp, obs)
def test_invalid_method(self):
dm = skbio.DistanceMatrix(
[[0.00, 0.25, 0.25], [0.25, 0.00, 0.00], [0.25, 0.00, 0.00]],
ids=['sample1', 'sample2', 'sample3'])
md = qiime2.MetadataCategory(
pd.Series(['a', 'b', 'b'],
name='a or b',
index=['sample1', 'sample2', 'sample3']))
with self.assertRaises(ValueError):
with tempfile.TemporaryDirectory() as output_dir:
beta_group_significance(output_dir, dm, md, method='bad!')
def test_error_on_non_numeric_metadata(self):
dm = skbio.DistanceMatrix(
[[0.00, 0.25, 0.25], [0.25, 0.00, 0.00], [0.25, 0.00, 0.00]],
ids=['sample1', 'sample2', 'sample3'])
md = qiime2.MetadataCategory(
pd.Series([1.0, 2.0, 'hello-world'],
name='number',
index=['sample1', 'sample2', 'sample3']))
with tempfile.TemporaryDirectory() as output_dir:
with self.assertRaisesRegex(ValueError, 'Non-numeric data was'):
beta_correlation(output_dir, dm, md)
def test_extra_metadata(self):
dm = skbio.DistanceMatrix(
[[0.00, 0.25, 0.25], [0.25, 0.00, 0.00], [0.25, 0.00, 0.00]],
ids=['sample1', 'sample2', 'sample3'])
md = qiime2.MetadataCategory(
pd.Series(['a', 'b', 'b', 'c'],
name='a or b',
index=['sample1', 'sample2', 'sample3', 'sample4']))
with tempfile.TemporaryDirectory() as output_dir:
beta_group_significance(output_dir, dm, md, permutations=42)
index_fp = os.path.join(output_dir, 'index.html')
self.assertTrue('<td>2</td>' in open(index_fp).read())
def test_missing_feature_ids(self):
feature_mc = qiime2.MetadataCategory(
pd.Series(['g0', 'g1', 'g2', 'g1', 'g2', 'extra'],
index=['a', 'c', 'd', 'e', 'f', 'g']))
data = np.array([[1, 0, 0], [1, 10, 10], [0, 0, 100], [5, 5, 5],
[0, 1, 100], [7, 8, 9]])
# g is missing on purpose
table = biom.Table(data,
sample_ids=['s1', 's2', 's3'],
observation_ids=['a', 'b', 'c', 'd', 'e', 'f'])
with self.assertRaisesRegex(ValueError, 'metadata.*missing: {\'b\'}'):
group(table, axis='feature', metadata=feature_mc, mode='sum')
def test_filtered_samples_str_metadata(self):
dm = skbio.DistanceMatrix(
[[0.00, 0.25, 0.25, 0.66], [0.25, 0.00, 0.00, 0.66],
[0.25, 0.00, 0.00, 0.66], [0.66, 0.66, 0.66, 0.00]],
ids=['sample1', 'sample2', 'sample3', 'sample4'])
md = qiime2.MetadataCategory(
pd.Series(['a', 'b', 'b', ''],
name='a or b',
index=['sample1', 'sample2', 'sample3', 'sample4']))
with tempfile.TemporaryDirectory() as output_dir:
beta_group_significance(output_dir, dm, md)
index_fp = os.path.join(output_dir, 'index.html')
self.assertTrue('Warning' in open(index_fp).read())
def test_str_casting(self):
md = qiime2.MetadataCategory(
pd.Series(['1', '2', '3', '4'],
name='number',
index=['sample1', 'sample2', 'sample3', 'sample4']))
exp = skbio.DistanceMatrix(
[[0.0, 1.0, 2.0, 3.0], [1.0, 0.0, 1.0, 2.0], [2.0, 1.0, 0.0, 1.0],
[3.0, 2.0, 1.0, 0.0]],
ids=['sample1', 'sample2', 'sample3', 'sample4'])
obs = distance_matrix(md)
self.assertEqual(exp, obs)
def test_missing_sample_ids(self):
sample_mc = qiime2.MetadataCategory(
pd.Series(['g0', 'g2', 'g0', 'g2'], index=['s1', 's3', 's4',
's6']))
data = np.array([[0, 1, 2, 3], [10, 11, 12, 13], [100, 110, 120, 130]])
table = biom.Table(data,
sample_ids=['s1', 's2', 's4', 's5'],
observation_ids=['x', 'y', 'z'])
with self.assertRaisesRegex(ValueError, 'metadata.*missing:') as e:
group(table, axis='sample', metadata=sample_mc, mode='sum')
self.assertIn('s2', str(e.exception))
self.assertIn('s5', str(e.exception))
def test_reorder(self):
sample_mc = qiime2.MetadataCategory(
pd.Series(['c', 'b', 'a'], index=['c', 'b', 'a']))
data = np.array([[1, 2, 3], [30, 20, 10]])
table = biom.Table(data,
sample_ids=['a', 'b', 'c'],
observation_ids=['x', 'y'])
expected = biom.Table(np.array([[3, 2, 1], [10, 20, 30]]),
sample_ids=['c', 'b', 'a'],
observation_ids=['x', 'y'])
result = group(table, axis='sample', metadata=sample_mc, mode='sum')
self.assertEqual(expected, result)
def test_superset_sample_group(self):
sample_mc = qiime2.MetadataCategory(
pd.Series(['g0', 'g1', 'g2', 'g0', 'g1', 'g2'],
index=['s1', 's2', 's3', 's4', 's5', 's6']))
data = np.array([[0, 1, 2, 3], [10, 11, 12, 13], [100, 110, 120, 130]])
table = biom.Table(data,
sample_ids=['s1', 's2', 's4', 's5'],
observation_ids=['x', 'y', 'z'])
expected = biom.Table(np.array([[2, 4], [22, 24], [220, 240]]),
sample_ids=['g0', 'g1'],
observation_ids=['x', 'y', 'z'])
result = group(table, axis='sample', metadata=sample_mc, mode='sum')
self.assertEqual(expected, result)
def test_evaluate_composition_metadata_not_superset(self):
incomplete_md = qiime2.MetadataCategory(
pd.DataFrame({'mock_id': ['there_can_only_be_one']},
index=['s3'])['mock_id'])
with self.assertRaisesRegex(ValueError, "Missing samples in metadata"):
_evaluate_composition(self.exp_one_sample,
self.obs,
depth=7,
palette='Set1',
plot_tar=True,
plot_tdr=True,
plot_r_value=True,
plot_r_squared=True,
plot_observed_features=True,
plot_observed_features_ratio=True,
metadata=incomplete_md)
def test_superset_feature_group(self):
feature_mc = qiime2.MetadataCategory(
pd.Series(['g0', 'g0', 'g1', 'g2', 'g1', 'g2', 'extra'],
index=['a', 'b', 'c', 'd', 'e', 'f', 'g']))
data = np.array([[1, 0, 0], [1, 10, 10], [0, 0, 100], [5, 5, 5],
[0, 1, 100], [7, 8, 9]])
# g is missing on purpose
table = biom.Table(data,
sample_ids=['s1', 's2', 's3'],
observation_ids=['a', 'b', 'c', 'd', 'e', 'f'])
expected = biom.Table(np.array([[2, 10, 10], [0, 1, 200], [12, 13,
14]]),
sample_ids=['s1', 's2', 's3'],
observation_ids=['g0', 'g1', 'g2'])
result = group(table, axis='feature', metadata=feature_mc, mode='sum')
self.assertEqual(expected, result)
def test_subsample_higher_than_seqs_count(self):
barcodes = self.barcodes[:1]
sequences = self.sequences[:1]
bsi = BarcodeSequenceFastqIterator(barcodes, sequences)
barcode_map = pd.Series(['AAAA'], index=['sample1'])
barcode_map = qiime2.MetadataCategory(barcode_map)
demux_data = emp_single(bsi, barcode_map)
with tempfile.TemporaryDirectory() as output_dir:
result = summarize(output_dir, _PlotQualView(demux_data,
paired=False), n=50)
self.assertTrue(result is None)
plot_fp = os.path.join(output_dir, 'quality-plot.html')
with open(plot_fp, 'r') as fh:
html = fh.read()
self.assertIn('<strong>Warning:</strong>', html)
