def test_kmer_map_to_metadata(self):
known = pd.DataFrame(data=[[
'Batman', 'Batman', 'Batman', 'Gotham', 'WANTCAT', 'CATCATCAT', 50.
],
[
'Superman', 'Superman', 'Superman',
'Metropolis', 'CATDAD', 'DADCAT', 50.
]],
columns=[
'db-seq', 'seq-name', 'kmer', 'region',
'fwd-primer', 'rev-primer', 'kmer-length'
],
index=pd.Index(['0', '1'], name='id'))
filepath = os.path.join(self.base_dir, 'kmer-map.tsv')
format = KmerMapFormat(filepath, mode='r')
test = t._2(format)
self.assertTrue(isinstance(test, Metadata))
columns = dict(test.columns)
npt.assert_array_equal(list(columns.keys()), [
'db-seq', 'seq-name', 'kmer', 'region', 'fwd-primer', 'rev-primer',
'kmer-length'
])
for k, v in columns.items():
if k == 'kmer-length':
self.assertEqual(v.type, 'numeric')
else:
self.assertEqual(v.type, 'categorical')
pdt.assert_frame_equal(known, test.to_dataframe())
def test_kmer_map_delayed_frame(self):
known = pd.DataFrame(data=[[
'Batman', 'Batman', 'Gotham', 'WANTCAT', 'CATCATCAT', 50
], ['Superman', 'Superman', 'Metropolis', 'CATDAD', 'DADCAT', 50]],
columns=[
'seq-name', 'kmer', 'region', 'fwd-primer',
'rev-primer', 'kmer-length'
],
index=pd.Index(['Batman', 'Superman'],
name='db-seq'))
filepath = os.path.join(self.base_dir, 'kmer-map.tsv')
format = KmerMapFormat(filepath, mode='r')
test = t._3(format)
self.assertTrue(isinstance(test, dd.DataFrame))
pdt.assert_frame_equal(known, test.compute())
def _4(obj: pd.DataFrame) -> KmerMapFormat:
ff = KmerMapFormat()
obj.to_csv(str(ff), sep='\t')
return ff