def test_calculate_matches_complex(self):
auto = EPGProfile.from_file(example_file('Ap10-haba-auto.csv'))
manual = EPGProfile.from_file(example_file('Ap10-haba-manual.csv'))
stats = EPGComparator((auto, manual)).compare()
expected = dict(
c=dict(n=71, full_matches=45, loffset=.01, roffset=.02),
pd=dict(n=49, full_matches=37, loffset=0., roffset=.04),
np=dict(n=22, full_matches=20, loffset=.02, roffset=0.))
for waveform in expected:
for name, val in expected[waveform].items():
self.assertAlmostEqual(val, stats.get(waveform, mean)[name],
2, name)
def test_calculate_matches(self):
auto = EPGProfile.from_file(example_file('Ap1-haba-auto.csv'))
manual = EPGProfile.from_file(example_file('Ap1-haba-manual.csv'))
stats = EPGComparator((auto, manual)).compare()
expected = dict(
c=dict(n=33, full_matches=21, loffset=.03, roffset=.06),
pd=dict(n=16, full_matches=8, loffset=0., roffset=0.),
np=dict(n=26, full_matches=20, loffset=.13, roffset=.04))
for waveform in expected:
for name, val in expected[waveform].items():
self.assertAlmostEqual(val, stats.get(waveform, mean)[name],
2, name)
def test_creating_confusion_matrix(self):
profile = EPGProfile.from_file(example_file('simple-manual.csv'))
other = EPGProfile.from_file(example_file('simple-auto.csv'))
stats = EPGComparator((profile, other)).compare()
self.assertSequenceEqual(
(1953, 2, 0, 6, 1677, 0, 450, 0, 316),
tuple(stats.cfm))
self.assertEqual(2409, sum(stats.cfm.column('c')))
self.assertEqual(1679, sum(stats.cfm.column('np')))
self.assertEqual(316, sum(stats.cfm.column('pd')))
def test_calculate_offset(self):
auto = EPGProfile.from_file(example_file('simple-auto.csv'))
manual = EPGProfile.from_file(example_file('simple-manual.csv'))
auto, manual = _EPGFingerprint(auto), _EPGFingerprint(manual)
expected = dict(
pd=[[20], [10]],
c=[[10, 0], [-446, 2]],
np=[[-6, 2], [0, 0]])
offsets = EPGComparator._calculate_offsets((auto, manual),
fractional=False)
self.assertDictEqual(expected, offsets)
