def test1(self):
"""Easier test case with known alignment
In this case we know that bt[156:] are matched to nt[:] within
an error term of 100ms.
"""
nt = np.loadtxt('test/ntimes')
bt = np.loadtxt('test/btimes')
fit = EventSync.sync(bt, nt, min_acceptable_error=.01,
gross_delay_scales_start=None, gross_delay_scales_factor=2.0,
clock_sync_guess=.9966361, n_scales=4)
self.assertEqual(len(fit), 804)
self.assertTrue(np.all(np.diff(fit.xi2yi_table) == -156))
def test2(self):
"""Harder test case with imperfect alignment
In this case the `mt` have error bars of about 500ms and also
there are many missing and spurious entries.
A failure on this test is not a deal breaker since it is probably
not well-posed. It would be nice if it could fit at least 40 though,
which is 80% of the values in `mt`.
"""
bt = np.loadtxt('test/btimes')
mt = np.loadtxt('test/mtimes')
fit = EventSync.sync(bt, mt, min_acceptable_error=.5,
gross_delay_scales_start=None, gross_delay_scales_factor=2.0,
clock_sync_guess=.9966361, n_scales=4)
self.assertTrue(len(fit) > 40)
self.assertTrue(fit.xi2yi[42] == 0)
self.assertTrue(fit.xi2yi[580] == 51)
def test9(self):
"""Max positive delay"""
sig1 = np.array([0, 0, 1])
sig2 = np.array([1, 0, 0, 0])
d, p = EventSync.estimate_delay(sig1, sig2)
self.assertTrue(d == 2)
def test9(self):
"""Unequal lengths, delay -1, sig2 longer"""
sig1 = np.array([0, 1, 0])
sig2 = np.array([0, 0, 1, 0])
d, p = EventSync.estimate_delay(sig1, sig2)
self.assertTrue(d == -1)
def test5(self):
"""Equal lengths, delay 1"""
sig1 = np.array([0, 0, 1])
sig2 = np.array([0, 1, 0])
d, p = EventSync.estimate_delay(sig1, sig2)
self.assertTrue(d == 1)
def test1(self):
"""Equal lengths, zero delay"""
sig1 = np.array([0, 1, 0])
sig2 = np.array([0, 1, 0])
d, p = EventSync.estimate_delay(sig1, sig2)
self.assertTrue(d == 0)