def test_convolve_hrf():
# Check that using an HRF convolution on two events is the same as doing the
# same convolution on each of them and summing
t = Term('t')
glover_t = hrf.glover(t)
# dt is exactly representable in floating point here
glover_conv = TimeConvolver(glover_t, [0, 26], 1. / 32)
event0 = blocks(((1, 2), ), (2., ))
event1 = blocks(((15, 20), ), (1., ))
conved0 = glover_conv.convolve(event0, [0, 3])
conved1 = glover_conv.convolve(event1, [14, 21])
times = np.arange(0, 50, 0.1)
e0_e1 = lambdify_t(conved0)(times) + lambdify_t(conved1)(times)
# Same thing in one shot
events = blocks(((1, 2), (15, 20)), (2., 1))
conved = glover_conv.convolve(events, [0, 21])
assert_almost_equal(lambdify_t(conved)(times), e0_e1)
# Same thing with convolve_functions
conved_cf = convolve_functions(events, glover_t, [0, 21], [0, 26], 1 / 32.)
assert_almost_equal(lambdify_t(conved_cf)(times), e0_e1)
define,
events,
blocks,
interp,
linear_interp,
step_function,
convolve_functions,
)
from .. import hrf
from nose.tools import (assert_equal, assert_false, raises, assert_raises)
from numpy.testing import (assert_array_equal, assert_array_almost_equal,
assert_almost_equal)
t = Term('t')
def test_define():
expr = sympy.exp(3 * t)
assert_equal(str(expr), 'exp(3*t)')
newf = define('f', expr)
assert_equal(str(newf), 'f(t)')
f = lambdify_t(newf)
tval = np.random.standard_normal((3, ))
assert_almost_equal(np.exp(3 * tval), f(tval))
def test_events():
# test events utility function
h = Function('hrf')