def test_time(n_samples=1000, n_features=100, leaf_size=1, k=20):
X = numpy.random.random([n_samples, n_features])
print "---------------------------------------------------"
print "%i neighbors of %i points in %i dimensions:" % (k, n_samples,
n_features)
print " (leaf size = %i)" % leaf_size
print " -------------"
t0 = time()
BT = BallTree(X, leaf_size)
print " Ball Tree construction : %.3g sec" % (time() - t0)
d, nbrs1 = BT.query(X, k)
print " total (construction+query) : %.3g sec" % (time() - t0)
print " -------------"
t0 = time()
KDT = cKDTree(X, leaf_size)
print " KD tree construction : %.3g sec" % (time() - t0)
d, nbrs2 = KDT.query(X, k)
print " total (construction+query) : %.3g sec" % (time() - t0)
print " -------------"
print " neighbors match: ",
print compare_nbrs(nbrs1, nbrs2)
print " -------------"
return np.all(nbrs1 == nbrs2)
if __name__ == '__main__':
n_samples = 1000
leaf_size = 1 # leaf size
k = 20
BT_results = []
KDT_results = []
for i in range(1, 10):
print 'Iteration %s' %i
n_features = i*100
X = np.random.random([n_samples, n_features])
t0 = time()
BT = BallTree(X, leaf_size)
d, nbrs1 = BT.query(X, k)
delta = time() - t0
BT_results.append(delta)
t0 = time()
KDT = cKDTree(X, leaf_size)
d, nbrs2 = KDT.query(X, k)
delta = time() - t0
KDT_results.append(delta)
# this checks we get the correct result
assert compare_nbrs(nbrs1, nbrs2)
xx = 100 * np.arange(1, 10)
pl.plot(xx, BT_results, label='scikits.learn (BallTree)')
