def test_math(self):
nTime = 4
nFrequency = 2
shutil.rmtree(os.path.dirname(os.path.realpath(__file__)) +
'/test3.treasure',
ignore_errors=True)
treasure = omni.Treasure(os.path.dirname(os.path.realpath(__file__)) +
'/test3.treasure',
nlst=nTime,
nfreq=nFrequency)
treasure.add_coin(('xx', np.array([0, 2, 3])))
treasure.update_coin(('xx', np.array([0, 2, 3])),
(treasure.lsts + treasure.lsts[1] *
(nTime / 2. + .5)) % (2 * np.pi),
np.outer(np.arange(nTime), np.arange(nFrequency)),
np.ones((nTime, nFrequency)))
predict_result = np.outer(
np.roll(np.append(
[0], (np.arange(nTime - 1) + np.arange(1, nTime)) / 2.),
nTime / 2,
axis=0), np.arange(nFrequency))
#print (treasure.lsts + treasure.lsts[1] * (nTime/2. + .5))%(2*np.pi), np.outer(np.arange(nTime), np.arange(nFrequency))
#print treasure.get_coin(('xx', np.array([0,2,3]))).mean
#print predict_result
#print predict_result - treasure.get_coin(('xx', np.array([0,2,3]))).mean
np.testing.assert_almost_equal(predict_result,
treasure.get_coin(
('xx', np.array([0, 2, 3]))).mean,
decimal=14)
def test_IO(self):
nTime = 3
nFrequency = 5
shutil.rmtree(os.path.dirname(os.path.realpath(__file__)) +
'/test.treasure',
ignore_errors=True)
shutil.rmtree(os.path.dirname(os.path.realpath(__file__)) +
'/test2.treasure',
ignore_errors=True)
treasure = omni.Treasure(os.path.dirname(os.path.realpath(__file__)) +
'/test.treasure',
nlst=nTime,
nfreq=nFrequency)
treasure.add_coin(('xx', np.array([0, 2, 3])))
treasure.add_coin(('xx', np.array([1, 2, 3])))
self.assertEqual(
treasure.coin_name(('xx', np.array([1, 2, 3]))),
os.path.dirname(os.path.realpath(__file__)) +
'/test.treasure//xx1.coin')
treasure2 = treasure.duplicate_treasure(
os.path.dirname(os.path.realpath(__file__)) + '/test2.treasure')
treasure.burn()
treasure2.add_coin(('xx', np.array([1, 2, 3])))
treasure2.add_coin(('xx', np.array([1, 2, 4])))
self.assertEqual(
treasure2.coin_name(('xx', np.array([1, 2, 4]))),
os.path.dirname(os.path.realpath(__file__)) +
'/test2.treasure//xx2.coin')
self.assertEqual(treasure2.coinShape, (nTime, nFrequency, 10))
treasure2.burn()
def test_probability(self):
nTime = 10
nFrequency = 1
shutil.rmtree(os.path.dirname(os.path.realpath(__file__)) +
'/test3.treasure',
ignore_errors=True)
treasure = omni.Treasure(os.path.dirname(os.path.realpath(__file__)) +
'/test3.treasure',
nlst=nTime,
nfreq=nFrequency)
treasure.add_coin(('xx', np.array([0, 2, 3])))
treasure.add_coin(('xx', np.array([1, 2, 3])))
nupdate = 4
update_lsts = np.append(
(treasure.lsts[-nupdate / 2:] + np.pi / 2 / nTime),
(treasure.lsts[:nupdate / 2] + np.pi / 2 / nTime))
nan_prob = .1
trials = 10000
for i in range(int(trials / (1 - nan_prob))):
#print i
vis_re = (np.random.randn(nupdate) * (np.arange(nupdate) + 1) +
range(nupdate)).reshape(nupdate, 1)
vis_im = (np.random.randn(nupdate) * (np.arange(nupdate) + 1) +
range(nupdate)).reshape(nupdate, 1)
epsilons = (np.arange(nupdate, dtype='float') + 1).reshape(
nupdate, 1)
if random.random() < nan_prob:
vis_re[:nupdate / 2] = vis_re[:nupdate / 2] + np.nan
if random.random() < nan_prob:
vis_re[-nupdate / 2:] = vis_re[-nupdate / 2:] + np.nan
treasure.update_coin(('xx', np.array([1, 2, 3])), update_lsts,
vis_re + 1.j * vis_im, epsilons**2)
#print epsilons**2
c = treasure.get_coin(('xx', np.array([1, 2, 3])))
#print c.count, c.mean, c.weighted_mean
#print c.variance_re, c.variance_im
#print c.weighted_variance
self.assertTrue(abs(c.count[1] - trials) < 3 * trials**.5)
self.assertTrue(abs(c.count[-1] - trials) < 3 * trials**.5)
sigma1 = (1 / 16. * epsilons[-2]**2 + 9 / 16. * epsilons[-1]**2)**.5
sigma2 = (1 / 16. * epsilons[0]**2 + 9 / 16. * epsilons[1]**2)**.5
for var in [c.weighted_variance, c.variance_re, c.variance_im]:
weighted_sigma = (var * trials)**.5
#print weighted_sigma, sigma1, sigma2
self.assertTrue(
abs(weighted_sigma[1] - sigma1) / sigma1 < 3 * trials**-.5)
self.assertTrue(
abs(weighted_sigma[-1] - sigma2) / sigma2 < 3 * trials**-.5)
self.assertTrue(
abs(c.mean[1] - 2.75 - 2.75j) < 1.414 * 3 * sigma1 * trials**-.5)
self.assertTrue(
abs(c.weighted_mean[1] - 2.75 - 2.75j) < 1.414 * 3 * sigma1 *
trials**-.5)
self.assertTrue(
abs(c.mean[-1] - .75 - .75j) < 1.414 * 3 * sigma2 * trials**-.5)
self.assertTrue(
abs(c.weighted_mean[-1] - .75 - .75j) < 1.414 * 3 * sigma2 *
trials**-.5)
treasure.burn()
help='Max value of plotting.')
o.add_option('--min',
action='store',
type='float',
default=None,
help='Min value of plotting.')
o.add_option('-u',
'--ubl_order',
action='store',
type='int',
default=0,
help='Which baseline to plot starting from shortest as 0.')
opts, args = o.parse_args(sys.argv[1:])
treasure = omni.Treasure(args[0])
item = args[1]
if item not in [
'count', 'variance_re', 'variance_im', 'weighted_mean', 'mean',
'weighted_variance'
]:
print "%s not recognized." % item
exit()
if item in ['mean', 'weighted_mean']:
if len(args) < 3 or args[2] not in ['amp', 'phs']:
print "Please specify amp or phs"
exit()
for p, pol in enumerate(treasure.ubls.keys()):
c = treasure.get_coin_now((pol, treasure.ubls[pol][np.argsort(
np.linalg.norm(treasure.ubls[pol], axis=1))[opts.ubl_order]]))
import numpy as np
import commands, os, time, math, ephem
import omnical.calibration_omni as omni
import omnical._omnical as _O
import optparse, sys
import matplotlib.pyplot as plt
import numpy.fft as fft
import scipy.signal as ss
import scipy.linalg as sla
import numpy.linalg as la
Cspeed = .299792458 #m/ns
treasure = omni.Treasure("/home/omniscope/data/PAPER/2015PSA64.treasure")
pol = 'xx'
#for ubl in treasure.ubls[pol]:
#f=treasure.seal_name((pol, ubl))
#if treasure.get_coin((pol, ubl)) is None:
#print ubl,f
#exit()
#data = []
#for ubl in treasure.ubls[pol]:
#c = treasure.get_coin_now((pol, ubl))
#if c is not None:
#plt.plot(abs(fft.fft(c.weighted_mean[10, 20:170])))
#plt.show()
#exit()
#for i, ubl in enumerate(treasure.ubls[pol][np.argsort(la.norm(treasure.ubls[pol],axis=1))[[4,24,44,64,84]]]):
if not os.path.isfile(model_noise_file):
raise IOError("model noise file %s does not exist" %
model_noise_file)
model_noises[pol] = omni.load_omnichisq(model_noise_file)
if np.max(omni.get_omnitime(model_noises[pol])) >= TPI or np.min(
omni.get_omnitime(model_noises[pol])) < 0:
raise ValueError(
"Times stored in noise model %s is outside the range [0, 2pi)."
% model_noise_file)
have_model_noises = True
chisq_leniency = opts.chisq_leniency
have_model_treasure = False
if opts.model_treasure is not None:
if os.path.isdir(os.path.expanduser(opts.model_treasure)):
model_treasure = omni.Treasure(os.path.expanduser(opts.model_treasure))
for pol in opts.pol.split(','):
if pol not in model_treasure.ubls.keys():
raise ValueError(
"Polarization %s not found in the model treasure file %s."
% (pol, os.path.expanduser(opts.model_treasure)))
else:
raise IOError("Model treasure folder not found: %s." %
os.path.expanduser(opts.model_treasure))
have_model_treasure = True
delay_compression = opts.delay_compression
if delay_compression is not None:
if delay_compression % 2 != 1 or delay_compression <= 0:
raise ValueError(
'Delay compression bin number must be odd and positive. Input: %i.'