def test_output_filename_correct(self, filename):
"""
Tests that the output filename is correct.
"""
create_random_data(filename, seconds=1., sample_rate=10.)
self.assertTrue(os.path.exists(filename))
def test_noise(self, filename):
"""
Tests that the noise is added correctly(its mean and std_dev are correct).
"""
seconds = 1.
sample_rate = 1.e6
baseline = 1.
noise_loc = 1.
noise_scale = 0.6
noise = stats.norm(loc=noise_loc, scale=noise_scale)
create_random_data(filename,
seconds=seconds,
sample_rate=sample_rate,
baseline=baseline,
noise=noise)
reader = get_reader_from_filename(filename)
data_all = reader.get_all_data()
data = data_all[0]
mean = np.mean(data)
mean_should_be = baseline + noise_loc
self.assertAlmostEqual(
mean, mean_should_be, 1,
"Unexpected mean. Wanted {0}, got {1}.".format(
mean_should_be, mean))
std_dev = np.std(data)
self.assertAlmostEqual(
noise_scale, std_dev, 1, "Unexpected standard deviation. "
"Wanted {0}, got {1}".format(noise_scale, std_dev))
reader.close()
def test_output_filename_correct(self, filename):
"""
Tests that the output filename is correct.
"""
create_random_data(filename, seconds=1., sample_rate=10.)
self.assertTrue(os.path.exists(filename))
def test_noise(self, filename):
"""
Tests that the noise is added correctly(its mean and std_dev are correct).
"""
seconds = 1.
sample_rate = 1.e6
baseline = 1.
noise_loc = 1.
noise_scale = 0.6
noise = stats.norm(loc=noise_loc, scale=noise_scale)
create_random_data(filename, seconds=seconds, sample_rate=sample_rate, baseline=baseline,
noise=noise)
reader = get_reader_from_filename(filename)
data_all = reader.get_all_data()
data = data_all[0]
mean = np.mean(data)
mean_should_be = baseline + noise_loc
self.assertAlmostEqual(mean, mean_should_be, 1,
"Unexpected mean. Wanted {0}, got {1}.".format(mean_should_be, mean))
std_dev = np.std(data)
self.assertAlmostEqual(noise_scale, std_dev, 1, "Unexpected standard deviation. "
"Wanted {0}, got {1}".format(noise_scale, std_dev))
reader.close()
def test_opening_existing_with_overwrite(self, filename):
"""
Tests that we can overwrite an existing file with overwrite=True
"""
# create a blank file
f = open(filename, mode='w')
f.close()
seconds = 1.
sample_rate = 1.
baseline = 1.
create_random_data(filename, seconds=seconds, sample_rate=sample_rate, baseline=baseline, overwrite=True)
self.assertTrue(os.path.exists(filename))
def _test_number_of_events_help(name):
# Need to seed, otherwise each process will generate the same random data!
np.random.seed()
# Check that setting the event rate gives correct number of events.
seconds = 5.
sample_rate = 1.e6
baseline = 1.
event_rate = 50.
event_duration = stats.norm(loc=100.e-6, scale=5.e-6)
event_depth = stats.norm(loc=-1., scale=.05)
noise = stats.norm(scale=0.02)
# create a list of file_names so we can average the number of events.
if os.path.exists(name):
os.remove(name)
n_e_r = create_random_data(filename=name, seconds=seconds, sample_rate=sample_rate,
baseline=baseline, noise=noise,
event_rate=event_rate, event_durations=event_duration,
event_depths=event_depth)
save_file_name = name[:-len('.h5')] + '_Events.h5'
if os.path.exists(save_file_name):
os.remove(save_file_name)
event_database = find_events([name], save_file_names=[save_file_name])[0]
ed = EventDatabase(event_database)
n_events_found = ed.get_event_count()
os.remove(name)
ed.close()
os.remove(event_database)
return n_e_r, n_events_found, name
def test_opening_existing_with_overwrite(self, filename):
"""
Tests that we can overwrite an existing file with overwrite=True
"""
# create a blank file
f = open(filename, mode='w')
f.close()
seconds = 1.
sample_rate = 1.
baseline = 1.
create_random_data(filename,
seconds=seconds,
sample_rate=sample_rate,
baseline=baseline,
overwrite=True)
self.assertTrue(os.path.exists(filename))
def test_baseline_no_noise_no_events(self, filename):
"""
Tests that with only the baseline, we create the right data.
"""
seconds = 1.
sample_rate = 1.e6
baseline = 1.
event_rate = 0
event_duration = stats.norm(loc=100.e-6, scale=10.e-6)
event_depth = stats.norm(loc=.5, scale=.05)
data_should_be = np.zeros(int(seconds * sample_rate)) + baseline
create_random_data(filename=filename, seconds=seconds, sample_rate=sample_rate, baseline=baseline,
event_rate=event_rate, event_durations=event_duration, event_depths=event_depth)
_test_params_equality(self, filename, data_should_be, sample_rate)
def test_event_params(self, filename):
"""
Tests that setting the event depth and duration give correct values.
"""
seconds = 5.
sample_rate = 1.e6
baseline = 10.
event_rate = 100.
event_duration_loc = 1.e-4
event_duration_scale = 5.e-6
event_duration = stats.norm(loc=event_duration_loc, scale=event_duration_scale)
event_depth_loc = -1.
event_depth_scale = .05
event_depth = stats.norm(loc=-1., scale=.05)
noise = stats.norm(scale=0.01)
n_events_returned = create_random_data(filename, seconds=seconds, sample_rate=sample_rate,
baseline=baseline, event_rate=event_rate,
event_durations=event_duration, event_depths=event_depth,
noise=noise)
save_file_name = filename[:-len('.h5')] + 'Events.h5'
if os.path.exists(save_file_name):
os.remove(save_file_name)
event_database = find_events([filename], save_file_names=[save_file_name])[0]
ed = EventDatabase(event_database)
count = ed.get_event_count()
count_should_be = event_rate * seconds
diff = abs(count - count_should_be)
self.assertLessEqual(diff, 100, "Unexpected number of events. "
"Expected {0}, was {1}.".format(count_should_be, count))
table_sample_rate = ed.get_sample_rate()
durations = [x['event_length'] / table_sample_rate for x in ed.get_event_table().iterrows()]
depths = [x['current_blockage'] for x in ed.get_event_table().iterrows()]
mean_duration = np.mean(durations)
self.assertAlmostEqual(event_duration_loc, mean_duration, 5, "Unexpected mean event duration. "
"Wanted {0}, got {1}.".format(event_duration_loc,
mean_duration))
mean_depth = np.mean(depths)
self.assertAlmostEqual(event_depth_loc, mean_depth, 1, "Unexpected mean event depth. "
"Wanted {0}, got {1}.".format(event_depth_loc,
mean_depth))
ed.close()
os.remove(event_database)
def test_baseline_no_noise_no_events(self, filename):
"""
Tests that with only the baseline, we create the right data.
"""
seconds = 1.
sample_rate = 1.e6
baseline = 1.
event_rate = 0
event_duration = stats.norm(loc=100.e-6, scale=10.e-6)
event_depth = stats.norm(loc=.5, scale=.05)
data_should_be = np.zeros(int(seconds * sample_rate)) + baseline
create_random_data(filename=filename,
seconds=seconds,
sample_rate=sample_rate,
baseline=baseline,
event_rate=event_rate,
event_durations=event_duration,
event_depths=event_depth)
_test_params_equality(self, filename, data_should_be, sample_rate)
def _test_number_of_events_help(name):
# Need to seed, otherwise each process will generate the same random data!
np.random.seed()
# Check that setting the event rate gives correct number of events.
seconds = 5.
sample_rate = 1.e6
baseline = 1.
event_rate = 50.
event_duration = stats.norm(loc=100.e-6, scale=5.e-6)
event_depth = stats.norm(loc=-1., scale=.05)
noise = stats.norm(scale=0.02)
# create a list of file_names so we can average the number of events.
if os.path.exists(name):
os.remove(name)
n_e_r = create_random_data(filename=name,
seconds=seconds,
sample_rate=sample_rate,
baseline=baseline,
noise=noise,
event_rate=event_rate,
event_durations=event_duration,
event_depths=event_depth)
save_file_name = name[:-len('.h5')] + '_Events.h5'
if os.path.exists(save_file_name):
os.remove(save_file_name)
event_database = find_events([name], save_file_names=[save_file_name])[0]
ed = EventDatabase(event_database)
n_events_found = ed.get_event_count()
os.remove(name)
ed.close()
os.remove(event_database)
return n_e_r, n_events_found, name
def test_event_params(self, filename):
"""
Tests that setting the event depth and duration give correct values.
"""
seconds = 5.
sample_rate = 1.e6
baseline = 10.
event_rate = 100.
event_duration_loc = 1.e-4
event_duration_scale = 5.e-6
event_duration = stats.norm(loc=event_duration_loc,
scale=event_duration_scale)
event_depth_loc = -1.
event_depth_scale = .05
event_depth = stats.norm(loc=-1., scale=.05)
noise = stats.norm(scale=0.01)
n_events_returned = create_random_data(filename,
seconds=seconds,
sample_rate=sample_rate,
baseline=baseline,
event_rate=event_rate,
event_durations=event_duration,
event_depths=event_depth,
noise=noise)
save_file_name = filename[:-len('.h5')] + 'Events.h5'
if os.path.exists(save_file_name):
os.remove(save_file_name)
event_database = find_events([filename],
save_file_names=[save_file_name])[0]
ed = EventDatabase(event_database)
count = ed.get_event_count()
count_should_be = event_rate * seconds
diff = abs(count - count_should_be)
self.assertLessEqual(
diff, 100, "Unexpected number of events. "
"Expected {0}, was {1}.".format(count_should_be, count))
table_sample_rate = ed.get_sample_rate()
durations = [
x['event_length'] / table_sample_rate
for x in ed.get_event_table().iterrows()
]
depths = [
x['current_blockage'] for x in ed.get_event_table().iterrows()
]
mean_duration = np.mean(durations)
self.assertAlmostEqual(
event_duration_loc, mean_duration, 5,
"Unexpected mean event duration. "
"Wanted {0}, got {1}.".format(event_duration_loc, mean_duration))
mean_depth = np.mean(depths)
self.assertAlmostEqual(
event_depth_loc, mean_depth, 1, "Unexpected mean event depth. "
"Wanted {0}, got {1}.".format(event_depth_loc, mean_depth))
ed.close()
os.remove(event_database)