def test_compute_metrics():
project_path = get_project_root()
project_path = get_project_root()
refer_path = project_path / 'res_experiments' / 'predictions' / '20170905_030404.fits'
refer, names = open_param_file(refer_path)
pred_path = project_path / 'res_experiments' / 'predictions' / '20170905_030404_common.fits'
save_path = project_path / 'res_experiments' / 'predictions' / 'test.csv'
predicted = fits.open(pred_path)
predicted_data = predicted[0].data
df = compute_metrics(refer, predicted_data, names, save_path)
# df.to_csv(save_path)
assert df.shape == (11, 3)
def test_compute_mean_spectrum():
filename = get_project_root() / 'data' / 'parameters_base.fits'
batch_size = 1000
nbatches = 2
MEAN, STD = compute_mean_spectrum(filename,
batch_size=batch_size,
nbatches=nbatches)
def sample_from_database(self):
# filename = '/Users/irinaknyazeva/Projects/Solar/InverseProblem/data/parameters_base.fits'
project_path = get_project_root()
filename = project_path / 'data' / 'small_parameters_base.fits'
source = 'database'
sobj = SpectrumDataset(param_path=filename, source=source)
sample = sobj[0]
return sample
def test_init_refer_dataset(self):
project_path = get_project_root()
filename = project_path / 'data' / 'hinode_source' / '20140926_170005.fits'
source = 'refer'
sobj = SpectrumDataset(param_path=filename, source=source)
assert isinstance(sobj.param_source, list)
assert isinstance(sobj[0]['X'][1], float)
assert sobj.__len__() == 446976
assert 224 == sobj[0]['X'][0].size
assert 11 == sobj[0]['Y'].size
def test_init_dataset(self):
project_path = get_project_root()
filename = project_path / 'data' / 'small_parameters_base.fits'
param_array = fits.open(filename)[0].data[:10]
transform = mlp_batch_rescale()
# transform=None
sobj = PregenSpectrumDataset(data_arr=param_array, transform=transform)
sample = sobj[0]
assert sample['X'][1].shape[0] == 1
assert sample['Y'].shape[0] == 11
assert True
def test_init_database_dataset(self):
project_path = get_project_root()
# filename = project_path / 'data' / 'parameters_base_new.fits'
filename = project_path / 'data' / 'small_parameters_base.fits'
source = 'database'
sobj = SpectrumDataset(param_path=filename, source=source)
assert sobj.param_source.shape[1] == 11
assert isinstance(sobj[0]['X'][1], float)
assert isinstance(sobj[0]['X'][0], np.ndarray)
assert isinstance(sobj[0]['Y'], np.ndarray)
assert sobj.__len__() == sobj.param_source.shape[0]
assert 224 == sobj[0]['X'][0].size
assert 11 == sobj[0]['Y'].size
def test_create_small_dataset():
filename = get_project_root() / 'data' / 'parameters_base.fits'
savename = get_project_root() / 'data' / 'small_parameters_base.fits'
create_small_dataset(filename, savename, size=10000)
params = fits.open(savename)[0].data
assert params.shape[0] == 10000
def test_download_from_google_disc():
file_id = '19jkSXHxAPWZvfgo5oxSmvEagme5YLY33'
dest_path = str(get_project_root() / 'data' / 'small_parameters_base.fits')
download_from_google_disc(fileid=file_id, dest=dest_path)
parameter_base = fits.open(dest_path)[0].data
assert type(parameter_base) == np.ndarray
def test_get_project_root():
root_path = get_project_root()
root_path1 = Path(__file__).parent.parent
assert root_path1 == root_path
def test_init_dataset(self):
project_path = get_project_root()
filename = project_path / 'data' / 'small_parameters_base.fits'
param_array = fits.open(filename)[0].data[:10]
sobj = SpectrumDataset(data_arr=param_array)
assert True
def test_open_param_file():
project_path = get_project_root()
real_path = project_path / 'res_experiments' / 'predictions' / '20170905_030404.fits'
refer, names = open_param_file(real_path)
assert refer.shape == (512, 485, 11)
assert len(names) == 11