def test_required_cols3():
"""extracts the required columns from a test file 3"""
t = [{
'stacks': {
'metrics': [
'bh_photo_z_validation.kstest',
'bh_photo_z_validation.npoisson',
'bh_photo_z_validation.log_loss'
],
'metric_bins': [{
'MAG_DETMODEL_I': '[ 17.5, 19, 22, 25]'
}],
'weights':
'WEIGHTS_LSS',
'truth_bins': [{
'Z_SPEC': 'numpy.linspace(0, 2, 4)'
}],
'tolerance': [0.7, 20],
'truths':
'Z_SPEC'
}
}]
cols = pval.required_cols(t, 'pdf')
corr_cols = ['COADD_OBJECTS_ID', 'Z_SPEC', 'MAG_DETMODEL_I', 'WEIGHTS_LSS']
print cols, corr_cols
for c in corr_cols:
np.testing.assert_equal(c in cols, True)
def test_required_cols1():
"""extracts the required columns from a test file 1"""
t = [{
'metrics': ['bh_photo_z_validation.sigma_68'],
'truths': 'Z_SPEC',
'weights': 'WEIGHTS',
'predictions': ['Z_MC'],
'bins': [{
'Z_MC': '[0, 0.5, 1.0, 2.0]'
}]
}]
cols = pval.required_cols(t, 'point')
corr_cols = ['COADD_OBJECTS_ID', 'Z_SPEC', 'Z_MC', 'WEIGHTS']
print cols, corr_cols
for c in corr_cols:
np.testing.assert_equal(c in cols, True)
def test_required_cols2():
"""extracts the required columns from a test file 2"""
t = [{
'individual': {
'metrics': ['bh_photo_z_validation.eval_pdf_point'],
'truths': 'Z_SPEC',
'tolerance': [0.7, 20],
'weights': 'WEIGHTS',
'bins': [{
'MAG_DETMODEL_I': '[ 17.5, 19, 22, 25]'
}]
}
}]
cols = pval.required_cols(t, 'pdf')
corr_cols = ['COADD_OBJECTS_ID', 'Z_SPEC', 'MAG_DETMODEL_I', 'WEIGHTS']
print cols, corr_cols
for c in corr_cols:
np.testing.assert_equal(c in cols, True)
#First point predictions
ptype = 'point'
#do we have any files of this type to work with?
if len(files[ptype]) > 0:
#results dictionary
res[ptype] = {}
#obtain the tests and required cols
tests = testProperties[ptype]
#check these test are "valid"
cont = pval.valid_tests(tests)
reqcols = pval.required_cols(tests, ptype)
#loop over all files
for f in files[ptype]:
#load a file, and complain if it's not formatted correctly.
d = load_file(f, reqcols)
res[ptype][f] = {}
#calculate all unweighted metrics for deltaz and deltaz/(1+z)
for tst in tests:
if test_name is None:
test_name = 'Test_randid' + str(np.random.randint(0, 1000))
if pval.key_not_none(tst, 'test_name'):