def setUp(self) -> None:
np.random.seed(0)
self.objects = pd.DataFrame(
index=[
'ZTF18abakgtm',
'ZTF18abvvcko',
'ZTF17aaaaaxg',
'ZTF18aaveorp'
],
columns=['corrected'],
data=np.random.choice(
a=[False, True],
size=(4,)
)
)
preprocess_ztf = DetectionsPreprocessorZTF()
raw_det_ZTF18abakgtm = pd.read_csv(
os.path.join(EXAMPLES_PATH, 'ZTF18abakgtm_det.csv'), index_col="oid")
raw_det_ZTF18abakgtm['sigmapsf_corr_ext'] = raw_det_ZTF18abakgtm['sigmapsf_corr']
raw_det_ZTF18abvvcko = pd.read_csv(
os.path.join(EXAMPLES_PATH, 'ZTF18abvvcko_det.csv'), index_col="oid")
raw_det_ZTF18abvvcko['sigmapsf_corr_ext'] = raw_det_ZTF18abvvcko['sigmapsf_corr']
raw_det_ZTF17aaaaaxg = pd.read_csv(
os.path.join(EXAMPLES_PATH, 'ZTF17aaaaaxg_det.csv'), index_col="oid")
raw_det_ZTF17aaaaaxg['sigmapsf_corr_ext'] = raw_det_ZTF17aaaaaxg['sigmapsf_corr']
raw_det_ZTF18aaveorp = pd.read_csv(
os.path.join(EXAMPLES_PATH, 'ZTF18aaveorp_det.csv'), index_col="oid")
raw_det_ZTF18aaveorp['sigmapsf_corr_ext'] = raw_det_ZTF18aaveorp['sigmapsf_corr']
keys = [
'mjd',
'fid',
'magpsf_corr',
'sigmapsf_corr_ext',
'isdiffpos',
'magpsf',
'sigmapsf',
'ra',
'dec',
'sgscore1',
'rb'
]
self.detections = pd.concat(
[raw_det_ZTF17aaaaaxg[keys],
raw_det_ZTF18abvvcko[keys],
raw_det_ZTF18abakgtm[keys],
raw_det_ZTF18aaveorp[keys]],
axis=0
)
self.detections = preprocess_ztf.get_magpsf_ml(
self.detections,
objects=self.objects
)
self.detections = self.detections[['fid', 'magpsf_ml']]
def test_non_numeric_columns(self):
# print(self.detections.head())
# print(self.detections.dtypes)
with self.assertRaises(TypeError):
_ = self.detections.sigmapsf_corr > 0.0
preprocessor = DetectionsPreprocessorZTF()
preprocessed_detections = preprocessor.preprocess_detections(
self.detections)
def setUp(self) -> None:
self.preprocessor_ztf = DetectionsPreprocessorZTF()
self.raw_data_ZTF18abakgtm = pd.read_csv(
os.path.join(EXAMPLES_PATH, 'ZTF18abakgtm_det.csv'), index_col="oid")
self.raw_data_ZTF18abakgtm['sigmapsf_corr_ext'] = self.raw_data_ZTF18abakgtm['sigmapsf_corr']
self.fake_objects = pd.DataFrame(
columns=['corrected'],
index=['ZTF18abakgtm'],
data=[[True]]
)
def setUp(self) -> None:
self.preprocess_ztf = DetectionsPreprocessorZTF()
self.fake_objects = pd.DataFrame(
index=['ZTF17aaaaaxg', 'ZTF18abvvcko', 'ZTF18abakgtm'],
data=[[True], [False], [True]],
columns=['corrected'])
self.raw_det_ZTF18abakgtm = pd.read_csv(os.path.join(
EXAMPLES_PATH, 'ZTF18abakgtm_det.csv'),
index_col="oid")
self.raw_det_ZTF18abakgtm[
'sigmapsf_corr_ext'] = self.raw_det_ZTF18abakgtm['sigmapsf_corr']
self.raw_nondet_ZTF18abakgtm = pd.read_csv(os.path.join(
EXAMPLES_PATH, 'ZTF18abakgtm_nondet.csv'),
index_col="oid")
self.det_ZTF18abakgtm = self.preprocess_ztf.preprocess(
self.raw_det_ZTF18abakgtm, objects=self.fake_objects)
self.raw_det_ZTF18abvvcko = pd.read_csv(os.path.join(
EXAMPLES_PATH, 'ZTF18abvvcko_det.csv'),
index_col="oid")
self.raw_det_ZTF18abvvcko[
'sigmapsf_corr_ext'] = self.raw_det_ZTF18abvvcko['sigmapsf_corr']
self.raw_nondet_ZTF18abvvcko = pd.read_csv(os.path.join(
EXAMPLES_PATH, 'ZTF18abvvcko_nondet.csv'),
index_col="oid")
self.det_ZTF18abvvcko = self.preprocess_ztf.preprocess(
self.raw_det_ZTF18abvvcko, objects=self.fake_objects)
self.raw_det_ZTF17aaaaaxg = pd.read_csv(os.path.join(
EXAMPLES_PATH, 'ZTF17aaaaaxg_det.csv'),
index_col="oid")
self.raw_det_ZTF17aaaaaxg[
'sigmapsf_corr_ext'] = self.raw_det_ZTF17aaaaaxg['sigmapsf_corr']
self.raw_nondet_ZTF17aaaaaxg = pd.read_csv(os.path.join(
EXAMPLES_PATH, 'ZTF17aaaaaxg_nondet.csv'),
index_col="oid")
self.det_ZTF17aaaaaxg = self.preprocess_ztf.preprocess(
self.raw_det_ZTF17aaaaaxg, objects=self.fake_objects)
def setUp(self) -> None:
np.random.seed(0)
self.detections = pd.read_pickle(
os.path.join(EXAMPLES_PATH, "periodic_light_curves.pkl"))
self.labels = pd.read_pickle(
os.path.join(EXAMPLES_PATH, "periodic_light_curve_labels.pkl"))
self.detections["sigmapsf_corr_ext"] = self.detections["sigmapsf_corr"]
self.fake_objects = pd.DataFrame(
index=self.labels.index,
columns=["corrected"],
data=np.random.choice(a=[False, True], size=(len(self.labels), )),
)
self.detections = DetectionsPreprocessorZTF().get_magpsf_ml(
self.detections, objects=self.fake_objects)
class TestObjectsMethods(unittest.TestCase):
def setUp(self) -> None:
self.preprocessor_ztf = DetectionsPreprocessorZTF()
self.raw_data_ZTF18abakgtm = pd.read_csv(
os.path.join(EXAMPLES_PATH, 'ZTF18abakgtm_det.csv'), index_col="oid")
self.raw_data_ZTF18abakgtm['sigmapsf_corr_ext'] = self.raw_data_ZTF18abakgtm['sigmapsf_corr']
self.fake_objects = pd.DataFrame(
columns=['corrected'],
index=['ZTF18abakgtm'],
data=[[True]]
)
def test_preprocess_one_object(self):
self.assertEqual(self.preprocessor_ztf.has_necessary_columns(
self.preprocessor_ztf.get_magpsf_ml(
self.raw_data_ZTF18abakgtm,
self.fake_objects
)),
True)
self.assertEqual(self.raw_data_ZTF18abakgtm.index.name, "oid")
preprocessed_data = self.preprocessor_ztf.preprocess(
self.raw_data_ZTF18abakgtm, objects=self.fake_objects)
self.assertEqual(type(preprocessed_data), pd.DataFrame)
def setUp(self) -> None:
self.detections = pd.read_pickle(
os.path.join(EXAMPLES_PATH, 'periodic_light_curves.pkl'))
self.labels = pd.read_pickle(
os.path.join(EXAMPLES_PATH, 'periodic_light_curve_labels.pkl'))
self.detections['sigmapsf_corr_ext'] = self.detections['sigmapsf_corr']
self.fake_objects = pd.DataFrame(index=self.labels.index,
columns=['corrected'],
data=np.random.choice(
a=[False, True],
size=(len(self.labels), )))
self.detections = DetectionsPreprocessorZTF().get_magpsf_ml(
self.detections, objects=self.fake_objects)
self.logger = logging.getLogger()
self.logger.setLevel(logging.INFO)
def setUp(self) -> None:
np.random.seed(0)
self.objects = pd.DataFrame(
index=[
"ZTF18abakgtm", "ZTF18abvvcko", "ZTF17aaaaaxg", "ZTF18aaveorp"
],
columns=["corrected"],
data=np.random.choice(a=[False, True], size=(4, )),
)
preprocess_ztf = DetectionsPreprocessorZTF()
raw_det_ZTF18abakgtm = pd.read_csv(os.path.join(
EXAMPLES_PATH, "ZTF18abakgtm_det.csv"),
index_col="oid")
raw_det_ZTF18abakgtm["sigmapsf_corr_ext"] = raw_det_ZTF18abakgtm[
"sigmapsf_corr"]
raw_det_ZTF18abvvcko = pd.read_csv(os.path.join(
EXAMPLES_PATH, "ZTF18abvvcko_det.csv"),
index_col="oid")
raw_det_ZTF18abvvcko["sigmapsf_corr_ext"] = raw_det_ZTF18abvvcko[
"sigmapsf_corr"]
raw_det_ZTF17aaaaaxg = pd.read_csv(os.path.join(
EXAMPLES_PATH, "ZTF17aaaaaxg_det.csv"),
index_col="oid")
raw_det_ZTF17aaaaaxg["sigmapsf_corr_ext"] = raw_det_ZTF17aaaaaxg[
"sigmapsf_corr"]
raw_det_ZTF18aaveorp = pd.read_csv(os.path.join(
EXAMPLES_PATH, "ZTF18aaveorp_det.csv"),
index_col="oid")
raw_det_ZTF18aaveorp["sigmapsf_corr_ext"] = raw_det_ZTF18aaveorp[
"sigmapsf_corr"]
keys = [
"mjd",
"fid",
"magpsf_corr",
"sigmapsf_corr_ext",
"isdiffpos",
"magpsf",
"sigmapsf",
"ra",
"dec",
"sgscore1",
"rb",
]
self.detections = pd.concat(
[
raw_det_ZTF17aaaaaxg[keys],
raw_det_ZTF18abvvcko[keys],
raw_det_ZTF18abakgtm[keys],
raw_det_ZTF18aaveorp[keys],
],
axis=0,
)
self.detections = preprocess_ztf.get_magpsf_ml(self.detections,
objects=self.objects)
self.detections = self.detections[["sgscore1"]]
class TestObjectsMethods(unittest.TestCase):
def setUp(self) -> None:
self.preprocess_ztf = DetectionsPreprocessorZTF()
self.fake_objects = pd.DataFrame(
index=['ZTF17aaaaaxg', 'ZTF18abvvcko', 'ZTF18abakgtm'],
data=[[True], [False], [True]],
columns=['corrected'])
self.raw_det_ZTF18abakgtm = pd.read_csv(os.path.join(
EXAMPLES_PATH, 'ZTF18abakgtm_det.csv'),
index_col="oid")
self.raw_det_ZTF18abakgtm[
'sigmapsf_corr_ext'] = self.raw_det_ZTF18abakgtm['sigmapsf_corr']
self.raw_nondet_ZTF18abakgtm = pd.read_csv(os.path.join(
EXAMPLES_PATH, 'ZTF18abakgtm_nondet.csv'),
index_col="oid")
self.det_ZTF18abakgtm = self.preprocess_ztf.preprocess(
self.raw_det_ZTF18abakgtm, objects=self.fake_objects)
self.raw_det_ZTF18abvvcko = pd.read_csv(os.path.join(
EXAMPLES_PATH, 'ZTF18abvvcko_det.csv'),
index_col="oid")
self.raw_det_ZTF18abvvcko[
'sigmapsf_corr_ext'] = self.raw_det_ZTF18abvvcko['sigmapsf_corr']
self.raw_nondet_ZTF18abvvcko = pd.read_csv(os.path.join(
EXAMPLES_PATH, 'ZTF18abvvcko_nondet.csv'),
index_col="oid")
self.det_ZTF18abvvcko = self.preprocess_ztf.preprocess(
self.raw_det_ZTF18abvvcko, objects=self.fake_objects)
self.raw_det_ZTF17aaaaaxg = pd.read_csv(os.path.join(
EXAMPLES_PATH, 'ZTF17aaaaaxg_det.csv'),
index_col="oid")
self.raw_det_ZTF17aaaaaxg[
'sigmapsf_corr_ext'] = self.raw_det_ZTF17aaaaaxg['sigmapsf_corr']
self.raw_nondet_ZTF17aaaaaxg = pd.read_csv(os.path.join(
EXAMPLES_PATH, 'ZTF17aaaaaxg_nondet.csv'),
index_col="oid")
self.det_ZTF17aaaaaxg = self.preprocess_ztf.preprocess(
self.raw_det_ZTF17aaaaaxg, objects=self.fake_objects)
def turbofats_features(self):
turbofats_extractor = TurboFatsFeatureExtractor()
turbofats_fs = turbofats_extractor.compute_features(
self.det_ZTF18abakgtm)
expected_cols = [
'Amplitude', 'AndersonDarling', 'Autocor_length', 'Beyond1Std',
'Con', 'Eta_e', 'Gskew', 'MaxSlope', 'Mean', 'Meanvariance',
'MedianAbsDev', 'MedianBRP', 'PairSlopeTrend', 'PercentAmplitude',
'Q31', 'Rcs', 'Skew', 'SmallKurtosis', 'Std', 'StetsonK',
'Harmonics_mag_1', 'Harmonics_mag_2', 'Harmonics_mag_3',
'Harmonics_mag_4', 'Harmonics_mag_5', 'Harmonics_mag_6',
'Harmonics_mag_7', 'Harmonics_phase_2', 'Harmonics_phase_3',
'Harmonics_phase_4', 'Harmonics_phase_5', 'Harmonics_phase_6',
'Harmonics_phase_7', 'Harmonics_mse', 'Pvar', 'ExcessVar',
'GP_DRW_sigma', 'GP_DRW_tau', 'SF_ML_amplitude', 'SF_ML_gamma',
'IAR_phi', 'LinearTrend'
]
self.assertEqual(type(turbofats_fs), pd.DataFrame)
self.assertEqual(len(turbofats_fs), len(expected_cols))
self.assertListEqual(turbofats_fs.columns, expected_cols)
def test_color_features(self):
color_extractor = ColorFeatureExtractor()
color_fs = color_extractor.compute_features(self.det_ZTF17aaaaaxg)
self.assertEqual(type(color_fs), pd.DataFrame)
self.assertEqual(len(color_fs.columns),
len(color_extractor.get_features_keys()))
self.assertListEqual(list(color_fs.columns),
color_extractor.get_features_keys())
def test_galactic_coordinates_features(self):
galactic_extractor = GalacticCoordinatesExtractor()
galactic_fs = galactic_extractor.compute_features(
self.det_ZTF18abakgtm)
self.assertEqual(type(galactic_fs), pd.DataFrame)
self.assertEqual(len(galactic_fs.columns), 2)
self.assertListEqual(list(galactic_fs.columns),
galactic_extractor.get_features_keys())
def test_real_bogus_features(self):
rb_extractor = RealBogusExtractor()
rb_fs = rb_extractor.compute_features(self.det_ZTF18abakgtm)
self.assertEqual(type(rb_fs), pd.DataFrame)
def test_sg_score_features(self):
sg_score = SGScoreExtractor()
sg_fs = sg_score.compute_features(self.det_ZTF18abakgtm)
self.assertEqual(type(sg_fs), pd.DataFrame)
def test_sn_detections_features(self):
sn_det_extractor = SupernovaeDetectionFeatureExtractor()
sn_fs = sn_det_extractor.compute_features(self.det_ZTF18abakgtm)
self.assertEqual(type(sn_fs), pd.DataFrame)
def test_sn_non_detections_features(self):
sn_non_det_extractor = SupernovaeDetectionAndNonDetectionFeatureExtractor(
)
sn_non_det = sn_non_det_extractor.compute_features(
self.det_ZTF18abakgtm, non_detections=self.raw_nondet_ZTF18abakgtm)
self.assertEqual(type(sn_non_det), pd.DataFrame)
def test_wise_features_ok(self):
wise_extractor = WiseStaticExtractor()
wise_features = wise_extractor.compute_features(
self.det_ZTF17aaaaaxg, non_detections=self.raw_nondet_ZTF17aaaaaxg)
self.assertEqual(type(wise_features), pd.DataFrame)
self.assertEqual(len(wise_features.dropna()), 1)
def test_wise_features_not_found(self):
wise_extractor = WiseStaticExtractor()
wise_features = wise_extractor.compute_features(
self.det_ZTF18abakgtm, non_detections=self.raw_nondet_ZTF18abakgtm)
self.assertEqual(type(wise_features), pd.DataFrame)
self.assertEqual(len(wise_features.dropna()), 0)
unused_classes = ['TDE', 'ZZ']
rename_class_dictionary = {
'EA': 'EB',
'EB/EW': 'EB',
'RSCVn': 'Periodic-Other',
'SNIIb': 'SNII',
'SNIIn': 'SNII'
}
labels = labels[~labels.classALeRCE.isin(unused_classes)].copy()
labels['classALeRCE'] = labels['classALeRCE'].map(
rename_class_dictionary).fillna(labels['classALeRCE'])
# Intersecting labels and detections
valid_oids = detections.index.unique().intersection(labels.index.unique())
labeled_detections = detections.loc[valid_oids]
labels = labels.loc[valid_oids].copy()
valid_oids = valid_oids.intersection(non_detections.index.unique())
labeled_non_detections = non_detections.loc[valid_oids]
# ZTF preprocessing
preprocessor_ztf = DetectionsPreprocessorZTF()
labeled_detections = preprocessor_ztf.preprocess(labeled_detections)
# Save data
labeled_detections.to_pickle('dataset_detections.pkl')
labeled_non_detections.to_pickle('dataset_non_detections.pkl')
labels.to_pickle('dataset_labels.pkl')
import sys
import pandas as pd
from late_classifier.features import DetectionsPreprocessorZTF
detections = pd.read_pickle(sys.argv[1])
non_detections = pd.read_pickle(sys.argv[2])
preprocessor_ztf = DetectionsPreprocessorZTF()
detections = preprocessor_ztf.preprocess(detections)
oid_intersection = detections.index.unique().intersection(
non_detections.index.unique())
non_detections = non_detections.loc[oid_intersection]
detections.to_pickle(sys.argv[3])
non_detections.to_pickle(sys.argv[4])
