def corr(args):
if args.alg == 0:
c = PearsonCorrelation(args.h5database)
corr_matrix = c.find_correlations()
if args.out is not None:
with open(args.out, 'wb') as f:
pickle.dump(corr_matrix, f)
elif args.alg == 1:
c = FourierApproximation(args.h5database)
corr_matrix = c.find_correlations(args.k, args.T, args.B, args.e)
if args.out is not None:
with open(args.out, 'wb') as f:
pickle.dump(corr_matrix, f)
elif args.alg == 2:
c = BooleanCorrelation(args.h5database, args.validate)
boolean_corr_matrix = c.boolean_approximation(args.T)
if args.out is not None:
with open(args.out, 'wb') as f:
pickle.dump(boolean_corr_matrix, f)
def __init__(self, t_dataset_path: str, validation=False):
"""
:param t_dataset_path: original dataset path
"""
self.norm_ds_path = t_dataset_path
self.norm_ds = DatasetH5(t_dataset_path)
self.UB = np.full(shape=(len(self.norm_ds), len(self.norm_ds)),
fill_value=sys.maxsize,
dtype="float32",
order="C")
self.LB = np.zeros(shape=(len(self.norm_ds), len(self.norm_ds)),
dtype="float32",
order="C")
self.CB = np.zeros(shape=(len(self.norm_ds), len(self.norm_ds)),
dtype="b1",
order="C")
self.cache = [None] * len(self.norm_ds)
self.logger = logging.getLogger("Correlation2")
if validation:
self.c = PearsonCorrelation(self.norm_ds_path)
self.validation = validation
