sdss_dr7_dataset#

Module Contents#

Classes#

SDSSDR7Dataset

Dataset of SDSS DR7 light curves with metadata

Functions#

load_sdss_dr7_catalogs(bandpasses)

clean_shen_et_al_2008_catalog(in_path, bandpasses[, ...])

Clean the catalog in

Attributes#

train_dataset

sdss_dr7_dataset.load_sdss_dr7_catalogs(bandpasses)[source]#
sdss_dr7_dataset.clean_shen_et_al_2008_catalog(in_path, bandpasses, out_path=None)[source]#

Clean the catalog in http://faculty.washington.edu/ivezic/macleod/qso_dr7/Southern_format_DB.html

class sdss_dr7_dataset.SDSSDR7Dataset(agn_params, bp_params, bandpasses, out_dir, num_samples, rescale_x=1.0 / 3339 * 0.5 * 4.0, shift_x=-3339 * 0.5, metadata_kwargs=dict(keep_agn_mode='max_obs'), light_curve_kwargs=dict())[source]#

Bases: torch.utils.data.Dataset

Dataset of SDSS DR7 light curves with metadata

max_xfloat

Maximum observer-frame time in days

delta_xfloat

Resolution of DRW rendering in observer-frame days

bp_to_int[source]#
process()[source]#

Process the metadata and light curves

_process_metadata(keep_agn_mode)[source]#

Process the metadata storing the target labels

_load_metadata()[source]#

Read in the DRW and BH metadata

_apply_selection()[source]#

Apply the MacLeod et al 2010 selection

_set_keep_agn(keep_agn_mode)[source]#

Set which AGN IDs to keep

_save_metadata()[source]#
cadence()[source]#
_process_light_curves()[source]#

Process the light curves, which are the NN input

__getitem__(index)[source]#
get_normalizing_metadata(dataloader)[source]#

Compute and set normalizing mean and std for the params

data_loader: iterator

Loader for the params (not necessarily the same as loader for training set)

__len__()[source]#
sdss_dr7_dataset.train_dataset[source]#