coniferest.datasets package

coniferest.datasets.non_anomalous_outliers(inliers=1024, outliers=32, regions=None, seed=0)[source]
coniferest.datasets.plasticc_gp(**kwargs)[source]

Load PLAsTiCC dataset of GP-approximated light curves.

It is a subset of the PLAsTiCC dataset, light curves are approximated with vector Gaussian processes (Kornilov et al., 2023). Features are these GP approximations and GP parameters, all normalized.

The dataset is hosted on Hugging Face, `datasets’ package is required to load it.

Adopted from Ishida et al. 2021: https://ui.adsabs.harvard.edu/abs/2021A%26A…650A.195I/abstract

Parameters:

**kwargs – Arbitrary keyword arguments passed to datasets.load_dataset.

Returns:

  • data (2-D numpy.ndarray of float32) – 2-D array of features.

  • metadata (1-D numpy.ndarray of bool) – 1-D array of anomaly labels, True for anomalies.

coniferest.datasets.single_outlier(inliers=10000, rng=0)[source]
coniferest.datasets.ztf_m31(**kwargs)[source]

Load ZTF DR3 M31 light curve feature dataset.

The dataset is hosted on Hugging Face, `datasets’ package is required to load it.

Adopted from Malanchev et al. 2021: https://ui.adsabs.harvard.edu/abs/2021MNRAS.502.5147M/abstract https://zenodo.org/record/4318700

Parameters:

**kwargs – Arbitrary keyword arguments passed to datasets.load_dataset.

Returns:

  • data (2-D numpy.ndarray of float32) – 2-D array of light curve features.

  • metadata (1-D numpy.ndarray of uint64) – ZTF DR object IDs (OIDs) of the objects.