deepextractor.data.glitch_data

Utilities for downloading DeepExtractor glitch reconstructions from HuggingFace.

Module Contents

deepextractor.data.glitch_data.REPO_ID = 'tomdooney/deepextractor-glitch-reconstructions'[source]
deepextractor.data.glitch_data.download_glitch_data(data_dir: str | pathlib.Path = 'data/', include_derivatives: bool = False, force: bool = False) → dict[str, pathlib.Path][source]

Download DeepExtractor glitch reconstructions from HuggingFace.

Downloads time-domain reconstructions of seven LIGO O3 glitch classes (35,000 samples total: Blip, Fast Scattering, Koi Fish, Low Frequency Burst, Scattered Light, Tomte, Whistle) from tomdooney/deepextractor-glitch-reconstructions on HuggingFace.

These reconstructions were produced by DeepExtractor and are suitable for training generative models such as GlitchGAN.

Parameters:

  • data_dir – Directory to save downloaded files. Created if it does not exist.

  • include_derivatives – If True, also download the first-order time-derivative array required for cDVGAN training (adds ~2.1 GB).

  • force – If True, re-download files even if they already exist locally.

Returns:

Mapping of {"samples", "labels", "label_order"} (and "derivatives" if requested) to their local file paths.

Return type:

dict[str, Path]

Examples

>>> from deepextractor.data import download_glitch_data
>>> paths = download_glitch_data("data/glitches/")
>>> import numpy as np
>>> X = np.load(paths["samples"])          # (35000, 8192)
>>> y = np.load(paths["labels"])           # (35000, 7)
>>> classes = np.load(paths["label_order"], allow_pickle=True)