pardata.dataset.Dataset.download
- Dataset.download(*, check=True, verify_checksum=True)
Downloads, extracts, and removes dataset archive. It adds a directory write lock during execution. If
download_url
in the schema is a file path, then this method only extracts.- Parameters
check (bool) – Check to make sure the data files are not already present in
_data_dir
(passed in viadata_dir
in the constructorDataset
) by runningis_downloaded()
. If set toTrue
, raise an error if they are present and prevent a subsequent download. Set toFalse
to remove this safeguard, and subsequent calls todownload()
will then overwrite data files if they were previously downloaded to_data_dir
.verify_checksum (bool) – If
True
, verify sha512sum of the downloaded dataset.
- Raises
RuntimeError –
check
isTrue
and the dataset was previously downloaded as indicated byis_downloaded()
returningTrue
.NotADirectoryError –
Dataset._data_dir
(passed in viadata_dir
in the constructorDataset
) points to an existing file that is not a directory.OSError –
verify_checksum
isTrue
and the SHA512 checksum of a downloaded dataset doesn’t match the expected checksum.RuntimeError – The archive could not be extracted.
exceptions.DirectoryLockAcquisitionError – Failed to acquire the directory lock.
- Return type