Inconsistent environment variable name for Sleep Physionet dataset

The docstring of mne.datasets.sleep_physionet.age.fetch_data states the environment variable pointing to the Sleep Physionet dataset should be “MNE_DATASETS_PHYSIONET_SLEEP_PATH”: https://github.com/mne-tools/mne-python/blob/855edbf48835bcee80e773eb7047d0a4caa14a1c/mne/datasets/sleep_physionet/age.py#L47

However the code looks instead for “PHYSIONET_SLEEP_PATH”: https://github.com/mne-tools/mne-python/blob/855edbf48835bcee80e773eb7047d0a4caa14a1c/mne/datasets/sleep_physionet/_utils.py#L77

Given that other datasets, e.g. eegbci use the format “MNE_DATASETS_…_PATH”: https://github.com/mne-tools/mne-python/blob/855edbf48835bcee80e773eb7047d0a4caa14a1c/mne/datasets/eegbci/eegbci.py#L62 I suppose the code should be changed to what the docstring states?

I just wanted to double check that is correct before submitting a PR. Thanks!

[agramfort]

@agramfort I just tested it and actually the mne-config.json file was not updated when I fetched data and provided a specific path. When I called it again without a path the data was downloaded again and saved under ~/mne_data. Does that change what you’d recommend doing?

Yes I had noticed this in the past. The fetcher doesn’t behave like others.

So maybe we will not break backward compat?

@sappelhoff So that would entail making a more complete version of the existing _data_path function? https://github.com/mne-tools/mne-python/blob/855edbf48835bcee80e773eb7047d0a4caa14a1c/mne/datasets/sleep_physionet/_utils.py#L44

Yes but I fear it’s quiet some work. Not sure how high priority this is…

[drammock]

closed by #9657