Sample data set cannot be downloaded if data folder does not exist

The default location for MNE sample data is ~/mne_data. Apparently, if this folder does not exist the data set cannot be downloaded:

# make sure the folder "~/mne_data" does not exist!
import mne
from mne.datasets import sample


data_path = sample.data_path()
raw_fname = data_path + '/MEG/sample/sample_audvis_filt-0-40_raw.fif'

raw = mne.io.read_raw_fif(raw_fname)

I get the following error:

Downloading archive MNE-sample-data-processed.tar.gz to /Users/clemens/mne_data
Downloading https://files.osf.io/v1/resources/rxvq7/providers/osfstorage/59c0e26f9ad5a1025c4ab159?version=5&action=download&direct (1.54 GB)
  0%|          | Downloading : 0.00/1.54G [00:00<?,        ?B/s]Error while fetching file https://osf.io/86qa2/download?version=5. Dataset fetching aborted.
Traceback (most recent call last):
  File "/usr/local/lib/python3.9/site-packages/IPython/core/interactiveshell.py", line 3418, in run_code
    exec(code_obj, self.user_global_ns, self.user_ns)
  File "<ipython-input-2-9a4a5c6c7639>", line 1, in <module>
    runfile('/Users/clemens/Library/Application Support/JetBrains/PyCharm2020.3/scratches/scratch.py', wdir='/Users/clemens/Library/Application Support/JetBrains/PyCharm2020.3/scratches')
  File "/Applications/PyCharm.app/Contents/plugins/python/helpers/pydev/_pydev_bundle/pydev_umd.py", line 197, in runfile
    pydev_imports.execfile(filename, global_vars, local_vars)  # execute the script
  File "/Applications/PyCharm.app/Contents/plugins/python/helpers/pydev/_pydev_imps/_pydev_execfile.py", line 18, in execfile
    exec(compile(contents+"\n", file, 'exec'), glob, loc)
  File "/Users/clemens/Library/Application Support/JetBrains/PyCharm2020.3/scratches/scratch.py", line 5, in <module>
    data_path = sample.data_path()
  File "<decorator-gen-461>", line 24, in data_path
  File "/Users/clemens/Repositories/mne-python/mne/datasets/sample/sample.py", line 19, in data_path
    return _data_path(path=path, force_update=force_update,
  File "/Users/clemens/Repositories/mne-python/mne/datasets/utils.py", line 418, in _data_path
    remove_archive, full = _download(path, u, an, h)
  File "/Users/clemens/Repositories/mne-python/mne/datasets/utils.py", line 475, in _download
    _fetch_file(url, full_name, print_destination=False,
  File "<decorator-gen-3>", line 24, in _fetch_file
  File "/Users/clemens/Repositories/mne-python/mne/utils/fetching.py", line 117, in _fetch_file
    _get_http(url, temp_file_name, initial_size, timeout, verbose_bool)
  File "/Users/clemens/Repositories/mne-python/mne/utils/fetching.py", line 55, in _get_http
    with open(temp_file_name, mode) as local_file:
FileNotFoundError: [Errno 2] No such file or directory: '/Users/clemens/mne_data/MNE-sample-data-processed.tar.gz.part'

Everything works as expected if the folder ~/mne_data exists (even if it is just an empty folder). We should automatically create this folder if not present to prevent this error.

[mmagnuski]

I think that given we don’t know what caused the config directory to be unavailable it might be safer to throw a very informative error and ask the user to check what is wrong (and recreate the directory if that is what they want). This seems to be safer and least overkill (on mne side) to me actually.

[cbrnr]

BTW, using e.g. eegbci.load_data(1, 4) instead of the sample data set works as I expected - it automatically creates the folder even if the folder specified in the config doesn’t exist. I think that’s how it should be for all datasets. So now I can argue that we are already inconsistent, but probably the solution is to add load_data methods to all datasets?