MAINT: Report download time and size

[larsoner]

CircleCI has timed out a couple of times lately due to our 3h limit. It costs more $$$ to increase this limit and I think it's pretty reasonable anyway, so let's try to make it faster. 2h is spent in the doc build mostly due to examples/tutorials -- which we should certainly continue to try to optimize -- but recently 45min has been spent in dataset downloading. This PR adds some diagnostic download time/size logger to our fetching that should help us see it in logs, and might be nice for users.

[larsoner]

Looks like this might be one culprit, looking locally after watching CircleCI have a 10-minute timeout trying to download it:

hf_sef_evoked.tar.gz
https://zenodo.org/record/3523071/files/hf_sef_evoked.tar.gz
731 KB/s - 124 MB of 731 MB, 14 mins left

we already have a mirror on osf.io so I'll update the file there and push

[larsoner]

Okay @drammock this one is ready for review. Mirroring data on OSF.io took time from ~45 min down to ~15 min as you can see here (along with all timings for downloads -- I killed the build once that step was done):

https://app.circleci.com/pipelines/github/mne-tools/mne-python/18689/workflows/29272f16-bc51-4ce8-ae40-c7ef59c3795f/jobs/53962

[drammock]

this is a nice improvement. CIs are not happy though. Cirrus fails with

mne/datasets/sleep_physionet/tests/test_physionet.py:132: in test_sleep_physionet_age
    paths = age.fetch_data(subjects=[0], recording=[1], path=physionet_tmpdir)
<decorator-gen-588>:12: in fetch_data
    ???
mne/datasets/sleep_physionet/age.py:129: in fetch_data
    sz += os.path.getsize(psg_fname)
/opt/homebrew/Cellar/python@3.10/3.10.6_2/Frameworks/Python.framework/Versions/3.10/lib/python3.10/genericpath.py:50: in getsize
    return os.stat(filename).st_size
E   FileNotFoundError: [Errno 2] No such file or directory: '/private/var/folders/76/zy5ktkns50v6gt5g8r0sf6sc0000gn/T/pytest-of-admin/pytest-0/physionet_files0/physionet-sleep-data/SC4001E0-PSG.edf'

which looks related / legit; haven't looked at any other failures

[drammock]

there are also

 _______________________ test_manifest_check_download[1] ________________________
mne/datasets/tests/test_datasets.py:225: in test_manifest_check_download
    _manifest_check_download(manifest_path, destination, url, hash_)
mne/datasets/utils.py:570: in _manifest_check_download
    pooch.retrieve(
E   TypeError: _fake_zip_fetch() got an unexpected keyword argument 'downloader'

[drammock]

I think this PR broke circle:

WARNING: /home/circleci/project/tutorials/preprocessing/40_artifact_correction_ica.py failed to execute correctly: Traceback (most recent call last):
  File "/home/circleci/project/tutorials/preprocessing/40_artifact_correction_ica.py", line 539, in <module>
    fname = mne.datasets.eegbci.load_data(subj + 1, runs=[3])[0]
IndexError: list index out of range