Fix TQDM usage

[danshapero]

Newer versions of tqdm behave differently at a terminal vs in a jupyter notebook. Import from tqdm.auto instead so that the downloader looks right in either a notebook or the terminal.

[billbrod]

I would also like to see this fixed.

A hacky workaround is to do:

from tqdm.auto import tqdm
import sys

use_ascii = bool(sys.platform == "win32")
pbar = tqdm(total=1, ncols=79, unit_scale=True, delay=1e-5, leave=True, unit="B", ascii=use_ascii)

and pass that to the progressbar arg for fetch.

• If total isn't set, then if progressbar returns an error, which happens during initialization of the downloader.
• Setting delay>0 seems to be necessary to avoid printing out the empty progress bar before we start iterating through it
• The other values are from the progressbar that pooch uses internally

[danshapero]

Some of the checks are failing but it looks more like a connection issue? For example, the macos-13 python-3.9 check is failing with the message

ValueError: Archive with doi:10.6084/m9.figshare.14763051.v1 not found (see https://figshare.com/articles/dataset/Test_data_for_the_Pooch_library/14763051/1). Is the DOI correct?

but when I search for the string doi:10.6084/m9.figshare.14763051.v1 in the logs for one of the passing workflows it seems to have gone through just fine.

[remrama]

Some of the checks are failing but it looks more like a connection issue? For example, the macos-13 python-3.9 check is failing with the message

ValueError: Archive with doi:10.6084/m9.figshare.14763051.v1 not found (see https://figshare.com/articles/dataset/Test_data_for_the_Pooch_library/14763051/1). Is the DOI correct?

but when I search for the string doi:10.6084/m9.figshare.14763051.v1 in the logs for one of the passing workflows it seems to have gone through just fine.

See #456