BUG: Fix URL

[larsoner]

Before merging …

• Add config for 4107
• Add test that all tests show up in CircleCI config
• Add 4107 and whatever else we need to tests
• Changelog has been updated (docs/source/changes.md)

Okay the basic fix was easy. Draft because I'm not sure how this download has ever worked, as this error appears in the blame from 2 years ago when it was added. It appears not to be used on CircleCI -- should it be? If so, I'll add another little sanity check test here that any dataset in datasets.py has at least some matching code in .circleci/config.yml.

Closes #599
Closes #539

[larsoner]

I'll probably take care of #539 here as well since it should be easy enough hopefully

[agramfort]

Personally I prefer this dryer option. +1 merge after the CIs are green. Note that the new dataset shall be added to the documentation page too.

[larsoner]

I think there is a bug with 1971 with participants.tsv:

participant_id  participant_id  sex     age     handedness
sub-001 CQ2     F       18-38   R
sub-002 BT4     F       28      R
sub-003 CQ1     F       18-38   R
sub-004 CQ3     F       18-38   R
sub-005 CP9     M       18-38   R
sub-006 CA7     M       18-38   R
sub-007 CB6     M       32      R
sub-008 BV2     M       38      R
sub-009 CP8     F       18-38   R
sub-010 CP1     F       23      R
sub-011 CP2     M       18-38   R
sub-012 CO9     F       18-38   R
sub-013 CP5     M       18-38   R
sub-014 CL1     M       18-38   R
sub-015 BL7     M       28      R
sub-016 CO3     M       18-38   R
sub-017 CE2     M       18-38   R
sub-018 CV6     F       18-38   R
sub-019 CJ2     F       18-38   R
sub-020 CY8     M       18-38   R

It has two participant_id columns. I'll need to exclude it for now I think.

EDIT: Looks like those subject ids are not even in the original file: https://openneuro.org/datasets/ds001971/versions/1.1.1/file-display/participants.tsv / mne-tools/mne-bids#1065