Removing *subject_ids* argument in TUHAbnormal example.

[bruAristimunha]

Related with the issue #377.

Removing a deprecated parameter in the TUHAbnormal example.

[codecov]

Codecov Report

Merging #402 (62520ed) into master (d0abec3) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master     #402   +/-   ##
=======================================
  Coverage   84.20%   84.20%           
=======================================
  Files          46       46           
  Lines        3616     3616           
=======================================
  Hits         3045     3045           
  Misses        571      571           

[robintibor]

So is this correct @gemeinl ?

[gemeinl]

No, actually not correct.
The function load_example_data has an argument subject_ids which does not have any effect anymore, due to the changes introduced here. Therefore, the entire dataset will be loaded instead of just 10 recs.
We still need this argument, however, its meaning has changed. It is not subjects we are refering to but recordings. I suggest to rename the argument to n_recordings and give this as input to TUHAbnormal which has the an argument recording_ids (see

braindecode/braindecode/datasets/tuh.py

Lines 207 to 208 in 884a78b

	def __init__(self, path, recording_ids=None, target_name='pathological',
	preload=False, add_physician_reports=False, n_jobs=1):

).

[bruAristimunha]

Many thanks for the explication @gemeinl. I was wondering, can you review it again to make sure everything is correct?

[gemeinl]

I suggest to stick as close to the original code as possible.
So change the input argument n_subjects to n_recordings in load_example_data and create the recording_ids for TUHAbnormal inside as before.

[bruAristimunha]

Done @gemeinl. Can you run it once? I don't have this dataset.

[gemeinl]

Great, thanks a lot @bruAristimunha !

[bruAristimunha]

Can you merge? @robintibor

[cedricrommel]

Could you solve the conflicts @bruAristimunha before I merge?

[bruAristimunha]

Done @cedricrommel.