Minor bug related to subject naming in GUID file

[mtnhuck]

The template suggests that files be formatted as SID - GUID. When I tried to do this (say for subject "sub-sid000001" and a file formatted as "sub-sid000001 - MYGUID", it won't find the subject, but if I remove the leading "sub-" it works just fine with a resulting file formatted as "sid000001 - MYGUID". Just a heads up if anyone else has the same problem.
Cheers
Jeremy

[chrisgorgo]

That's surprising. How does your participants.tsv file look like? Does the dataset pass BIDS Validator https://github.com/INCF/bids-validator?

[mtnhuck]

The only error I get is for 1 fieldmap file related to a missing "EchoTime1" and "EchoTime2", which may be corrupt, all other files check out. The participants.tsv is formatted as
participant_id age sex group
sub-sid000001 18 M control
...

[yarikoptic]

I think it is just reflection of a fact that SID is undefined in BIDS format. It has participant_label (e.g. sid000001) and participant_id (e.g. sub-sid000001). For the purpose of this mapping, it should be participant_label.

FWIW, this is how our constructed fake GUIDs mapping file looked (the one we with @mtnhuck used before) and the shell command we used to generate it. this one didn't cause any trouble to the script

(git)hopa:~datalad[master]NDA/uploads
$> head sid-to-GUID.map
sid000064 - GUID0000
sid000079 - GUID0001
sid000100 - GUID0002
sid000139 - GUID0003
sid000148 - GUID0004
sid000178 - GUID0005
sid000182 - GUID0006
sid000259 - GUID0007
sid000264 - GUID0008
sid000269 - GUID0009
2 18406.....................................:Thu 11 Jan 2018 05:27:04 PM EST:.
(git)hopa:~datalad[master]NDA/uploads
$> k=0; for f in */sub-sid000*; do echo $f | sed -e 's,.*-,,g'; done | uniq | while read sid; do echo "$sid - GUID000$k"; k=$(($k + 1)); done >| sid-to-GUID.map

[mtnhuck]

Thanks to both of you on this. Its doesn't seem like a big deal given the easy workaround I had.
Cheers
J

[chrisgorgo]

So I guess there is a mistake in the readme?