Empty fields in participants.tsv

[eurunuela]

This discussion is related to #244.

I'm opening this discussion because we should decide whether we want to have a "n/a" (not answered) field or if phys2bids should only write the subject id in participants.tsv. The "n/a" would "advice" the users they should be adding that info following BIDS guidelines.

So, there are two possible options here (feel free to add more):

1. Only write subject ID and other fields provided in the YAML file.
2. Write the three BIDS recommended fields as "n/a" when they're not provided in the YAML

I'd like to know what you think @smoia @vinferrer @tsalo @robertoostenveld

[tsalo]

I say we take our cue from heudiconv, which adds those columns (along with "group") and puts a n/a if the information is unavailable.

[smoia]

@marcelzwiers, what is the behaviour of BIDScoin?

[marcelzwiers]

In Bidscoin I use the n/a value, like this (participants_table is a pandas dataframe):

    if participants_tsv.is_file():
        participants_table = pd.read_csv(participants_tsv, sep='\t')
        participants_table.set_index(['participant_id'], verify_integrity=True, inplace=True)
    else:
        participants_table = pd.DataFrame()
        participants_table.index.name = 'participant_id'

 [...]

participants_table.replace('','n/a').to_csv(participants_tsv, sep='\t', encoding='utf-8', na_rep='n/a')

[smoia]

Given that the two DICOM bidsifiers we're adopting as reference have a "n/a" value for missing field, I would do the same.
I think this is the current behaviour, right @eurunuela? If so, and there is no strong opinion against it, we can close this issue.

[eurunuela]

That's the current behavior. Closing the issue then.