[SCHEMA] Add new MEG files and fix entity tables

[tsalo]

Closes #810.

NOTE: This adds in support for objects/dictionaries within datatype rules for entities, which may impact code using the schema. Tagging @bids-standard/schema-users in case this might be a problem for any of you.

Changes proposed:

• Add new schema entries for crosstalk and calibration MEG files.
• Incorporate allowed values (enums) into filename patterns.
• Fix bugs in entity tables introduced by moi in [SCHEMA] Consolidate schema files by term type #883.
• Allow multiple lines of the same datatypes/suffixes if the entity patterns differ.

[tsalo]

On the one hand, the filename templates look awesome:

sub-<label>/
    [ses-<label>/]
        meg/
            sub-<label>[_ses-<label>]_task-<label>[_acq-<label>][_run-<index>][_proc-<label>][_split-<index>]_meg.<extension>
            sub-<label>[_ses-<label>]_task-<label>[_acq-<label>][_run-<index>][_proc-<label>][_split-<index>]_meg.json
            sub-<label>[_ses-<label>]_acq-<calibration>_meg.dat
            sub-<label>[_ses-<label>]_acq-<crosstalk>_meg.fif
            sub-<label>[_ses-<label>][_task-<label>][_acq-<label>][_space-<label>]_markers.mrk
            sub-<label>[_ses-<label>][_task-<label>][_acq-<label>][_space-<label>]_markers.sqd
            sub-<label>[_ses-<label>]_task-<label>[_acq-<label>][_run-<index>]_events.json
            sub-<label>[_ses-<label>]_task-<label>[_acq-<label>][_run-<index>]_events.tsv

On the other hand, the entity tables look terrible:

"also meg" was the best I came up with 😞, and I just realized that the way I wrote the code, it can only handle two rows with the same datatype/suffixes, so... thoughts?

[guiomar]

Can we name these files meg (shielding) or something similar? This goes for the calibration and crosstalk, right?

[tsalo]

@guiomar there's no mechanism to label the different groups of suffixes in the schema at the moment. We could add one, though I'm always hesitant to add new fields to the schema.

EDIT: Yes, the new row corresponds to both calibration and crosstalk files.

[tsalo]

As a more flexible, but not very informative, alternative, I've made it able to handle any number of duplicates and have dropped the "also":

[sappelhoff]

thanks! looks all good except the entity table, as you say :-)

As a more flexible, but not very informative, alternative, I've made it able to handle any number of duplicates and have dropped the "also"

Can we not make it MEG (calibration & crosstalk) ?

[Remi-Gau]

Can we not make it MEG (calibration & crosstalk) ?

That would be fine with me.

The rest looks good. :-D

[tsalo]

Can we not make it MEG (calibration & crosstalk) ?

@sappelhoff @Remi-Gau I can't think of a straightforward way to do that with the current schema, so I would need to add a new field, like "name", for labeling whole groups of suffixes in schema/rules/datatypes/*.yaml.

[Remi-Gau]

OK I am ok keeping like this for now

[sappelhoff]

I am also OK with the current state. I think merging this and opening an issue on how to modify the schema that the meg (meg) can become a meg (crosstalk & calibration) would be a good way to go forward.

Thanks @tsalo for your work here

[guiomar]

Hi! I think I missed when the calibration and crosstalk files were incorporated to the specs. I think they weren't on the first release. Why do they have the _meg suffix? Shouldn't it have been better to have something like shielding for a suffix to disentangle from actual meg data? These are only required for SSS in Elekta systems, right? Maybe some discussions related to that happened that I missed, sorry!

[sappelhoff]

Maybe some discussions related to that happened that I missed, sorry!

it happened here @guiomar :-)

• [ENH] Specify how to share cross-talk and fine-calibration for Neuromag/Elekta/MEGIN data #581

[guiomar]

Thanks @sappelhoff! And it's not that you didn't tag me!

[tsalo]

Is everyone okay with me merging?