Add functional phase data patterns to file_level_rules

[tsalo]

Closes #834. Adds support for functional phase data (added to the BIDS specification in bids-standard/bids-specification#128 and bids-standard/bids-specification#312).

[codecov]

Codecov Report

Merging #835 into master will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master     #835   +/-   ##
=======================================
  Coverage   89.08%   89.08%           
=======================================
  Files          81       81           
  Lines        2328     2328           
  Branches      427      427           
=======================================
  Hits         2074     2074           
  Misses        218      218           
  Partials       36       36

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 0745c4d...fc0f20a. Read the comment docs.

[effigies]

I would clump the suffixes together. bold, sbref, cbv and phase are all clustered on the big table, so let's put them together here.

Beyond this, @rwblair @nellh are there further considerations for adding a suffix?

[tsalo]

@effigies Done!

In reordering the tokens I noticed that _defacemask is included as a func token even though it's only included under anat in the Entity table. I can open a related issue either here (if it should be under anat) or in the specification repository (if the Entity table is wrong), but I wasn't sure which way it should go.

[rwblair]

I can't seem to find a reference to defacemasks for functional images in the standard even outside the entity table. If its removed then we can simplify the whole section by adding a 'func_ext' like other sections have for matching .nii nii.gz and .json, defacemask is the only one missing .json extension.

[effigies]

Perhaps check the git blame to see if there was a use case driving its inclusion?

[rwblair]

Looks like its origins are in #233. Implies that at some point in the spec that func defacemasks were acceptable.

[rwblair]

ds000113 has funcitonal defacemasks for every run.
ds001107 is using them as well

So I guess well keep them in the validator for now. I'll keep looking to see if there is any further discussion around this.

[tsalo]

Okay, that sounds good.

Regarding supporting functional phase data, is there anything else that needs to be changed? It seems to be working locally on my own complex (phase+mag), multi-echo data.

[rwblair]

@tsalo one last thing, could you add an example of a valid filename to bids-validator/tests/type.spec.js around line 41 to have it covered by some sort of test?