ENH: Support combined surface-volume reference spaces

[tsalo]

Closes #881. This is a ways away from being usable, but it might be good to get folks' thoughts on it before I put too much effort into it.

The basic idea is that users can request any combination of surface and volume that they might want as output CIFTIs.
For example, the common fsLR space with den-91k corresponds to CIFTIs with the left and right hemisphere surface components in fsLR space with 32k vertices in each hemisphere and the subcortical volumetric component in MNI152NLin6Asym space with 2 mm voxels. In my proposal, users would define that as --output-spaces fsLR:den-32k::MNI152NLin6Asym:res-02. This could also be extended to other combinations, such as:

fsnative:den-32k::native:res-native --> <source_entities>_space-fsnative_volspace-native_den-32k_res-native_bold.dtseries.nii
onavg:den-41k::MNI152NLin2009cAsym:res-02 --> <source_entities>_space-onavg_volspace-MNI152NLin2009cAsym_den-41k_res-02_bold.dtseries.nii
dhcpAsym:cohort-42:den-32k::MNIInfant:cohort-2:res-1 --> <source_entities>_space-dhcpAsym_cohort-42_volspace-MNIInfant_volcohort-2_den-32k_res-1_bold.dtseries.nii

To discuss:

• Do we want separate space and volspace, cohort and volcohort entities?
  • @effigies' proposal: Instead of <source_entities>_space-dhcpAsym_cohort-42_volspace-MNIInfant_volcohort-2_den-32k_res-1_bold.dtseries.nii, maybe <source_entities>_space-dhcpAsym42+MNIInfant2_den-32k_res-1_bold.dtseries.nii.
  • cohort, volspace, and volcohort are not part of BIDS at the moment, so this approach would require fewer changes to BIDS (though we'd still need to allow + symbols in entity values).

Changes proposed:

• Add a delimiter between surface and volume space strings (::) to allow combinations of surfaces and volumes for CIFTIs.
• Add volspace and volcohort keys to the spec dictionary.
• Add a cifti attribute to the Reference object to denote if a space is CIFTI or not.

[codecov]

Codecov Report

Attention: Patch coverage is 56.60377% with 23 lines in your changes missing coverage. Please review.

Project coverage is 69.32%. Comparing base (72327c5) to head (84c8e46).

Files with missing lines	Patch %	Lines
niworkflows/utils/spaces.py	56.60%	17 Missing and 6 partials ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #883      +/-   ##
==========================================
- Coverage   69.40%   69.32%   -0.09%     
==========================================
  Files          86       86              
  Lines        8483     8526      +43     
  Branches     1059     1079      +20     
==========================================
+ Hits         5888     5911      +23     
- Misses       2373     2387      +14     
- Partials      222      228       +6     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[tsalo]

This check would hopefully only be necessary in the short-term, since mixed standard/non-standard CIFTIs should be allowed.

[tsalo]

Alternatively, I could use a new dim value, like 2.5, instead of adding this attribute.