Ciftify

[ENC] fmriprep base layer for fmriprep_cifitfy Docker is now 1.3.2
[ENC] nose dependancy removed from CI tests
[FIX] markdown to rest conversion of the long descriptor in the setup.py

[ENC] added new ciftify_dlabel_to_vol script to project surface-based labels into subject-specific nifti volume rois
[FIX] fixed bugs in ciftify_meants where the meants.csv was shorter than the labels if the labels were located in a location with no data. (Now will output a row of zero's for missing data).
[ENC] added many smoke tests related for ciftify_meants and ciftify_seed_corr functions
[FIX] updated the PINT test fixtures to correspond with newer nilearn image clean outputs (note: we now confirm that cifitfy_clean_img outputs should be different if for nilearn 0.5 and later (but slightly better)

This is the final release before resubmitting revisions to NeuroImage. Enhancements include:

[ENC] adding the user an option in input a pre-calculated MNI transforms to `ciftify_recon_all`
[ENC] adding a new `ciftify_falff` script drafted by Delaney Johnson during her co-op term.

Incremental bug fixes for `ciftify_clean_img` and `cifti_vis_PINT`.
[FIX] heatmaps in cifti_vis_PINT have white lines in correct place (in newer python)
[FIX] ciftify_clean_img doc string now clearer
[FIX] ciftify_clean_img does not assume that nifti TR's are in ms
[FIX] ciftify_clean_img now exits gracefully when confound.tsv input is missing
[FIX] all meants functions now don't exclude voxels/vertices with means of zero

[ENC] - bidsapp now has fmriprep:1.3.0.post2 as the base layer
[ENC] - bidsapp can now be set to read BIDS common derivatives from a specified path (with fmriprep turned off) using `--read-from-derivatives` option
[ENC] - bidsapp can also read old fmriprep (pre 1.2.0) derivatives with `--older-fmriprep`
[ENC] - bidsapp uses python 3.7 base with pybids 0.7.0

[ENC] - ciftify_subject_fmri now runs boundary-based registration if the --FLIRT-to-T1w flag is given (previously) this was FLIRT (corratio-based)