RF: Update primary bold workflow to incorporate single shot resampling (#3114)

This is a pretty big one, so I'm pausing now. Things that this PR is
doing:

1) Adding single-shot reampling interfaces for fieldmaps and BOLD
volumes/series.
2) Gets rid of the old `single_subject_wf` and replaces it with what had
been `single_subject_fit_wf`.
3) Creates a new workflow in `workflows.bold.base` to wrap around fit
and further stages.
4) Adds an `init_bold_native_wf` to `workflow.bold.fit`, which handles
slice timing correction and multi-echo, and always returns both a
native-resampled file as well as something that can be resampled and the
transforms/fieldmaps it should use.
* For single-echo, the resample-able thing is the STC'd file, and you
get motion transforms and fieldmap ID back.
* For multi-echo, the resample-able thing is the optimal combination,
and motion transforms and fieldmap are dropped.

---------

Co-authored-by: Mathias Goncalves <[email protected]>

Author: effigies

docs/api.rst

@@ -14,13 +14,13 @@ you get started.
 Setting up your development environment
 ---------------------------------------
 We believe that *fMRIPrep* must be free to use, inspect, and critique.
-Correspondingly, you should be free to modify our software to improve it 
+Correspondingly, you should be free to modify our software to improve it
 or adapt it to new use cases and we especially welcome contributions to
 improve it or its documentation.
 
 We actively direct efforts into making the scrutiny and improvement processes
 as easy as possible.
-As part of such efforts, we maintain some 
+As part of such efforts, we maintain some
 `tips and guidelines for developers <https://www.nipreps.org/devs/devenv/>`__
 to help minimize your burden if you want to modify the software.
 
@@ -35,3 +35,4 @@ Workflows
 
 .. automodule:: fmriprep.workflows.base
 .. automodule:: fmriprep.workflows.bold
+.. automodule:: fmriprep.workflows.bold.fit

docs/workflows.rst

@@ -80,6 +80,8 @@ single reference template (see `Longitudinal processing`_).
         skull_strip_fixed_seed=False,
         t1w=['sub-01/anat/sub-01_T1w.nii.gz'],
         t2w=[],
+        msm_sulc=True,
+        precomputed={},
     )
 
 .. important::
@@ -266,7 +268,8 @@ packages, including FreeSurfer and the `Connectome Workbench`_.
 
     from smriprep.workflows.surfaces import init_surface_recon_wf
     wf = init_surface_recon_wf(omp_nthreads=1,
-                               hires=True)
+                               hires=True,
+                               precomputed={})
 
 See also *sMRIPrep*'s
 :py:func:`~smriprep.workflows.surfaces.init_surface_recon_wf`