FIX: Remove redundant fmap_boldref resample in desc-fmapCoreg_bold.svg report

[lobennett]

Closes #3639.

Changes proposed in this pull request

• Remove the fmap_boldref ApplyTransforms node from init_func_fit_reports_wf.
• Connect inputnode.fieldmap directly to sdcreg_report.fieldmap.

This restores the wiring that existed before #3387, which is now correct because #3467 rewired func_fit_reports_wf.inputnode.fieldmap to receive boldref_fmap.out_file — a fieldmap already reconstructed onto the BOLD reference grid.

Why

fmap_boldref was added in #3387 to resample a raw RAS fieldmap into BOLD space for the FieldmapReportlet. After #3467 introduced boldref_fmap (a ReconstructFieldmap node) and rewired its output into the reports workflow at fit.py:500, the resample became redundant. Removing fmap_boldref restores correct rendering. Actual SDC is unaffected — unwarp_boldref (fit.py:614) and the time-series resample (apply.py) use ReconstructFieldmap outputs directly. See the linked issue for figures and a centroid trace.

Impact

• Reports: desc-fmapCoreg_bold.svg now plots the fieldmap at the correct location regardless of BOLD/fmap acquisition geometry.
• Workflow: one node removed, two connections deleted, one connection added. grep -rn fmap_boldref returns zero matches after the change.
• Other outputs: none affected. fmap_boldref was reports-only.

Testing

• Built init_func_fit_reports_wf under sdc_correction=True/False × freesurfer=True/False and walked the workflow graph: fmap_boldref is gone, and sdcreg_report.fieldmap is connected directly from inputnode.fieldmap post-fix; the other three required inputs (reference, moving, mask) remain wired as before.
• Reproduced the bug on a real subject with fmriprep 25.2.5 (figures in the linked issue) and verified the fix produces correctly-located overlays by invoking FieldmapReportlet with the proposed wiring.
• ruff check, ruff format --check, and codespell pass on the modified file.

Documentation that should be reviewed

None.

[effigies]

Thanks for the detailed analysis and PR! I've triggered the tests. I still need to digest these reports and the referenced PRs, but this is much very much appreciated.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 74.24%. Comparing base (acdb0fd) to head (8b7b3ac).
⚠️ Report is 11 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #3640      +/-   ##
==========================================
- Coverage   74.57%   74.24%   -0.33%     
==========================================
  Files          62       62              
  Lines        4975     4954      -21     
  Branches      637      532     -105     
==========================================
- Hits         3710     3678      -32     
- Misses       1131     1144      +13     
+ Partials      134      132       -2     

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[effigies]

This is correct, but a bit disturbing. In general, we've tried to indicate the type and space of the thing being passed between workflows via the name, and it seems that I didn't follow that in #3467, which makes me worry that we aren't being consistent elsewhere.

I've updated the deeply outdated docstring so at least the space is documented here.

cc @mgxd as a heads up.