STYLE: Call write once instead of looping

[jhlegarreta]

Call write once instead of looping by concatenating all strings that need to be written.

Avoid using writelines as ImageOpener objects do not have that attribute.

Fixes:

-                for val in bvals:
-                    fid.write(f'{val} ')
+                fid.writelines(f'{val} ' for val in bvals)

and other similar errors raised for example in:
https://github.com/nipy/nibabel/actions/runs/16092302666/job/45410655457?pr=1422#step:7:23

Documentation:
https://docs.astral.sh/ruff/rules/for-loop-writes/

[codecov]

Codecov Report

❌ Patch coverage is 75.00000% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 95.34%. Comparing base (b5ecf48) to head (db08acf).
⚠️ Report is 15 commits behind head on master.

Files with missing lines	Patch %	Lines
nibabel/cmdline/parrec2nii.py	66.66%	1 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #1423   +/-   ##
=======================================
  Coverage   95.34%   95.34%           
=======================================
  Files         209      209           
  Lines       29780    29776    -4     
  Branches     3357     3356    -1     
=======================================
- Hits        28393    28390    -3     
+ Misses        948      947    -1     
  Partials      439      439           

☔ 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.

[jhlegarreta]

Typecheck issues being addressed in PR #1424.

@effigies This is ready to be merged.

[effigies]

@jhlegarreta Thanks for your patience. I've looked into this. While the bvals one is almost certainly a trivial increase in memory usage, the purpose of writing the blocks of zeros using a loop was to avoid increasing memory usage. The writelines() change recommended by ruff does not concatenate internally, instead making multiple os.write() calls, so passing a generator expression should not cause a slowdown or increase in memory use.

Applied in #1433.