FIX: Use seconds as time units when converting PARRECHeader to Nifti1Header

[effigies]

parrec2nii produces NIfTI files with TRs in seconds but records the units as msec

#692 proposed retaining the PAR/REC msec units by removing a division by 1000.. This PR instead proposes keeping things in seconds.

cc @Roosted7 @epongpipat

Replaces #692.

[effigies]

Would be glad if anybody wants to weigh in on which of these two approaches is best. I think this one will specifically fix conversions and not break anybody's workflows that have been depending on PARRECImage providing TRs in seconds. I can see an argument for keeping things close to the image, but I think standardizing on seconds makes more sense.

[codecov]

Codecov Report

Merging #931 into maint/3.1.x will decrease coverage by 0.08%.
The diff coverage is 100.00%.

@@               Coverage Diff               @@
##           maint/3.1.x     #931      +/-   ##
===============================================
- Coverage        91.86%   91.77%   -0.09%     
===============================================
  Files               97       97              
  Lines            12360    12360              
  Branches          2177     2177              
===============================================
- Hits             11354    11344      -10     
- Misses             675      681       +6     
- Partials           331      335       +4     

Impacted Files	Coverage Δ
nibabel/parrec.py	92.50% <100.00%> (ø)
nibabel/environment.py	75.00% <0.00%> (-20.00%)	⬇️
nibabel/casting.py	85.28% <0.00%> (-0.87%)	⬇️
nibabel/dft.py	80.36% <0.00%> (-0.62%)	⬇️
nibabel/nifti1.py	92.12% <0.00%> (-0.31%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 97a0467...7759954. Read the comment docs.

[effigies]

Anybody care to have a look? If not I'll exercise supreme maintainer powers and hit merge tomorrow.

@Roosted7 @epongpipat I would appreciate your opinions on this one.

[matthew-brett]

Hmm - it's a tough call, I agree. In general, we try to report the file from nibabel as a native reader would have reported it, but here I guess we allowed ourselves to make the file appear more NIfTI-like. This fix seems better for back-compatibility, but you could also make an argument for converting to seconds only when converting the image to NIfTI format (so, not in _calc_zooms, but in as_analyze...). But - I do agree that we shouldn't make odd milliseconds NIfTI images, they would be very confusing.

[effigies]

How about going with this for a bug fix, and we can move PARRECImage to report msec TR in a feature release?

Maybe it would be worth revisiting #567. I ran into some conceptual problems trying to think through the API there.

[matthew-brett]

Sounds reasonable ...