Normalize flip_angles in DCEAcquisitionParams.__post_init__

[ParikshitGupta2605]

Normalization now happens in DCEAcquisitionParams.post_init, the VFA-specific empty check is restored (load-bearing, not dead code since fixed-T1/t1_assumed acquisitions legitimately construct with empty flip_angles), redundant casts in runner.py/vfa.py are cleaned up, and tests are broadened to cover construction-time validation directly (None, scalar, non-numeric, NaN/inf, 2D array, empty-with-t1_assumed). Ready for another look.

[ParikshitGupta2605]

Thanks for the earlier feedback, @ltorres6 and @MohamedNasser8 — I've reopened this with the updated approach.

What changed:

Moved the fix from vfa.py into DCEAcquisitionParams.post_init (osipy/common/types.py), so flip_angles is normalized into a validated list[float] once, at construction — accepting lists, tuples, and NumPy arrays, and raising a clear ValueError for None, non-numeric, non-1D, or non-finite input.
Empty flip_angles is intentionally still allowed at the dataclass level — turns out fixed-T1 acquisitions using t1_assumed legitimately construct DCEAcquisitionParams with flip_angles=[] (confirmed via dce_pipeline.py's branch on t1_assumed is not None), so enforcing non-empty at the dataclass level would've broken that workflow.
Restored the empty check inside compute_t1_vfa itself, since that's the actual place where non-empty flip_angles are required, it's load-bearing now, not dead code.
Cleaned up the now-redundant list/np.asarray casts in runner.py and vfa.py that were doing defensive re-casting the dataclass now handles.
Broadened the tests to cover construction-time validation directly (None, scalar, non-numeric, NaN/inf, 2D array, empty-with-t1_assumed), plus a regression test confirming compute_t1_vfa still rejects empty flip_angles.

[MohamedNasser8]

Thanks @ParikshitGupta2605! Left a couple of comments to fix

[ParikshitGupta2605]

@MohamedNasser8 - addressed both: Pushed both fixes, threads resolved. Tests are passing locally. Let me know if you need anything else!

[ParikshitGupta2605]

Hi @MohamedNasser8, I'd like to get more involved with OSIPI and contribute to the project. Would it be possible to send me an invite to the OSIPI Slack workspace so I can introduce myself in the project channel, as mentioned in the contributing guidelines? Thank you!

[MohamedNasser8]

Hey @ParikshitGupta2605 you can join by sending email to ismrmosipi@gmail.com, they will invite you to slack workspace.
Sorry for the delay, didn't notice it.

[ParikshitGupta2605]

Thanks for letting me know! No worries about the delay.