Current Master Failing CI

[jbteves]

Summary

For some reason in 4a56876 we're failing CI, which is bizarre since it only changed the license type. According to the failed report view here the 4-echo test generated mismatched outputs, where we expect s0v.nii and two other files to be generated; however, they are not.

Next Steps

@tsalo @emdupre @rmarkello any thoughts on what's happening?

[tsalo]

The error message is misleading. s0v.nii is not missing- it's just that no BOLD components are being detected so none of the denoised files are written out. I don't know what's causing the issue, however something similar is also impacting #358.

[dowdlelt]

Looking through the artifacts, its finding plenty of components, its just ignoring almost all of them:

Its odd how 'close' the kappa and rho values are.
Not sure how this could be related to the license change at all. Really doesn't seem like it could be. The mask and components all still look 'good'.

EDIT:
This is insanity: Component 1 from successful test and failed test are effectively the exact same, but kappa and rho values changed...