Description

Minor bugs when testing TestDataSets

Requirements

• When processing is requested, there should be no warning about the lack of an output structure
• Incorrect warning Longitudinal registration: x.P does not contain subjectIDs ...
• If a fieldname is correct but on incorrect location, we now incorrectly warn that the field is deprecated. This warning should be changed to:
  x.bUseBasilQuantification was moved to x.Q.bUseBasilQuantification, please tweak dataPar.json accordingly
• Missing data that are set to NaN should be subtle warnings: fprintf(2, ) instead of warning( )
• if x.Q.BackgroundSuppressionNumberPulses is missing, this is now always assumed to be 0, whereas for 3D sequences this should default to 5 (GE) or 4 (Philips/Siemens)
• Add transformation to standard space for BASIL output
• Update xASL_qc_TestExploreASL with the new log output folder
• Avoiding genpath for bids-matlab in ExploreASL_Initialize: this goes wrong with older Matlab versions

A subtle warning should be included to give the user a heads-up that GE eASL is acquired as Hadamard but processed as multi-PLD: fprintf(2, “GE eASL detected, acquired as Hadamard but exported as multiPLD”).
Jan: Actually, after checking the flavor and fixing NEXes, there's not a robust and reliable way to detect that it is Hadamard and not multiPLD with similar PLDs. But I've fixed the TotalAcquiredPairs for Hadamard, so all should now work seamlessly whatever case we have.

For Siemens, we cannot assume that we can extract TotalAcquiredPairs from the number of exported volumes.
Jan Yes, but that's the best shot we have. Besides, normally, TotalAcquiredPairs are taken from NEXes from DICOM where it is usually quite reliable with GE and Siemens. So the estimation from the number of repetitions only happens when all deltaMs are averaged and TotalAcquiredPairs is 1. So there we really have to start counting. But that's a situation really only happening rarely.

Release notes

Minor fixes in warnings. Added to a different issue.