@jan-petr: will run the flavors tonight and I'll post the results tomorrow.

@MichaelStritt see my first results

• Plenty of datasets differ in acknowledgement. Despite the ignoring. My guess is that if both reference and source have the field, then this is ignored, but if only one of the compared sets have it, then it complained about the difference. So it is not about the content but about a difference in existence.
• Siemens_PCASL_2DEPI_VB17A_BSupN_1 has different asl.nii. By coincidence, this is a dataset from the same center as the one I was adding for the newest tests. My guess is that previously, we were ordering this dataset wrong and this now got fixed :) You can see that in the reference - the motion across repetitions is sketchy. And in the new version, this is much more smooth. Do you agree with this? If yes, then we have to replace the reference asl.nii in the flavors with this newly converted asl.nii
• The only other issue is with Hadamard - but Beatriz knows about this and is fixing it in some other issue, so we can ignore that.

@MichaelStritt Can you please upload the new asl.nii to flavors and fix the acknowledgement issue in this issue?

@jan-petr: here is the print out of the bids comparison: bids_comparison.txt

Plenty of datasets differ in acknowledgement. Despite the ignoring. My guess is that if both reference and source have the field, then this is ignored, but if only one of the compared sets have it, then it complained about the difference. So it is not about the content but about a difference in existence.

I saw the acknowledgement thing. will fix it immediately, should be easy 👍

Fix: 7ef0051

Siemens_PCASL_2DEPI_VB17A_BSupN_1 has different asl.nii. By coincidence, this is a dataset from the same center as the one I was adding for the newest tests. My guess is that previously, we were ordering this dataset wrong and this now got fixed :) You can see that in the reference - the motion across repetitions is sketchy. And in the new version, this is much more smooth. Do you agree with this? If yes, then we have to replace the reference asl.nii in the flavors with this newly converted asl.nii

I saw that too. That's really weird.

Dataset: Siemens_PCASL_2DEPI_VB17A_BSupN_1
File:      \sub-Sub1\perf\sub-Sub1_asl.nii.gz
           RMSE of NIFTIs above threshold.
Warning: RMSE of NIFTIs above threshold: \sub-Sub1\perf\sub-Sub1_asl.nii.gz    
> In xASL_bids_CompareStructures>printDifferencesAsWarnings (line 217)
  In xASL_bids_CompareStructures (line 206)
  In xASL_test_Flavors (line 137) 

I checked out both rawdata NIfTIs in ImageJ and it seems like the new version has less "motion". It really seems like there's something wrong with the ordering in the old one. I personally would say the new one is correct.

Fix: https://github.com/ExploreASL/FlavorDatabase/commit/4f4fd0781440f7c67477b0c8365a846aa3cfaa5d

The only other issue is with Hadamard - but Beatriz knows about this and is fixing it in some other issue, so we can ignore

Agree, we shouldn't fix this here and rather wait for @BeatrizPadrela to fix this 😛