ApplyCal flulljones - order of ph, amp soltab important?

[henedler]

Hi,
I was puzzled for a while with a case where the combination of fulljones and invert=False in the ApplyCal step produced strange results. The weird thing was that my solutions appeared to be good, so applying them to the data (invert=True) lead to good results.

Eventually I figured that the only difference in the parsets is the order of the phase and amplitude soltab, so one time I used cor.soltab=[amplitude000,phase000], the other cor.soltab=[phase000,amplitude000]. Changing the order seems to do the trick.
From the documentation, I would not assume the order of the soltabs to be important.

[rvweeren]

So @henedler when invert=True order does not matter? and with invert=False which order produces the correct result?

[henedler]

Hi @rvweeren, I did not test the case invert=False in combination with the order [phase000,amplitude000], but I would expect that the behavior is similar.
I think this could be more a case of an ambiguous documentation than a bug, my initial understanding was that one provides two soltab names in ApplyCal and DP3 figures internally which one is the amps/phases of the fulljones. But it might be that the first one is always assumed to contain the amplitudes and the second one the phases.

[rvweeren]

@henedler thanks for the info. I had a look here  https://www.astron.nl/citt/DP3/steps/ApplyCal.html#

"The name or names of the H5 soltab. Currently only used when correction=fulljones, in which case soltab should list two names (amplitude and phase soltab)"

That does indeed not suggest to a user that the order is relevant. So if soltab=[amplitude000,phase000] is indeed the only correct way for fulljones the documentation should make this very clear. In fact, I would even say DP3 should give an error if a user tries soltab=[phase000,amplitude000] with fulljones applycal. Can one of the developers give some more info on order and fulljones? (I'm getting worried about the fullones applycal now in my own scripts as I was not paying attention to order)

[henedler]

Reopend on request of Reinout, since the issue persists - it would be good to either explicitly check the order of amp/phase soltabs or at least update the documentation of the Applycal step under the point "soltab".