cyriltasse / eht-imaging Goto Github PK

So following Neal's example, I'm trying to create empty observation file, then fill it with visibilities. But it's weird - there are holes ion the uv track - do you also see that on your side?

OK - things are plugged in now - seens to be running as expected. I've just simulated a point source at the phase center, did a dirty image of it, and used that as the prior (so a pretty crappy prior)

ClosureImager.py --MSName 0000.MS --FITSName test.dirty.fits

Starting model:

Very quickly it converges to a point - which is good

More tests after the week-end...

ok - running a test on 3C295 dataset given by Etienne (rotated, massively averaged) using the lastes commit on the UseMS branch. I've first deconvolved with DDFacet and used the output model image as the prior

ClosureImager.py --MSName L242400_SB150_uv.dppp.phaseshift.avg150.MS --ColName DATA --FITSName test.model01.fits --FlagAnts RS,CS

The 0th iteration gives

but then

after some iterations. Obvioulsy it wrong. I might do something wrong. Particularly - I don't really know how to handle flagged data... More later but if you have insights feel free to share.

Has anyone tried to clone and run it?

Do you see anything wrong Neal with the options I'm using here at my first call to the solver?

EHT.imager_func(self.EHT_obs,self.EHT_im_prior,self.EHT_im_prior,1.0,d1='bs',d2='cphase',s1='gs',maxit=1000)