search

msimet / Stile

Stile: the Systematics Tests In Lensing pipeline
BSD 3-Clause "New" or "Revised" License
9 stars 6 forks source link

#67 Star size residual correlation functions #72

Closed msimet closed 9 years ago

msimet commented 9 years ago

And here's the star size residual correlation function: a scalar correlation function of (psf_sigma - sigma)/psf_sigma. Pretty straightforward, except for revealing a couple of bugs in pieces of code that we haven't really touched on before...

HironaoMiyatake commented 9 years ago

I went through the code, and wrote one comment about the short name of this test. I found a bug; data was put into a method to calculate a correlation function, but it should be new_data.

After @msimet check out those, I think this branch can be merged into master.

msimet commented 9 years ago

I changed the name of the test, as suggested.

I also made one other change (to make sure the sigma columns are calculated for data sets other than the first one, if they're present). Do you want to look at the code to make sure it looks okay to you? The commit history looks a little weird, unfortunately--my text editor didn't reload the file when I thought it did, so I ended up changing some things back that I didn't mean to...

HironaoMiyatake commented 9 years ago

I think it would be better to change the denominator of new_data['sigma'] = (new_data['psf_sigma'] - new_data['sigma'])/new_data['sigma'] to new_data['psf_sigma'], to make it less noiser (this is what I do for other fractional residuals).

Also you use objects_list = ['star'] , but do not we want to use objects_list = ['star PSF'] ? At least rho1 is the latter, so it would be nice to make it consistent.

msimet commented 9 years ago

Okay, made both these changes--does that look good?

HironaoMiyatake commented 9 years ago

Yes, I think it's ready to merge!

rmandelb commented 9 years ago

Thinking ahead to the meeting next week, do you have any plots that you could show using this code on the old or new reductions?

rmandelb commented 9 years ago

(I realize you don't want to put that on here since it's proprietary data, but could you e-mail them?)

msimet commented 9 years ago

I tested this on a few little patches but haven't run the whole of the old (or new) reductions. I can certainly do that by meeting next week, though, assuming Coma gets up and running...

msimet commented 9 years ago

(I'm gonna merge this, but we can keep talking on the closed issue if we need to)