Modifications for partial sky

[b-thorne]

@zonca These are the changes I think will need to be made to allow the code to consider only partial maps. Most likely an incomplete list, but I hope it helps! Thanks so much!

• Initialization:
  • Any use of the code is initialized using the models function in nominal.py. I imagine just adding a kwarg for pixel_indices to that would be the thing to do.
• Inputs:
  • Input templates are read-in in nominal.py. Modifying the models function in this file will do a lot of the work.
• Processing: Several models generate maps of random values on the fly (e.g. CMB, HD17). These will need to be checked to make sure they are either generated for the required number of pixels when possible, or for the whole sky and then reduced for e.g. the CMB.
  • components.py:
  • CMB class, taylens function.
  • Dust class, hensley_draine_2017 function through the draw_uval function, which draws a map of random values. Unfortunately, due to a work around we had to use in order to speed it up this model also relies on the initialise_hd_dust_model_bandpass function in pysm.py, which initializes the model when bandpass-integration is requested.
  • instrument.py:
  • noiser method of instrument class.
• Outputs:
  • Maps are only saved to file when the Instrument.observe is used on an instance of the Sky object. So Instrument.writer will have to be able to pad the chosen pixels with zeros when doing only partial skies.

[zonca]

thanks @bthorne93 I started working on this, I'm checking the other functions to make sure they don't need to be modified.

I'm checking decorrelation in components.py, I believe it doesn't need modifications, right?

[b-thorne]

@zonca The current decorrelation model is meant to just multiply by a scalar at each frequency, so in theory it shouldn't need to be modified.

[zonca]

I think we can close this issue, this has been merged in #11