Lensing induced covariance

[raphkou]

I implemented the lensing induced covariance in the CMB primary covariance block. I also evaluated the impact of the N1 bias on the cross-covariance between CMB primary power spectra and reconstructed CMB lensing, which turned out to be negligible.

[itrharrison]

Thanks for this! The notebook is very good at showing how we get these cov terms, which is definitely useful.

It would be great if it used the framework in the SmoothTheory example we are building more fully.

So in the SmoothTheory example there are a set of scripts to create the different probe data files:

https://github.com/simonsobs/SOLikeT/tree/16da3950342ace4a2a878c1c6f0478bd172b3ada/examples/smooth/scripts

which use some yamls which specifiy fiducial settings for the SmoothTheory example (cosmological and accuracy parameters etc)
https://github.com/simonsobs/SOLikeT/tree/16da3950342ace4a2a878c1c6f0478bd172b3ada/examples/smooth/yamls

Can we change what is here so that your notebook / scripts construct the cross covariance for those data files and model, then construct a new multi-probe sacc with the full cross-covariance?

[raphkou]

After discussing together, we clarified that we want to update this notebook/script to use the fiducial settings of the SmoothTheory example (cosmology and accuracy parameters). But those fiducial settings should be updated to match the ones already used to compute the auto-covariance.
We also want to compute other cross-covariance terms between CMB primary, CMB lensing, galaxy shear (which only enters through its cross-correlation with CMB lensing).