Open ThomasTram opened 3 months ago
Hi Thomas,
Thanks a lot for doing this, would you like to go ahead and add it to the private devel branch? I can also do it, but will be quite busy the next three weeks and might not get around to it soon. Then once we have everything there (e.g. DESI BAO and new WMAP Python 3 wrapper from @schoeneberg , NPIPE, ground based CMB, etc) we can put out a new release.
If there are any other fixes or improvements you've done that haven't been included yet (e.g. PR #372 ) you can also add those. Thanks a lot!
Best, Thejs
@brinckmann I actually forgot that I had access to the private repo. I can add the changes yes.
I implemented the BK18 likelihood based on the previous BK14 likelihood. I initially wanted to do it by inheriting from the BK14 likelihood class, but in the end a lot of small things had changed so I decided to copy-paste it and then edit it. There are actually two likelihoods, one using only B-modes (which is the default) called
BK18lf_dust
and one that includes E-modes which is calledBK18lf_dust_incEE
. The two likelihoods are identical and differ only in the datafiles and the maps used inside the .data file.To validate the likelihood I recreated Figure 4 of BKXIII in 2110.00483. Using this script:
I get these figures:![image](https://github.com/brinckmann/montepython_public/assets/5593952/960d83e4-cf01-40d5-b17a-f82a74010f54)
which at least by eye agree.