Closed grburgess closed 5 years ago
In the .txt file, it is -2 loglike (as getdist expects). It does actually explicitly say this in the Readme, is there any reason you'd expect this to be otherwise?
Having a look in the getdist source code: https://github.com/cmbant/getdist/blob/master/getdist/mcsamples.py#L186
It calls this loglike in the input and the object member is names loglike.
Then here:
https://github.com/cmbant/getdist/blob/master/getdist/mcsamples.py#L457
it switches and calls this -2loglike.
This seems to be a naming issue with getdist so I will ping them. It could just be really confusing for people not paying attention. The output that one would get from the posterior object in polychord is -2loglike, but the object is called loglike. I only noticed when wrapping this up in our modeling code that the loglikes between multinest and polychord were really different because I naively accepted the member name as a good description :)
It might be worth it to add a note in the docs somewhere.
I see! Could you point out the misleading documentation in PolyChord (or better yet, submit a pull request with the documentation that you think would make it clearer)?
Already working on a PR
I'm trying to figure out if the value of loglike in the getDist output is -2 loglike or loglike. From the root.txt file, it looks like -2 loglike is being passed in.
Cheers