Root finding generating trouble about montepython_public

[brinckmann]

Hi,

Ah, I remember the P(k) de-wiggling algorithm issues, I worked on that many years ago and I agree they were never the most stable. I'm sorry you had problems with it. Luckily more recent P(k) analyses seem to have moved beyond this as a modelling method, but that also means the likelihood hasn't really been kept up to date (especially with the evolution of python/scipy) as we haven't seen any new data using a similar format. I seem to recall there are a few algorithms out there that can do it, but I don't know which work best, and it is entirely possible some of the other ones work better. Thank you for reporting back on an alternative way to do this, that could prove useful if anyone needs to revisit these old likelihoods or have new likelihoods that use de-wiggling. We always appreciate people reporting back on things like this, it's very helpful.

Best,
Thejs