Open LucaNap opened 3 years ago
I have found out that, with a GP model, everytime results.rv.evaluate() is called the "results" are updated and this makes it impossible to call the function again (if for example one is trying to plot points from two or more instruments, or if one is trying to make more than one plot using results.rv.evaluate).
P.S. Even more strange is that this error can't be bypassed by making a copy of "results" (dataset.fit) because the copy also gets overwritten after results.rv.evaluate() is called.
Update: I've found a workaround, which relies on using
import copy
results_1 = copy.deepcopy(results)
everytime results.rv.evaluate is called (now becomes results_1).
Hi @LucaNap,
Yes, this is actually not a desirable output of calling the evaluate
method for sure, and is in my to-do to look at. I hope to have it fixed for the next juliet
version; will thus leave this open for now!
Thanks for bringing this up with such a detailed report!
Néstor
Finally got to this -- fixed on dev
. Will be on the next juliet
version --- but users that need this/find this problem can go and install the dev
version.
Hello, I am having an issue trying to plot the fit of radial velocities only when (rv) GPs are involved. While trying to replicate the example in Juliet's documentation:
I get the following error:
It looks like the error is related to GPregressors, but of course the evaluation can't be done without it.