通过 python 中的 rpy2 在 BLME r package 中传递自定义 function

Question

Issue: In Python, I am trying to pass a dnorm() R function to the argument cov.prior of the R package called Blme . The example is written on page 4 for the custom prior. I wanted to use Gaussian as my covariance prior distribution, but I don't know how to wrap the dnorm() function with the custom() in order to make it work.

Here is my attempt 1

from rpy2 import robjects 
from rpy2.robjects.packages import importr

stats = importr("stats")
blme = importr("blme")
lmc = robjects.r(f"lmerControl({control})")
dnorm = stats.dnorm
covprior = robjects.r(f"custom({dnorm})")
blme.blmer(formula, data=data, REML=REML, control=lmc, contrasts=contrasts, **{cov.prior=covprior})

Here is my attempt 2

from rpy2 import robjects 
from rpy2.robjects.packages import importr

stats = importr("stats")
blme = importr("blme")
lmc = robjects.r(f"lmerControl({control})")
robjects.r(''' penalityFn <- function(sigma){dnorm(sigma, 1, 1)}''')
covprior = robjects.r(f"custom({penalityFn})")
blme.blmer(formula, data=data, REML=REML, control=lmc, contrasts=contrasts, **{cov.prior=covprior})

Answer 1

I figured it out.

from rpy2 import robjects 
from rpy2.robjects.packages import importr
blme = importr("blme")
lmc = robjects.r(f"lmerControl({control})")
r_string = robjects.r("""penalityFn <- function(sigma){dnorm(sigma, 1, 1)}""")
covprior = ("custom(penalityFn)")
blme.blmer(formula, data=data, REML=REML, control=lmc, contrasts=contrasts, **{cov.prior=covprior})