logDensity 'y'

[kellijohnson-NOAA]

I do not know where 'y' comes from in the logDensity function:
avg.pos[j] = log(mean(dgamma(y[nonZeros[j]], shape = gamma.a, rate=gamma.b))) and when I try and run the function on the results from a model I get the following error:
Error in dgamma(y[nonZeros[j]], shape = gamma.a, rate = gamma.b) : object 'y' not found

[ericward-noaa]

That function wasn't updated during the last major code revision, but we need to do so. 'y' should be part of the data frame (basically the same as the estimation code).

[kellijohnson-NOAA]

I am pretty sure I was able to fix the issues being caused because of non-attached data, but I am wondering if the calcs should support ECE distributions? @ericward-noaa would it be okay if I change this

obj$likelihood == "gamma"

to

grepl("gamma", obj$likelihood)

and the same for the lognormal?

[ericward-noaa]

Yep -- totally fine with me, thanks for updating this!