evaluate() uses model even if x is missing about dismo

[rhijmans]

If you look at the first example, you see that evaluate only uses arguments p and a, both a vector.

[stragu]

Looking at the function definition, I guess I hit the case when x is missing and a and p are not vectors. In this particular case, it goes to the else branch and will use the model to predict. Hence my question.

library("dismo")
# with matrices instead
p <- matrix(rnorm(50), ncol = 2)
a <- matrix(rnorm(50), ncol = 2)
e <- evaluate(p=p, a=a)

Error in predict(model, data.frame(p), ...) :
argument "model" is missing, with no default

I guess the rest of the documentation does hint at that, but the description of the model argument got me confused as I took it on its own. Should it say "Not used when x is missing (and both a and p are vectors)"?

[rhijmans]

Thanks for the clarification. I regret the way I designed this function. It is too complex, it should have been at least two functions. (even if it is not as bad as stats::reshape). I have updated the docs per your suggestion.

[stragu]

No problem, thank you for the prompt reply! And thank you for all your work for R Spatial 👍 :)