Bug Report – MaxEnt() throws error when using args = "betamultiplier=2"

[Rimagination]

Dear developer team,

I encountered an error when trying to pass a regularization parameter to the function MaxEnt(). The following code triggers an error:

  model <- MaxEnt(x = env, p = occ, args = "betamultiplier=2")

Error message:
Error in strsplit(nodata, "=")[[1]] : subscript out of bounds

However, when I pass two or more arguments as a character vector, the function works properly. For example:
mod1 <- MaxEnt(
x = env,
p = occ,
args = c("jackknife=true", "responsecurves=true")
)

This suggests that the function is not handling a single string input to args correctly — possibly expecting it to always be a vector of length ≥ 2 or mismanaging nodata internally.
I hope this issue can be fixed in a future release to allow single-argument args as intended.
Let me know if you need a reproducible example or session info.

Best regards,
Liang

[billy34]

There is a rather strange block of code that provoke this error

x <- stats::na.omit(x)
nodata <- grep("nodata=", args, value=TRUE)
if (length(args) == 1) {
	nodata = as.numeric(strsplit(nodata, "=")[[1]][2])
	x[is.na(x)] <- nodata			
} else {
	x[is.na(x)] <- -9999 
}

As a workaround, you can add another arg with it's default value. It will force x[is.na(x)] <- -9999 that won't do anything as NAs where already removed by x <- stats::na.omit(x).

for example add verbose arg with it's default value false --> args = c("addsamplestobackground=false", "verbose=false")