It took me a long time to figure out how to replicate the forecast from stockassessment.org exactly on a pc. Maybe it is possible to include the line RNGkind(sample.kind = "Rounding") to the standard forecast script to make it robust to both pc and stockassessment.org. But an alternative is to allow users to download the forecast object directly from stockassessment.org.
I have adapted the fitfromweb function to do this (below). Is this something you would consider including in the package?
forecastfromweb <- function (stockname, character.only = FALSE, return.all = FALSE)
{
if (!character.only)
stockname <- as.character(substitute(stockname))
con <- url(sub("SN", stockname, "https://stockassessment.org/datadisk/stockassessment/userdirs/user3/SN/run/forecast.RData"))
e <- new.env()
nam <- load(con, e)
close(con)
if (return.all)
return(e)
hasFC <- grepl("^FC$", nam)
if (!any(hasFC)) {
warning(sprintf("No object named FC. Found %s in model.RData.",
paste0(nam, collapse = ", ")))
return(NA)
}
else if (length(nam) > 1) {
message(sprintf("Not returning %s from model.RData",
paste0(nam[!hasFC], collapse = ", ")))
}
return(e$FC)
}
It took me a long time to figure out how to replicate the forecast from stockassessment.org exactly on a pc. Maybe it is possible to include the line
RNGkind(sample.kind = "Rounding")to the standard forecast script to make it robust to both pc and stockassessment.org. But an alternative is to allow users to download the forecast object directly from stockassessment.org.I have adapted the fitfromweb function to do this (below). Is this something you would consider including in the package?