在GitHub上下载的代码怎么在R中运行，需要配什么环境？

求问在GitHub上下载的代码怎么在R中运行，需要配什么环境？
这些都是什么意思，也看不懂，小菜一枚，求大咖解答
#' Fit a single forecasting model.
#' @param model A forecasting model; see ?maple_models for more details.
#' @param deaths Deaths matrix; see ?maple for more details.
#' @param population Population matrix; see ?maple for more details.
#' @param forecast.horizon The number of years to produce projections for.
#' @param num.threads The number of threads passed to INLA; see ?maple for more details.
#' @return If the model is fitted using INLA, an INLA object. See ?INLA::inla for more details.
#' For Lee Carter models, a list is returned with the following entries
#' \describe{
#' \item{logrates}{The logarithm of the fitted death rates.}
#' \item{rates}{The fitted death rates.}
#' \item{alphas}{The intercepts.}
#' \item{betas}{The age-specific components in the Lee Carter model. This is a matrix whose columns correspond to age groups and rows to principal components.}
#' \item{gammas}{The time-varying components in the Lee Carter model. This is a matrix whose columns correspond to years and rows to principal components.}
#' \item{gammas.fit}{A list of length equal to the number of principal components. Each item in the list is the random walk fitted as an Arima object; see ?stats::arima for more details.}
#' \item{gammas.pred}{The extrapolated values of gammas.fit used for projections.}
#' \item{pct.var}{The proportion of variance explained by the principal component decomposition.}
#' \item{num.pcs}{The number of principal components used.}
#' \item{forecast.horizon}{The number of years forecast ahead.}
#' }
#' @examples
#' maple_fit_model(maple_models()[[1]], deaths = maple.deaths,
#' population = maple.population, forecast.horizon = 20)
#' @export
maple_fit_model <- function(model, deaths, population, forecast.horizon, num.threads) {
UseMethod("maple_fit_model")
}

#' @export
maple_fit_model.inla.model <- function(model, deaths, population, forecast.horizon,
num.threads = inla.getOption("num.threads"), ...) {
inla.dat <- prep_inla_dat(deaths = deaths,
population = population,
model = model,
forecast.horizon = forecast.horizon)
lhd.wght <- NULL
if (!is.null(model$likelihood.weight.rate)) {
inla.setOption(enable.inla.argument.weights = TRUE)
lhd.wght <- calc_likelihood_weights(inla.dat$year, model$likelihood.weight.rate)
}
fit <- inla(model$fml,
family = "poisson",
data = inla.dat,
E = population,
control.predictor = list(link = 1),
control.compute = list(dic = TRUE, config = TRUE),
weights = lhd.wght,
num.threads = num.threads
)
if (!is.null(model$likelihood.weight.rate)) {
inla.setOption(enable.inla.argument.weights = FALSE)
}
fit
}

#' @export
maple_fit_model.lc.model <- function(model, deaths, population, forecast.horizon, ...) {
ages <- as.numeric(rownames(deaths))
years <- as.numeric(colnames(deaths))
forecast.years <- ncol(deaths) + seq(forecast.horizon)

rates <- deaths / population

if (any(is.na(rates))) stop("Cannot run Lee-Carter model beacause of NAs in data.")

logrates <- log(rates)
alphas <- rowMeans(logrates)
clogrates <- logrates - alphas
clogrates.svd <- svd(clogrates)
betas <- t(clogrates.svd$u[, seq_len(model$num.pcs)])
gammas <- betas %*% clogrates
# Normalise so that each beta sums to 1
gammas <- gammas * rowSums(betas)
betas <- betas / rowSums(betas)
gammas.fit <- apply(gammas, 1, arima, order = c(0, 1, 0), xreg = seq_len(ncol(gammas)))
# In-sample fits
logrates.fit <- alphas + Reduce(`+`, lapply(seq_len(nrow(gammas)),
                                            function(i) outer(betas[i, ], gammas[i, ])))
# Forecasts
gammas.pred <- lapply(gammas.fit, predict,
                      newxreg = ncol(gammas) + seq(forecast.horizon),
                      n.ahead = forecast.horizon)
gammas.pred.mean <- lapply(gammas.pred, function(x) as.numeric(x$pred))
gammas.pred.se <- lapply(gammas.pred, function(x) as.numeric(x$se))
logrates.pred <- alphas + Reduce(`+`, lapply(seq_along(gammas.pred.mean),
                                             function(i) outer(betas[i, ], gammas.pred.mean[[i]])))

logfit <- cbind(logrates.fit, logrates.pred)
rownames(logfit) <- ages
colnames(logfit) <- c(years, years[length(years)] + seq(forecast.horizon))
fit <- exp(logfit)

l <- structure(list(
        logrates = logfit,
        rates = fit,
        alphas = alphas,
        betas = betas,
        gammas = gammas,
        gammas.fit = gammas.fit,
        gammas.pred = gammas.pred,
        pct.var = sum(clogrates.svd$d[seq_len(model$num.pcs)] ^ 2 / sum(clogrates.svd$d ^ 2)),
        num.pcs = model$num.pcs,
        forecast.horizon = forecast.horizon
    ), class = "lc")
l

}