R/various_defaults_func.R
In ks905383/quantproj: Project Climate Data Based on Changes in Distribution Shape
Defines functions
set.defaults
Documented in
set.defaults
#' Set default inputs for quantile mapping
#'
#' \code{set.defaults} creates a list object that stores the default
#' input parameters for the package \code{quantproj}, to be
#' used as an input for most other functions. It also creates the
#' directories set as defaults (and their subdirectories used in
#' various bits of code in this package) if they do not yet exist.
#'
#' @section Directory structure:
#' Generally, this package uses three main directories (though
#' \code{aux.dir} may be the same as either of the data directories,
#' the user is discouraged from setting the two data directories equal
#' or using pre-existing directories, for file dependency reasons):
#' \describe{
#' \item{\code{aux_dir}}{a directory containing two sub-directories,
#' \code{bases} for saved spline basis functions (the
#' predictors in the quantile regression), and \code{run_logs}
#' for logging progress on the bigger scripts}
#' \item{\code{base.data.dir}}{the directory in which to place the raw data to
#' be projected. Contains a sub-directory, \code{output}, for the resultant
#' projections.}
#' \item{\code{mod.data.dir}}{the directory in which to place the model data
#' whose distributional changes are used to project the data in
#' \code{base.data.dir}. Contains a sub-directory, \code{params}, in which the
#' quantile fit parameters are stored.}
#' }
#'
#' @section Note on file structure:
#' This package assumes all files follow CMIP5 standard file naming
#' conventions; in other words, it assumes that variable, data frequency, model
#' name, and file year range are all encoded in the filenames themselves (see
#' the
#' \href{https://cmip.llnl.gov/cmip5/docs/cmip5_data_reference_syntax.pdf}{CMIP5
#' syntax guide} for more information). The filename format is:
#' \code{[variable]_[frequency]_[model]_[experiment]_[run/ensemble
#' member]_[timeframe](_[suffix])}
#'
#' @param base.data.dir the directory with raw data to be projected
#' @param mod.data.dir the directory with raw model ensemble data used to project
#' @param aux.dir directory for auxiliary files - saved basis functions, etc.
#'
#' @param fn.suffix optional add-on to all output filenames
#'
#' @param lat.clip,lon.clip if desired, a \code{c(min,max)} vector giving bounds for a lon/lat box; only data within this box will be loaded and processed
#'
#' @param filevar variable shorthand (CMIP5 syntax, def: \code{"tas"} for near-surface air temperature)
#' @param freq data frequency shorthand (CMIP5 syntax, def: \code{"day"} for daily)
#' @param mod.name name of model ensemble (def: \code{"LENS"})
#' @param base.name name of data product to be projected (def: \code{"ERA-INTERIM"})
#'
#' @param base.year.range desired time frame of the "base" time period to project (def: \code{c(1979,2010)})
#' @param mod.year.range desired time frame of the time frame to be processed (quantiles calculated) for the model ensmeble; make sure this range includes \code{base.year.range} (def: \code{c(1979,2099)})
#' @param proj.year.range desired time frame to project to (def: \code{c(2011,2099)})
#'
#' @param q_norm normalizing quantiles (must be a \code{3 x 1} vector, i.e. \code{c(0.1,0.5,0.9)})
#' @param q_bulk bulk quantiles (must be a \code{[n_bulk_quantiles x 1]} vector, i.e. \code{c(0.25,0.3,0.5,0.6,0.75)})
#' @param q_tail tail quantiles (must be a \code{[n_bulk_quantiles x 1]} vector, i.e. \code{c(0.5,0.9,0.95)}). These are fractions of the difference between \code{max(q_bulk)} and \code{0}/\code{1}!
#' @param norm.x.df,bulk.x.df,tail.x.df degrees of freedom for normalizing, bulk, and tail quantile regressions, in a \code{3 x 1} vector for \code{c([seasonal],[long-term],[interaction])}
#' @param get.volc include a volcanic predictor in the normalization (def: \code{FALSE})
#'
#' @param nboots number of bootstrap runs (def: \code{50})
#' @param block.size bootstrap block size (def: \code{40}) (ACROSS RUNS - SO 20 means each bootstrap block is across 20 randomly chosen runs)
#'
#' @param varnames a list with each element giving the possible names for the
#' variable with the same name as the list element. For example, the list
#' element \code{lat=c("lat","latitude","Latititude","latitude_1")} makes
#' it so that the latitude variable in the NetCDF files loaded through
#' \code{\link{get.ncdf}} can be named any of those names without the code
#' running into trouble.
#' @param search.str a search string used to look for the source NetCDF files
#' (def: "\code{[filevar]_day_.*nc}", with \code{[filevar]} set above)
#'
#'
#' @return a list object containing all of the parameters listed above, to be used
#' as an input in thd other functions in this package.
set.defaults <- function(
#----- FILE PATHS -----
# The directory in which code is stored
aux.dir=paste0(getwd(),"/aux/"),
# The directory with weather data to be projected
base.data.dir=paste0(getwd(),"/base_data/"),
# The directory with raw model ensemble data
mod.data.dir=paste0(getwd(),"/model_data/"),
# A custom suffix for output files
fn.suffix=character(),
#----- VARIABLE -----
# Variable shorthand (use CMIP5)
filevar="tas",
#----- MODEL NAMES AND IDENTIFIERS -----
# Name of the model ensmeble
mod.name="LENS",
# Name of the weather product
base.name="ERA-INTERIM",
#------ GEOGRAPHIC PARAMETERS -----
# Set to only process a subset of locations
lat.clip=numeric(),
lon.clip=numeric(),
#------ QUANTILE MAPPING PARAMETERS -----
# Base data process year range (this is the desired
# length of the processing - if the reanalysis files
# spread farther than this range, only a subset will
# be used)
base.year.range=c(1979,2010),
# Model process year range (this is the desired
# length of the processing, as above. Make sure this
# range includes the base.year.range above!)
mod.year.range=c(1979,2099),
# Years to project to
proj.year.range=c(2011,2099),
# Normalizing quantiles
q_norm=c(0.1,0.5,0.9),
# Normalizing degrees of freedom (model)
norm.x.df=c(14,6,3),
# Normalizing degrees of freedom (weather)
base.norm.x.df=c(10,1,0),
# Include a volcanic predictor in the normalization
get.volc=F,
# Bulk quantiles
q_bulk=c(0.10,0.18,0.25,0.35,0.42,0.50,0.58,0.65,0.75,0.82,0.90),
# Bulk degrees of freedom
bulk.x.df=c(14,6,3),
# Tail quantiles
q_tail=c(0.01,0.10,0.25,0.50,0.75),
# Tail degrees of freedom
tail.x.df=c(3,1,0),
#------ PROJECTION OPTIONS -----
# How to determine 'base' days for projection
index.type="resampling_rep",
resampling.timescale="year",
#------ BOOTSTRAPPING -----
bootstrapping=F,
nboots=50,
block.size=40,
#----- FILENAME/VARNAME ISSUES -----
varnames=list(lon=c("lon","longitude","Longitude","longitude_1"),
lat=c("lat","latitude","Latititude","latitude_1"),
time=c("time"),
run=c("run"),
loc=c("global_loc","loc","location","lonlat")),
search.str=character()
) {
# Warning if the years don't overlap.
if (min(base.year.range)<min(mod.year.range) || max(base.year.range)>max(mod.year.range)) {
warning(paste0('The year range chosen for the model processing (',
paste0(mod.year.range,collapse='-'),') does not ',
'include the years chosen for the weather data processing (',
paste0(base.year.range,collapse='-'),'), please reconsider.'))
}
# Set filename search string
if (length(search.str)==0) {
search.str <- paste0(filevar,"_day_.*nc")
}
# Throw a warning if the model name contains a . or a /
if (grepl("\\.|\\_",mod.name)) {
warning(paste0("The name of the projecting model (ensemble), ",
mod.name,
" contains a '.' or a '/'. Please avoid this (yes, rename your files if you have to), since it tends to mess up the file identification procedures."))
}
if (grepl("\\.|\\_",base.name)) {
warning(paste0("The name of the base dataset/model, ",
base.name,
" contains a '.' or a '/'. Please avoid this (yes, rename your files if you have to), since it tends to mess up the file identification procedures."))
}
# Create directories if they don't yet exist ---------
# Auxiliary directory and subdirectories
if (!dir.exists(aux.dir)) {dir.create(aux.dir); cat(paste0("\n",aux.dir," created for auxiliary files!"),fill=T)} else {cat(paste0("\n",aux.dir," set as location of auxiliary files!"),fill=T)}
if (!dir.exists(paste0(aux.dir,"bases/"))) {dir.create(paste0(aux.dir,"bases/")); cat(paste0(aux.dir,"bases/","created for predictor basis files!"),fill=T)} else {cat(paste0(aux.dir,"bases/ set as location of predictor basis files!"),fill=T)}
if (!dir.exists(paste0(aux.dir,"run_logs/"))) {dir.create(paste0(aux.dir,"run_logs/")); cat(paste0(aux.dir,"run_logs/ created for run logs!"),fill=T)} else {cat(paste0(aux.dir,"run_logs/ set as location of run logs!"),fill=T)}
# Base data directory
if (!dir.exists(base.data.dir)) {dir.create(base.data.dir); cat(paste0("\n",base.data.dir," created for data to be projected! Please make sure the only files in here that match the search string '",search.str,"' are the base data files you want to project."),fill=T)
} else {
cat(paste0("\n",base.data.dir," set as location of the data to be projected! Please make sure the only files in here that match the search string '",search.str,"' are the base data files you want to project."),fill=T)
}
if (!dir.exists(paste0(base.data.dir,"output/"))) {dir.create(paste0(base.data.dir,"output/")); cat(paste0(base.data.dir,"output/ created for the outputted, projected data!"),fill=T)} else {cat(paste0(base.data.dir,"output/ set as location for the outputted, projected data!"),fill=T)}
# Model/projecting data directory
if (!dir.exists(mod.data.dir)) {dir.create(mod.data.dir); cat(paste0("\n",mod.data.dir," created for projecting data! Please make sure the only files in here that match the search string '",search.str,"' are the model data files you want to use for projecting."),fill=T)
} else {
cat(paste0("\n",mod.data.dir," set as location for projecting data! Please make sure the only files in here that match the search string '",search.str,"' are the model data files you want to use for projecting."),fill=T)
}
if (!dir.exists(paste0(mod.data.dir,"params/"))) {dir.create(paste0(mod.data.dir,"params/")); cat(paste0(mod.data.dir,"params/ created for projecting data quantile fits!"),fill=T)} else {cat(paste0(mod.data.dir,"params/ set as location for projecting data quantile fits!"),fill=T)}
# Save -----------------------------------------------
defaults <- list()
for (var.idx in ls()[ls()!="defaults"]) {
defaults[[var.idx]] <- eval(parse(text=var.idx))
}
rm(var.idx)
return(defaults)
}
ks905383/quantproj documentation built on Nov. 1, 2020, 9:12 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Defines functions set.defaults
Documented in set.defaults
#' Set default inputs for quantile mapping
#'
#' \code{set.defaults} creates a list object that stores the default
#' input parameters for the package \code{quantproj}, to be
#' used as an input for most other functions. It also creates the
#' directories set as defaults (and their subdirectories used in
#' various bits of code in this package) if they do not yet exist.
#'
#' @section Directory structure:
#' Generally, this package uses three main directories (though
#' \code{aux.dir} may be the same as either of the data directories,
#' the user is discouraged from setting the two data directories equal
#' or using pre-existing directories, for file dependency reasons):
#' \describe{
#' \item{\code{aux_dir}}{a directory containing two sub-directories,
#' \code{bases} for saved spline basis functions (the
#' predictors in the quantile regression), and \code{run_logs}
#' for logging progress on the bigger scripts}
#' \item{\code{base.data.dir}}{the directory in which to place the raw data to
#' be projected. Contains a sub-directory, \code{output}, for the resultant
#' projections.}
#' \item{\code{mod.data.dir}}{the directory in which to place the model data
#' whose distributional changes are used to project the data in
#' \code{base.data.dir}. Contains a sub-directory, \code{params}, in which the
#' quantile fit parameters are stored.}
#' }
#'
#' @section Note on file structure:
#' This package assumes all files follow CMIP5 standard file naming
#' conventions; in other words, it assumes that variable, data frequency, model
#' name, and file year range are all encoded in the filenames themselves (see
#' the
#' \href{https://cmip.llnl.gov/cmip5/docs/cmip5_data_reference_syntax.pdf}{CMIP5
#' syntax guide} for more information). The filename format is:
#' \code{[variable]_[frequency]_[model]_[experiment]_[run/ensemble
#' member]_[timeframe](_[suffix])}
#'
#' @param base.data.dir the directory with raw data to be projected
#' @param mod.data.dir the directory with raw model ensemble data used to project
#' @param aux.dir directory for auxiliary files - saved basis functions, etc.
#'
#' @param fn.suffix optional add-on to all output filenames
#'
#' @param lat.clip,lon.clip if desired, a \code{c(min,max)} vector giving bounds for a lon/lat box; only data within this box will be loaded and processed
#'
#' @param filevar variable shorthand (CMIP5 syntax, def: \code{"tas"} for near-surface air temperature)
#' @param freq data frequency shorthand (CMIP5 syntax, def: \code{"day"} for daily)
#' @param mod.name name of model ensemble (def: \code{"LENS"})
#' @param base.name name of data product to be projected (def: \code{"ERA-INTERIM"})
#'
#' @param base.year.range desired time frame of the "base" time period to project (def: \code{c(1979,2010)})
#' @param mod.year.range desired time frame of the time frame to be processed (quantiles calculated) for the model ensmeble; make sure this range includes \code{base.year.range} (def: \code{c(1979,2099)})
#' @param proj.year.range desired time frame to project to (def: \code{c(2011,2099)})
#'
#' @param q_norm normalizing quantiles (must be a \code{3 x 1} vector, i.e. \code{c(0.1,0.5,0.9)})
#' @param q_bulk bulk quantiles (must be a \code{[n_bulk_quantiles x 1]} vector, i.e. \code{c(0.25,0.3,0.5,0.6,0.75)})
#' @param q_tail tail quantiles (must be a \code{[n_bulk_quantiles x 1]} vector, i.e. \code{c(0.5,0.9,0.95)}). These are fractions of the difference between \code{max(q_bulk)} and \code{0}/\code{1}!
#' @param norm.x.df,bulk.x.df,tail.x.df degrees of freedom for normalizing, bulk, and tail quantile regressions, in a \code{3 x 1} vector for \code{c([seasonal],[long-term],[interaction])}
#' @param get.volc include a volcanic predictor in the normalization (def: \code{FALSE})
#'
#' @param nboots number of bootstrap runs (def: \code{50})
#' @param block.size bootstrap block size (def: \code{40}) (ACROSS RUNS - SO 20 means each bootstrap block is across 20 randomly chosen runs)
#'
#' @param varnames a list with each element giving the possible names for the
#' variable with the same name as the list element. For example, the list
#' element \code{lat=c("lat","latitude","Latititude","latitude_1")} makes
#' it so that the latitude variable in the NetCDF files loaded through
#' \code{\link{get.ncdf}} can be named any of those names without the code
#' running into trouble.
#' @param search.str a search string used to look for the source NetCDF files
#' (def: "\code{[filevar]_day_.*nc}", with \code{[filevar]} set above)
#'
#'
#' @return a list object containing all of the parameters listed above, to be used
#' as an input in thd other functions in this package.
set.defaults <- function(
#----- FILE PATHS -----
# The directory in which code is stored
aux.dir=paste0(getwd(),"/aux/"),
# The directory with weather data to be projected
base.data.dir=paste0(getwd(),"/base_data/"),
# The directory with raw model ensemble data
mod.data.dir=paste0(getwd(),"/model_data/"),
# A custom suffix for output files
fn.suffix=character(),
#----- VARIABLE -----
# Variable shorthand (use CMIP5)
filevar="tas",
#----- MODEL NAMES AND IDENTIFIERS -----
# Name of the model ensmeble
mod.name="LENS",
# Name of the weather product
base.name="ERA-INTERIM",
#------ GEOGRAPHIC PARAMETERS -----
# Set to only process a subset of locations
lat.clip=numeric(),
lon.clip=numeric(),
#------ QUANTILE MAPPING PARAMETERS -----
# Base data process year range (this is the desired
# length of the processing - if the reanalysis files
# spread farther than this range, only a subset will
# be used)
base.year.range=c(1979,2010),
# Model process year range (this is the desired
# length of the processing, as above. Make sure this
# range includes the base.year.range above!)
mod.year.range=c(1979,2099),
# Years to project to
proj.year.range=c(2011,2099),
# Normalizing quantiles
q_norm=c(0.1,0.5,0.9),
# Normalizing degrees of freedom (model)
norm.x.df=c(14,6,3),
# Normalizing degrees of freedom (weather)
base.norm.x.df=c(10,1,0),
# Include a volcanic predictor in the normalization
get.volc=F,
# Bulk quantiles
q_bulk=c(0.10,0.18,0.25,0.35,0.42,0.50,0.58,0.65,0.75,0.82,0.90),
# Bulk degrees of freedom
bulk.x.df=c(14,6,3),
# Tail quantiles
q_tail=c(0.01,0.10,0.25,0.50,0.75),
# Tail degrees of freedom
tail.x.df=c(3,1,0),
#------ PROJECTION OPTIONS -----
# How to determine 'base' days for projection
index.type="resampling_rep",
resampling.timescale="year",
#------ BOOTSTRAPPING -----
bootstrapping=F,
nboots=50,
block.size=40,
#----- FILENAME/VARNAME ISSUES -----
varnames=list(lon=c("lon","longitude","Longitude","longitude_1"),
lat=c("lat","latitude","Latititude","latitude_1"),
time=c("time"),
run=c("run"),
loc=c("global_loc","loc","location","lonlat")),
search.str=character()
) {
# Warning if the years don't overlap.
if (min(base.year.range)<min(mod.year.range) || max(base.year.range)>max(mod.year.range)) {
warning(paste0('The year range chosen for the model processing (',
paste0(mod.year.range,collapse='-'),') does not ',
'include the years chosen for the weather data processing (',
paste0(base.year.range,collapse='-'),'), please reconsider.'))
}
# Set filename search string
if (length(search.str)==0) {
search.str <- paste0(filevar,"_day_.*nc")
}
# Throw a warning if the model name contains a . or a /
if (grepl("\\.|\\_",mod.name)) {
warning(paste0("The name of the projecting model (ensemble), ",
mod.name,
" contains a '.' or a '/'. Please avoid this (yes, rename your files if you have to), since it tends to mess up the file identification procedures."))
}
if (grepl("\\.|\\_",base.name)) {
warning(paste0("The name of the base dataset/model, ",
base.name,
" contains a '.' or a '/'. Please avoid this (yes, rename your files if you have to), since it tends to mess up the file identification procedures."))
}
# Create directories if they don't yet exist ---------
# Auxiliary directory and subdirectories
if (!dir.exists(aux.dir)) {dir.create(aux.dir); cat(paste0("\n",aux.dir," created for auxiliary files!"),fill=T)} else {cat(paste0("\n",aux.dir," set as location of auxiliary files!"),fill=T)}
if (!dir.exists(paste0(aux.dir,"bases/"))) {dir.create(paste0(aux.dir,"bases/")); cat(paste0(aux.dir,"bases/","created for predictor basis files!"),fill=T)} else {cat(paste0(aux.dir,"bases/ set as location of predictor basis files!"),fill=T)}
if (!dir.exists(paste0(aux.dir,"run_logs/"))) {dir.create(paste0(aux.dir,"run_logs/")); cat(paste0(aux.dir,"run_logs/ created for run logs!"),fill=T)} else {cat(paste0(aux.dir,"run_logs/ set as location of run logs!"),fill=T)}
# Base data directory
if (!dir.exists(base.data.dir)) {dir.create(base.data.dir); cat(paste0("\n",base.data.dir," created for data to be projected! Please make sure the only files in here that match the search string '",search.str,"' are the base data files you want to project."),fill=T)
} else {
cat(paste0("\n",base.data.dir," set as location of the data to be projected! Please make sure the only files in here that match the search string '",search.str,"' are the base data files you want to project."),fill=T)
}
if (!dir.exists(paste0(base.data.dir,"output/"))) {dir.create(paste0(base.data.dir,"output/")); cat(paste0(base.data.dir,"output/ created for the outputted, projected data!"),fill=T)} else {cat(paste0(base.data.dir,"output/ set as location for the outputted, projected data!"),fill=T)}
# Model/projecting data directory
if (!dir.exists(mod.data.dir)) {dir.create(mod.data.dir); cat(paste0("\n",mod.data.dir," created for projecting data! Please make sure the only files in here that match the search string '",search.str,"' are the model data files you want to use for projecting."),fill=T)
} else {
cat(paste0("\n",mod.data.dir," set as location for projecting data! Please make sure the only files in here that match the search string '",search.str,"' are the model data files you want to use for projecting."),fill=T)
}
if (!dir.exists(paste0(mod.data.dir,"params/"))) {dir.create(paste0(mod.data.dir,"params/")); cat(paste0(mod.data.dir,"params/ created for projecting data quantile fits!"),fill=T)} else {cat(paste0(mod.data.dir,"params/ set as location for projecting data quantile fits!"),fill=T)}
# Save -----------------------------------------------
defaults <- list()
for (var.idx in ls()[ls()!="defaults"]) {
defaults[[var.idx]] <- eval(parse(text=var.idx))
}
rm(var.idx)
return(defaults)
}
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.