Package 'sims'

Add Simulated Datasets

Description

Add Simulated Datasets

Usage

sims_add(path = ".", nsims = 1)

Arguments

path	A string specifying the path to the directory to add the data sets to.
nsims	A count of the number of additional datasets to generate.

Value

A character vector of the names of the files created.

---

Check Simulated Data

Description

Checks the simulated data argument values in the '.sims.rds' file.

Usage

sims_check(path = ".")

Arguments

path	A string of the path to the directory with the simulated data.

Details

The checks include whether number and names of the data files in the directory are consistent with the number of simulations.

Value

An informative error or invisible list of the argument values.

Examples

set.seed(10)
sims_simulate("a <- runif(1)",
  save = TRUE, path = tempdir(), exists = NA,
  ask = FALSE
)
(sims_check(tempdir()))

---

Copy Simulated Datasets

Description

Copy Simulated Datasets

Usage

sims_copy(
  path_from = ".",
  path_to = paste0(path_from, "_copy"),
  exists = FALSE,
  ask = getOption("sims.ask", TRUE),
  silent = FALSE
)

Arguments

path_from	A string of the path to the directory containing the simulated datasets.
path_to	A string of the path to the directory to copy the simulated dataset to.
exists	A flag specifying whether path_to should already exist. If exists = NA it doesn't matter. If the directory already exists sims compatible files are deleted if exists = TRUE or exists = NA otherwise an error is thrown.
ask	A flag specifying whether to ask before deleting files.
silent	A flag specifying whether to suppress warnings.

Value

A character vector of the names of the files copied.

---

Simulated Datasets

Description

Gets the simulated datasets as an nlist::nlists_object(). There is no guarantee that all the datasets will fit in memory.

Usage

sims_data(path = ".")

Arguments

path	A string of the path to the directory with the simulated data.

Value

An nlist::nlists_object() of the simulated datasets.

Examples

set.seed(10)
sims_simulate("a <- runif(1)",
  nsims = 10L, path = tempdir(),
  exists = NA, ask = FALSE
)
library(nlist)
sims_data(tempdir())

---

Simulated Data Files

Description

Gets the names of the simulated data files.

Usage

sims_data_files(path = ".")

Arguments

path	A string of the path to the directory with the simulated data.

Value

A character vector of the names of the simulated data files.

Examples

set.seed(10)
sims_simulate("a <- runif(1)",
  nsims = 10L, path = tempdir(),
  exists = NA, ask = FALSE
)
sims_data_files(tempdir())

---

Simulated Data Argument Values

Description

Gets the simulated data argument values in the '.sims.rds' file.

Usage

sims_info(path = ".")

Arguments

path	A string of the path to the directory with the simulated data.

Value

A named list of the values in file.path(path, '.sims.rds').

Examples

set.seed(10)
sims_simulate("a <- runif(1)", path = tempdir(), exists = NA, ask = FALSE)
sims_info(tempdir())

---

Sims Random R Distributions

Description

Gets the names of the R random variate generating functions listed in Distributions().

Usage

sims_rdists()

Value

A character vector.

Examples

sims_rdists()

---

Simulate Datasets

Description

Simulates datasets using JAGS or R code. By default returns the datasets as an nlist::nlists_object(). If path is provided then the datasets are written to the directory as individual .rds files.

Usage

sims_simulate(
  code,
  constants = nlist::nlist(),
  parameters = nlist::nlist(),
  monitor = ".*",
  stochastic = NA,
  latent = NA,
  nsims = 1,
  save = FALSE,
  path = ".",
  exists = FALSE,
  rdists = sims_rdists(),
  ask = getOption("sims.ask", TRUE),
  silent = FALSE
)

Arguments

code	A string of the JAGS or R code to generate the data. The JAGS code must not be in a data or model block.
constants	An nlist object (or list that can be coerced to nlist) specifying the values of nodes in code. The values are included in the output dataset.
parameters	An nlist object (or list that can be coerced to nlist) specifying the values of nodes in code. The values are not included in the output dataset.
monitor	A character vector (or regular expression if a string) specifying the names of the nodes in code to include in the dataset. By default all nodes are included.
stochastic	A logical scalar specifying whether to monitor deterministic and stochastic (NA), only deterministic (FALSE) or only stochastic nodes (TRUE).
latent	A logical scalar specifying whether to monitor observed and latent (NA), only latent (TRUE) or only observed nodes (FALSE).
nsims	A whole number between 1 and 1,000,000 specifying the number of data sets to simulate. By default 1 data set is simulated.
save	A flag specifying whether to return the data sets as an nlists object or save in path. If save = NA the datasets are returned as an nlists object and saved in path.
path	A string specifying the path to the directory to save the data sets in.
exists	A flag specifying whether the path directory should already exist (if exists = NA it doesn't matter).
rdists	A character vector specifying the R functions to recognize as stochastic.
ask	A flag specifying whether to ask before deleting sims compatible files.
silent	A flag specifying whether to suppress warnings.

Details

JAGS code is identified by the presence of '~' indicating a stochastic variable node. Otherwise code is assumed to be R code and stochastic variable nodes are those where assignment is immediately succeeded by a call to one of the functions named in rdists.

Both constants and parameters must be ⁠[nlist::nlist_object]s⁠ (or lists that can be coerced to such) . The only difference between constants and parameters is that the values in constants are appended to the output data while the values in parameters are not. Neither constants or parameters can include missing values nor can they have elements with the same name. Elements which are not in code are dropped with a warning (unless silent = TRUE in which case the warning is suppressed).

Each set of simulated data set is written as a separate .rds file. The files are labelled data0000001.rds, data0000002.rds, data0000003.rds etc. The argument values are saved in the hidden file .sims.rds.

sims compatible files are those matching the regular expression ⁠^((data\\\\d\{7,7\})|([.]sims))[.]rds$⁠.

Parallelization is implemented using the future package.

Value

By default an nlist::nlists_object() of the simulated data. Otherwise if path is defined saves the datasets as individual .rds files and returns TRUE.

See Also

sims_rdists()

Examples

set.seed(101)
sims_simulate("a <- runif(1)", path = tempdir(), exists = NA, ask = FALSE)

---