Tidy posterior summaries from multiple MCMCs per model

Run multiple MCMCs on simulated datasets and return posterior summaries and the effective number of parameters for each run.

Usage

tar_jags_rep_summary(
  name,
  jags_files,
  parameters.to.save,
  data = list(),
  variables = NULL,
  summaries = NULL,
  summary_args = NULL,
  batches = 1L,
  reps = 1L,
  combine = TRUE,
  n.cluster = 1,
  n.chains = 3,
  n.iter = 2000,
  n.burnin = as.integer(n.iter/2),
  n.thin = 1,
  jags.module = c("glm", "dic"),
  inits = NULL,
  RNGname = c("Wichmann-Hill", "Marsaglia-Multicarry", "Super-Duper", "Mersenne-Twister"),
  jags.seed = NULL,
  stdout = NULL,
  stderr = NULL,
  progress.bar = "text",
  refresh = 0,
  tidy_eval = targets::tar_option_get("tidy_eval"),
  packages = targets::tar_option_get("packages"),
  library = targets::tar_option_get("library"),
  format = "qs",
  format_df = "fst_tbl",
  repository = targets::tar_option_get("repository"),
  error = targets::tar_option_get("error"),
  memory = "transient",
  garbage_collection = targets::tar_option_get("garbage_collection"),
  deployment = targets::tar_option_get("deployment"),
  priority = targets::tar_option_get("priority"),
  resources = targets::tar_option_get("resources"),
  storage = targets::tar_option_get("storage"),
  retrieval = targets::tar_option_get("retrieval"),
  cue = targets::tar_option_get("cue"),
  description = targets::tar_option_get("description")
)

Arguments

name

Symbol, base name for the collection of targets. Serves as a prefix for target names.

jags_files

Character vector of JAGS model files. If you supply multiple files, each model will run on the one shared dataset generated by the code in data. If you supply an unnamed vector, tools::file_path_sans_ext(basename(jags_files)) will be used as target name suffixes. If jags_files is a named vector, the suffixed will come from names(jags_files).

parameters.to.save

Model parameters to save, passed to R2jags::jags() or R2jags::jags.parallel(). See the argument documentation of the R2jags::jags() and R2jags::jags.parallel() help files for details.

data

Code to generate the data list for the JAGS model. Optionally include a .join_data element to join parts of the data to correspondingly named parameters in the summary output. See the vignettes for details.

variables

Character vector of model parameter names. The output posterior summaries are restricted to these variables.

summaries

List of summary functions passed to ... in posterior::summarize_draws() through $summary() on the CmdStanFit object.

summary_args

List of summary function arguments passed to .args in posterior::summarize_draws() through $summary() on the CmdStanFit object.

batches

Number of batches. Each batch runs a model reps times.

reps

Number of replications per batch. Ideally, each rep should produce its own random dataset using the code supplied to data.

combine

Logical, whether to create a target to combine all the model results into a single data frame downstream. Convenient, but duplicates data.

n.cluster

Number of parallel processes, passed to R2jags::jags() or R2jags::jags.parallel(). See the argument documentation of the R2jags::jags() and R2jags::jags.parallel() help files for details.

n.chains

Number of MCMC chains, passed to R2jags::jags() or R2jags::jags.parallel(). See the argument documentation of the R2jags::jags() and R2jags::jags.parallel() help files for details.

n.iter

Number if iterations (including warmup), passed to R2jags::jags() or R2jags::jags.parallel(). See the argument documentation of the R2jags::jags() and R2jags::jags.parallel() help files for details.

n.burnin

Number of warmup iterations, passed to R2jags::jags() or R2jags::jags.parallel(). See the argument documentation of the R2jags::jags() and R2jags::jags.parallel() help files for details.

n.thin

Thinning interval, passed to R2jags::jags() or R2jags::jags.parallel(). See the argument documentation of the R2jags::jags() and R2jags::jags.parallel() help files for details.

jags.module

Character vector of JAGS modules to load, passed to R2jags::jags() or R2jags::jags.parallel(). See the argument documentation of the R2jags::jags() and R2jags::jags.parallel() help files for details.

inits

Initial values of model parameters, passed to R2jags::jags() or R2jags::jags.parallel(). See the argument documentation of the R2jags::jags() and R2jags::jags.parallel() help files for details.

RNGname

Choice of random number generator, passed to R2jags::jags() or R2jags::jags.parallel(). See the argument documentation of the R2jags::jags() and R2jags::jags.parallel() help files for details.

jags.seed

The jags.seed argument of the tar_jags_rep*() functions is deprecated. See the "Seeds" section for details.

stdout

Character of length 1, file path to write the stdout stream of the model when it runs. Set to NULL to print to the console. Set to R.utils::nullfile() to suppress stdout. Does not apply to messages, warnings, or errors.

stderr

Character of length 1, file path to write the stderr stream of the model when it runs. Set to NULL to print to the console. Set to R.utils::nullfile() to suppress stderr. Does not apply to messages, warnings, or errors.

progress.bar

Type of progress bar, passed to R2jags::jags() or R2jags::jags.parallel(). See the argument documentation of the R2jags::jags() and R2jags::jags.parallel() help files for details.

refresh

Frequency for refreshing the progress bar, passed to R2jags::jags() or R2jags::jags.parallel(). See the argument documentation of the R2jags::jags() and R2jags::jags.parallel() help files for details.

tidy_eval

Logical, whether to enable tidy evaluation when interpreting command and pattern. If TRUE, you can use the "bang-bang" operator !! to programmatically insert the values of global objects.

packages

Character vector of packages to load right before the target runs or the output data is reloaded for downstream targets. Use tar_option_set() to set packages globally for all subsequent targets you define.

library

Character vector of library paths to try when loading packages.

format

Character of length 1, storage format of the data frames of posterior summaries and other data frames returned by targets. We recommend efficient data frame formats such as "feather" or "aws_parquet". For more on storage formats, see the help file of targets::tar_target().

format_df

Character of length 1, storage format of the data frame targets such as posterior draws. We recommend efficient data frame formats such as "feather" or "aws_parquet". For more on storage formats, see the help file of targets::tar_target().

repository

Character of length 1, remote repository for target storage. Choices:

"local": file system of the local machine.
"aws": Amazon Web Services (AWS) S3 bucket. Can be configured with a non-AWS S3 bucket using the endpoint argument of tar_resources_aws(), but versioning capabilities may be lost in doing so. See the cloud storage section of https://books.ropensci.org/targets/data.html for details for instructions.
"gcp": Google Cloud Platform storage bucket. See the cloud storage section of https://books.ropensci.org/targets/data.html for details for instructions.
A character string from tar_repository_cas() for content-addressable storage.

Note: if repository is not "local" and format is "file" then the target should create a single output file. That output file is uploaded to the cloud and tracked for changes where it exists in the cloud. The local file is deleted after the target runs.

error

Character of length 1, what to do if the target stops and throws an error. Options:

"stop": the whole pipeline stops and throws an error.
"continue": the whole pipeline keeps going.
"null": The errored target continues and returns NULL. The data hash is deliberately wrong so the target is not up to date for the next run of the pipeline. In addition, as of version 1.8.0.9011, a value of NULL is given to upstream dependencies with error = "null" if loading fails.
"abridge": any currently running targets keep running, but no new targets launch after that.
"trim": all currently running targets stay running. A queued target is allowed to start if:
1. It is not downstream of the error, and
2. It is not a sibling branch from the same tar_target() call (if the error happened in a dynamic branch).
The idea is to avoid starting any new work that the immediate error impacts. error = "trim" is just like error = "abridge", but it allows potentially healthy regions of the dependency graph to begin running. (Visit https://books.ropensci.org/targets/debugging.html to learn how to debug targets using saved workspaces.)

memory

Character of length 1, memory strategy. Possible values:

"auto": new in targets version 1.8.0.9011, memory = "auto" is equivalent to memory = "transient" for dynamic branching (a non-null pattern argument) and memory = "persistent" for targets that do not use dynamic branching.
"persistent": the target stays in memory until the end of the pipeline (unless storage is "worker", in which case targets unloads the value from memory right after storing it in order to avoid sending copious data over a network).
"transient": the target gets unloaded after every new target completes. Either way, the target gets automatically loaded into memory whenever another target needs the value.

For cloud-based dynamic files (e.g. format = "file" with repository = "aws"), the memory option applies to the temporary local copy of the file: "persistent" means it remains until the end of the pipeline and is then deleted, and "transient" means it gets deleted as soon as possible. The former conserves bandwidth, and the latter conserves local storage.

garbage_collection

Logical: TRUE to run base::gc() just before the target runs, FALSE to omit garbage collection. In the case of high-performance computing, gc() runs both locally and on the parallel worker. All this garbage collection is skipped if the actual target is skipped in the pipeline. Non-logical values of garbage_collection are converted to TRUE or FALSE using isTRUE(). In other words, non-logical values are converted FALSE. For example, garbage_collection = 2 is equivalent to garbage_collection = FALSE.

deployment

Character of length 1. If deployment is "main", then the target will run on the central controlling R process. Otherwise, if deployment is "worker" and you set up the pipeline with distributed/parallel computing, then the target runs on a parallel worker. For more on distributed/parallel computing in targets, please visit https://books.ropensci.org/targets/crew.html.

priority

Numeric of length 1 between 0 and 1. Controls which targets get deployed first when multiple competing targets are ready simultaneously. Targets with priorities closer to 1 get dispatched earlier (and polled earlier in tar_make_future()).

resources

Object returned by tar_resources() with optional settings for high-performance computing functionality, alternative data storage formats, and other optional capabilities of targets. See tar_resources() for details.

storage

Character string to control when the output of the target is saved to storage. Only relevant when using targets with parallel workers (https://books.ropensci.org/targets/crew.html). Must be one of the following values:

"main": the target's return value is sent back to the host machine and saved/uploaded locally.
"worker": the worker saves/uploads the value.
"none": targets makes no attempt to save the result of the target to storage in the location where targets expects it to be. Saving to storage is the responsibility of the user. Use with caution.

retrieval

Character string to control when the current target loads its dependencies into memory before running. (Here, a "dependency" is another target upstream that the current one depends on.) Only relevant when using targets with parallel workers (https://books.ropensci.org/targets/crew.html). Must be one of the following values:

"main": the target's dependencies are loaded on the host machine and sent to the worker before the target runs.
"worker": the worker loads the target's dependencies.
"none": targets makes no attempt to load its dependencies. With retrieval = "none", loading dependencies is the responsibility of the user. Use with caution.

cue

An optional object from tar_cue() to customize the rules that decide whether the target is up to date.

description

Character of length 1, a custom free-form human-readable text description of the target. Descriptions appear as target labels in functions like tar_manifest() and tar_visnetwork(), and they let you select subsets of targets for the names argument of functions like tar_make(). For example, tar_manifest(names = tar_described_as(starts_with("survival model"))) lists all the targets whose descriptions start with the character string "survival model".

Value

tar_jags_rep_summary() returns list of target objects. See the "Target objects" section for background. The target names use the name argument as a prefix, and the individual elements of jags_files appear in the suffixes where applicable. As an example, the specific target objects returned by tar_jags_rep_dic(name = x, jags_files = "y.jags") are as follows.

x_file_y: reproducibly track the JAGS model file. Returns a character vector of length 1 with the path to the JAGS model file.
x_lines_y: read the contents of the JAGS model file for safe transport to parallel workers. Returns a character vector of lines in the model file.
x_data: use dynamic branching to generate multiple JAGS datasets from the R expression in the data argument. Each dynamic branch returns a batch of JAGS data lists.
x_y: run JAGS on each dataset from x_data. Each dynamic branch returns a tidy data frame of summaries for each batch of data.
x: combine all the batches from x_y into a non-dynamic target. Suppressed if combine is FALSE. Returns a long tidy data frame with all summaries from all the branches of x_y.

Details

The MCMC targets use R2jags::jags() if n.cluster is 1 and R2jags::jags.parallel() otherwise. Most arguments to tar_jags() are forwarded to these functions.

Seeds

Rep-specific random number generator seeds for the data and models are automatically set based on the batch, rep, parent target name, and tar_option_get("seed"). This ensures the rep-specific seeds do not change when you change the batching configuration (e.g. 40 batches of 10 reps each vs 20 batches of 20 reps each). Each data seed is in the .seed list element of the output, and each JAGS seed is in the .seed column of each JAGS model output.

Target objects

Most stantargets functions are target factories, which means they return target objects or lists of target objects. Target objects represent skippable steps of the analysis pipeline as described at https://books.ropensci.org/targets/. Please read the walkthrough at https://books.ropensci.org/targets/walkthrough.html to understand the role of target objects in analysis pipelines.

For developers, https://wlandau.github.io/targetopia/contributing.html#target-factories explains target factories (functions like this one which generate targets) and the design specification at https://books.ropensci.org/targets-design/ details the structure and composition of target objects.

Examples

if (identical(Sys.getenv("TAR_JAGS_EXAMPLES"), "true")) {
targets::tar_dir({ # tar_dir() runs code from a temporary directory.
targets::tar_script({
library(jagstargets)
# Do not use a temp file for a real project
# or else your targets will always rerun.
tmp <- tempfile(pattern = "", fileext = ".jags")
tar_jags_example_file(tmp)
list(
  tar_jags_rep_summary(
    your_model,
    jags_files = tmp,
    data = tar_jags_example_data(),
    parameters.to.save = "beta",
    batches = 2,
    reps = 2,
    stdout = R.utils::nullfile(),
    stderr = R.utils::nullfile()
  )
)
}, ask = FALSE)
targets::tar_make()
})
}