taf.boot: Boot TAF Analysis

In TAF: Transparent Assessment Framework for Reproducible Research

Boot TAF Analysis

Description

Process metadata files ‘SOFTWARE.bib’ and ‘DATA.bib’ to set up software and data files required for the analysis.

Usage

taf.boot(software = TRUE, data = TRUE, clean = TRUE, force = FALSE,
  taf = NULL, quiet = FALSE)

Arguments

software	whether to process ⁠SOFTWARE.bib⁠.
data	whether to process ⁠DATA.bib⁠.
clean	whether to clean directories during the boot procedure.
force	whether to remove existing ⁠boot/data⁠, ⁠boot/library⁠, and ⁠boot/software⁠ directories before the boot procedure.
taf	a convenience flag where taf = TRUE sets software, data, clean, and force to TRUE, as is done on the TAF server. Any other value of taf is ignored.
quiet	whether to suppress messages reporting progress.

Details

If clean = TRUE then:

1. clean.software and clean.library are run if ‘SOFTWARE.bib’ is processed.

2. clean.data is run if ‘DATA.bib’ is processed.

The default behavior of taf.boot is to skip downloading of remote files (GitHub resources, URLs, boot scripts) and also skip installing R packages from GitHub if the files seem to be already in place. This is done to speed up the boot procedure as much as possible. To override this and guarantee that all data and software files are updated, pass force = TRUE to download and install everything declared in ⁠SOFTWARE.bib⁠ and ⁠DATA.bib⁠.

Value

Logical vector indicating which metadata files were processed.

Note

This function should be called from the top directory of a TAF analysis. It looks for a directory called ‘boot’ and prepares data files and software according to metadata specifications.

The boot procedure consists of the following steps:

1. If a ⁠boot/SOFTWARE.bib⁠ metadata file exists, it is processed.

2. If a ⁠boot/DATA.bib⁠ metadata file exists, it is processed.

After the boot procedure, software and data have been documented and are ready to be used in the subsequent analysis. Specifically, the procedure populates up to three new directories:

• ⁠boot/data⁠ with data files.

• ⁠boot/library⁠ with R packages compiled for the local platform.

• ⁠boot/software⁠ with software files, such as R packages in ⁠tar.gz⁠ source code format.

From version 4.2 onwards, the term boot is preferred for what used to be called bootstrap, mainly to avoid confusion with statistical bootstrap. To taf.boot() is similar to booting a computer, readying the components required for subsequent computations. Help pages now refer to boot, but all TAF functions fully support existing analyses that have a legacy bootstrap folder.

Model settings and configuration files can be set up within ⁠DATA.bib⁠, see TAF Wiki.

See Also

draft.data and draft.software can be used to create initial draft versions of ‘DATA.bib’ and ‘SOFTWARE.bib’ metadata files.

taf.library loads a package from ⁠boot/library⁠.

TAF-package gives an overview of the package.

Examples

## Not run: 
taf.boot()

## End(Not run)

TAF documentation built on March 31, 2023, 6:51 p.m.