tugboat (version 0.1.4)

create: Create a Dockerfile

Description

This function will crawl all files in the current project/directory and (attempt to) detect all R packages and store these in a lockfile. From this lockfile, it will create a corresponding Dockerfile. It will also copy the full contents of the current directory/project into the Docker image. The directory in the Docker container containing the current directory contents will be /current-directory-name. For example if your analysis directory is named incredible_analysis, the corresponding location in the generated Docker image will be /incredible_analysis.

Usage

create(
  project = here::here(),
  as = file.path(project, "Dockerfile"),
  FROM = NULL,
  ...,
  exclude = NULL,
  verbose = FALSE,
  optimize_pak = TRUE
)

Value

The Dockerfile contained as a string vector. Each vector element corresponds to a line in the Dockerfile.

Arguments

project

The project directory. If no project directory is provided, by default, the here package will be used to determine the active project. If no project is currently active, then here defaults to the working directory where initially called.

as

The file path to write to. The default value is file.path(project, "Dockerfile").

FROM

Docker image to start FROM. Default is FROM r-base:R.version.

...

Additional arguments which are passed directly to renv::snapshot. Please see the documentation for that function for all relevant details.

exclude

A vector of strings specifying all paths (files or directories) that should NOT be included in the Docker image. By default, all files in the directory will be included. NOTE: the file and directory paths should be relative to the project directory. They do NOT need to be absolute paths.

verbose

A boolean indicating whether or not to print the resulting Dockerfile to the console. Default value is FALSE.

optimize_pak

A boolean indicating whether or not to try to optimize package installations with pak. Defaults to TRUE. This should rarely be changed from its default value. However, sometimes this optimization may cause build failures. When encountering a build error, a good first step can be to set optimize_pak = FALSE and see if the error persists.

See Also

here::here; this will be used by default to determine the current project directory.

renv::snapshot which this function relies on to find all R dependencies and create a corresponding lockfile.

Examples

if (FALSE) {
# Create a Dockerfile based on the rocker/rstudio image.
# Write the Dockerfile locally to here::here("Dockerfile").
# Copy all files except the /data and /examples directories.
dock <- create(
  project = here::here(),
  FROM = "rocker/rstudio",
  exclude = c("/data", "/examples")
)
}