Create a Dockerfile based on either a sessionInfo, a workspace or a file.
dockerfile(from = utils::sessionInfo(),
image = getImageForVersion(getRVersionTag(from)),
maintainer = Sys.info()[["user"]], save_image = FALSE,
envir = .GlobalEnv, env = list(generator = paste("containerit",
utils::packageVersion("containerit"))), soft = FALSE,
offline = FALSE, copy = NULL, container_workdir = "/payload/",
cmd = "R", entrypoint = NULL, add_self = FALSE, silent = FALSE,
predetect = TRUE, versioned_libs = FALSE,
versioned_packages = FALSE, filter_baseimage_pkgs = FALSE)The source of the information to construct the Dockerfile. Can be a sessionInfo object, a path to a file within the working direcotry, a DESCRIPTION file, or the path to a workspace). If NULL then no automatic derivation of dependencies happens. If a DESCRIPTION file, then the minimum R version (e.g. "R (3.3.0)") is used for the image version and all "Imports" are explicitly installed.
('>From-object or character) Specifes the image that shall be used for the Docker container (FROM instruction).
By default, the image selection is based on the given session. Alternatively, use getImageForVersion(..) to get an existing image for a manually defined version of R, matching the version with tags from the base image rocker/r-ver (see details about the rocker/r-ver at 'https://hub.docker.com/r/rocker/r-ver/'). Or provide a correct image name yourself.
Specify the maintainer of the Dockerfile. See documentation at 'https://docs.docker.com/engine/reference/builder/#maintainer'. Defaults to Sys.info()[["user"]]. Can be removed with NULL.
When TRUE, it calls save.image in the current working directory and copys the resulting .RData file to the container's working directory. The created file in the local working director will not be deleted.
Alternatively, you can pass a list of objects to be saved, which may also include arguments to be passed down to save, e.g. save_image = list("object1", "object2"). You can configure the name of the file the objects are saved to by adding a file name to the list of arguments, e.g. save_image = list("objectA", save_image_filename = "mydata.RData"), in which case the file path must be in UNIX notation. Note that you may not use save_image_filename for other objects in your session!
save will be called with envir.
The environment for save_image.
optionally specify environment variables to be included in the image. See documentation: 'https://docs.docker.com/engine/reference/builder/#env
(boolean) Whether to include soft dependencies when system dependencies are installed, default is no.
(boolean) Whether to use an online database to detect system dependencies or use local package information (slower!), default is no.
whether and how a workspace should be copied - values: "script", "script_dir" or a list of relative file paths to be copied, or NA ot disable copying of files
the working directory in the container, defaults to /payload/ and must end with /. Can be skipped with value NULL.
The CMD statement that should be executed by default when running a parameter. Use CMD_Rscript(path) in order to reference an R script to be executed on startup, CMD_Render(path) to render an R Markdown document, or Cmd(command) for any command. If character is provided it is passed wrapped in a Cmd(command).
the ENTRYPOINT statement for the Dockerfile
Whether to add the package containerit itself if loaded/attached to the session
Whether or not to print information during execution
Extract the required libraries based on library calls using the package automagic before running a script/document
[EXPERIMENTAL] Whether it shall be attempted to match versions of linked external libraries
Whether it shall be attempted to match versions of R packages
Do not add packages from CRAN that are already installed in the base image. This does not apply to non-CRAN dependencies, e.g. packages install from GitHub.
An object of class Dockerfile
Use the current sessionInfo) to create a Dockerfile.
Given an existing path to a directory, the method tries to automatically find the main R file within that directory.
Files are searched recursively. The following types are supported:
regular R script files, identified by file names ending in .R
weaved documents, identified by file names ending in R Markdown (.Rmd)
After identifying the main file, the process continues as described in the section file.
If both types are found, documents are given priority over scripts.
If multiple files are found, the first file as returned by dir will be used.
Given an executable R script or document, create a Dockerfile to execute this file.
This executes the whole file to obtain a complete sessionInfo object, see section "Based on sessionInfo", and copies required files and documents into the container.
# NOT RUN {
dockerfile <- dockerfile()
print(dockerfile)
# }
Run the code above in your browser using DataLab