Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/mdneuzerling/usedocker
This is an abandoned package that I intended to use for simplifying the creation of Dockerfiles for R projects. Please don't use this package. Use dockyard instead.
https://github.com/mdneuzerling/usedocker
Last synced: 9 days ago
JSON representation
This is an abandoned package that I intended to use for simplifying the creation of Dockerfiles for R projects. Please don't use this package. Use dockyard instead.
- Host: GitHub
- URL: https://github.com/mdneuzerling/usedocker
- Owner: mdneuzerling
- License: other
- Created: 2020-10-13T21:02:47.000Z (about 4 years ago)
- Default Branch: main
- Last Pushed: 2020-10-13T21:08:06.000Z (about 4 years ago)
- Last Synced: 2024-08-13T07:13:38.930Z (4 months ago)
- Language: R
- Homepage:
- Size: 12.7 KB
- Stars: 3
- Watchers: 3
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
- jimsghstars - mdneuzerling/usedocker - This is an abandoned package that I intended to use for simplifying the creation of Dockerfiles for R projects. Please don't use this package. Use dockyard instead. (R)
README
# usedocker
This is an **abandoned** package that I intended to use for simplifying the creation of Dockerfiles for R projects. Please don't use this package. Use [dockyard](https://github.com/thebioengineer/dockyard) instead.
While I don't think this package should ever be installed, it may be have some useful code that can be harvested.
## Design philosophy
My intention was to create a package to lower the barrier of entry to using Docker with R. Naturally, this would have leaned heavily on [the Rocker project](https://www.rocker-project.org/).
I had the following ideas in mind:
* Chain functions together with `%>%`, which would map to Docker commands separated by new lines
* Associate each function with a human-readable _description_. Each command would have default descriptions, although the user can provide their own.
* Take advantage of R's metaprogramming features to make it easier for user's to list arguments, eg. `install_r_packages(plumber)` rather than `install_r_packages("plumber")`
* Keep track of the "status" of the dockerfile using S3 and attributes, to warn users about potential mistakes such as installing an R package before installing R
## Example
``` r
library(usedocker)
library(magrittr)
example <- from_rocker(version = "3.6.2") %>%
install_system_dependencies(
make,
libsodium-dev,
libicu-dev,
libcurl4-openssl-dev,
libssl-dev
) %>%
install_r_packages(plumber, promises, future) %>%
docker_env(use_ssl = FALSE) %>%
docker_run(
"groupadd -r plumber && useradd --no-log-init -r -g plumber plumber",
description = "Creating non-root plumber user"
)
```
The Dockerfile prints nicely:
``` r
example
#> Start with rocker/r-ver:3.6.2
#> Install system packages: make, libsodium-dev, libicu-dev, libcurl4-openssl-dev, libssl-dev
#> Install R packages: plumber, promises, future
#> Declare environment variable: use_ssl = FALSE
#> Creating non-root plumber user
```
The raw commands can also be printed:
``` r
print_commands(example)
#> FROM rocker/r-ver:3.6.2
#> RUN apt-get update && apt-get install -y --no-install-recommends make libsodium-dev libicu-dev libcurl4-openssl-dev libssl-dev
#> RUN Rscript -e "install.packages('plumber','promises','future')"
#> ENV use_ssl = 'FALSE'
#> RUN groupadd -r plumber && useradd --no-log-init -r -g plumber plumber
```