Daniel Nüst

The best option is `sysreqs`. If that does not work I see this option: Ask users to create a PR based on a template for bespoke installation functions ``` containerit::bespoke_instructions

Show how you can use `containerit` and `stevedore` and _build arguments_ to put useful information into labels, e.g. build timestampe, build user, current git commit. Also show how to access...

See https://pak.r-lib.org/ _What are advantages for reproducibility?_

See https://www.jumpingrivers.com/blog/what-r-version-do-you-really-need-for-a-package/

Add new demo/vignette for passing parameters to the script

1

Create a demo script that accepts environment variables as parameters and show how these can be changed when running the script within the container.

Add a metadata extraction script to higher level containers (session extraction)

2

When packaging research into higher level containers, e.g. [ERC](http://dx.doi.org/10.1045/january2017-nuest), we most probably need some meta information. While the use can be asked for this, see #13, it would be better...

Add all dependencies to Dockerfile, even transitive ones

1

Potentially useful functions: - Using `pkggraph` we could explicitly list all R package dependencies of the loaded packages in the Dockerfile. See https://cran.r-project.org/web/packages/pkggraph/vignettes/vignette_pkggraph.html - `tools` package could also already do...

sysreqs has it: https://sysreqs.r-hub.io/pkg/runjags maybe just neet to test properly..

The package seems to cause some problems during precheck. https://cran.r-project.org/package=spam

https://pachyderm.readthedocs.io/en/latest/getting_started/beginner_tutorial.html https://github.com/pachyderm/pachyderm/tree/master/doc/examples/ml/iris