Skip to contents

Environment variables

Document generated: 2024-09-14 13:58:30 UTC+0000

Source: vignettes/scdrake_envvars.Rmd
scdrake_envvars.Rmd

On scdrake load or attach (?), several environment variables are read and corresponding options() are set. Those are mainly variables controlling paths to config directories. See get_scdrake_options() for more details and list of used environment variables.

We can take the SCDRAKE_SINGLE_SAMPLE_CONFIG_DIR environment variable as an example:

  • If it is set, the scdrake_single_sample_config_dir option is set on scdrake load or attach.
  • It it is not set, a default value "config/single_sample" will be used for the option above. You can see that the default value is consistent with the file structure of a newly initialized scdrake project.
  • Value of the scdrake_single_sample_config_dir option is the default one for many scdrake functions, e.g. update_single_sample_configs(dir = getOption("scdrake_pipeline_config_dir"), ...) or run_single_sample(dir = getOption("scdrake_pipeline_config_dir"), ...).

Environment variables are mainly used in _drake_single_sample.R and _drake_integration.R scripts. Those are scripts sourced by drake in a fresh new R session via callr when you call e.g. run_single_sample_r(). For this reason, environment variables are the only way how to pass dynamic parameters into such scripts. In case of scdrake, the most important are paths to directories with config files, which carry analysis/pipeline-specific parameters.