sal2bash: Translate SYSargsList back to a bash workflow

In tgirke/systemPipeR: systemPipeR: Workflow Environment for Data Analysis and Report Generation

Translate SYSargsList back to a bash workflow

Description

This function takes a SYSargsList object and translate it to an executable bash script, so one can run the workflow without loading SPR or using an R console.

Usage

sal2bash(sal, out_dir = ".", bash_path = "/bin/bash", stop_on_error = TRUE)

Arguments

sal	SYSargsList object.
out_dir	string, a relative or absolute path to a directory. If the directory does not exist, this function will first try to create it.
bash_path	string, the path to the bash executable program
stop_on_error	bool, should the bash script stop if any error happens in running

Details

## out files

1. The main executable bash file will be created to the root of 'out_dir'

2. All R steps will be stored as R scripts and along with other supporting files inside a folder called 'spr_bash' under 'out_dir'

3. Not all R steps will have an individual file. This function will "collapse" adjacent R steps into one file as much as possible. Namely, if there is no sysArgs steps in between, R steps will be merged into one file, otherwise they will be in different files.

## R steps

Similarly as running the workflow in R console, all R steps will share the same environment variables and loaded packages. This is done by loading and saving the R environment into a file 'spr_wf.RData' before and after the R script execution. Therefore, it will be good to keep all R steps bundle together as much as possible to avoid the package and environment loading/saving overhead time.

Initially, this environment only contains the SYSargsList object that was used to create the bash script. Note: the SYSargsList object name will be the same as what you pass to 'sal2bash'. If you have 'sal2bash(my_sal)', then in the 'spr_wf.RData' there will be an object called 'my_sal' saved there. It is important to keep using the same name for the SYSargsList object managing the workflow and in workflow running. For example, if you have an R step that requires to query a column from the outfiles of the SAL, 'getColumn(sal, "FileName1")', but you pass 'my_sal' to 'sal2bash(my_sal)', this will cause this R step cannot find the SAL object when run the workflow from bash.

## Execution

This way of execution is not able to handle complex dependency graphs. The original step dependencies from SAL object will be ignored, so all steps will be executed in a linear manner. It is recommended to adjust the workflow order before using this function.

Value

no return

Author(s)

Le Zhang and Daniela Cassol

Examples

file_path <- system.file("extdata/spr_simple_wf.Rmd", package="systemPipeR")
sal <- SPRproject(overwrite = TRUE)
sal <- importWF(sal, file_path)
sal2bash(sal)

tgirke/systemPipeR documentation built on Sept. 24, 2024, 9:48 a.m.