ivoryc / myscript_module Goto Github PK

do a demo using the wait function, to show that the program will wait for the script to finish.

IF it doesn't, then remove the line to add the _Success file, and make that part of the user script.

add a line to the script that does
#start of script
touch script.sh_Started
...
#end of script
touch script.sh_Success

This may just be a requirement for user scripts.
We could probably add a bit of code in checkDependencies() to check that a line like this is present, and if it is not suggest a line to the user.
Depending on the launch command,
Please modify script [script.sh] to include the following line at the end of the script, or the end of the main method:
if bash: touch script.sh_Success
if R: system("touch script.sh_Success")
if python:

The config option:
myScript.launcher=

is intended for things like "python", or "Rscript". Default is none (ie, bash)
It up to the user to make sure that the launch command is available in the running environment (cluster module, or local install, or docker image, etc).

Verify that this is working for "python" and "Rscript" and "Rscript -knit" (or whatever the arg is for knitting a markdown from the command line). Make corresponding demos, similar to the "someUserScript.sh" example.

With config:
myScript.resources=path/to/myFavLib.sh, path/leading/to/image.png, otherFile.weird, refTable.txt
or
myScript.resources=path/to/dir

Each resource will be copied to top level of module resources folder.
If a directory is supplied, the contents of that directory will be pulled in.

They can be called from the imported script using "../resources/"

In an upcoming change to the docker utility, modules will have to implement a method to return the docker image they need.
getDockerImage()

This probably create a general config property (docker.dockerImage=) that can be supplied in a module-specific form (myScript.dockerImage=), so the user will be able to supply the appropriate docker image for their script to run in.
Once that change is made, this module will need to implement the method.

99.9% of the time, the scripts our users supply will be one of bash, python, R, or java.
We can determine which of those it is based on the launch command or the file extension of the script. IF there is no docker image given in the config file, the method should choose one of the biolockj docker images (blj_basic_python, blj_basic_bash etc).

If it is run in docker mode without a configured docker image, so it has to use a default, it should give a info/warning message saying which image is used and why. We might also say that if this is the desired container to please put that in the config file.

running
file.setExecutable(true, false)
was not enough to run the script.

setting the file permissions with chmod a+x before the program started
was enough for the script to run.

If users make the file executable, will that always work?
What about the resources? Will they preserve their original permissions?