Hi all,
this is slightly connected to this post, but a more general problem.
I’m running a workflow, where I don’t know the explicit output name when calling renku run
using renku-python version 1.0.1. Usually renku does automatically detect all new files and I don’t need to worry about explicitly specifying the output. In this case I parse the path to the output file as command line option. Renku seems to detect the output path as output instead of the actual output file. This gives rise to problems when executing the workflow with a different set of parameter etc. and will also not work if files exist in the output directory.
In this project I have a test case (workflow param_5
):
The test script (src/parameter_outpath2.sh) has two input arguments -o
, which is the path to the output file and -p
with some parameter to write to to the output file’s content.
When I run it using:
renku run --name param_5 -- bash src/parameter_outpath2.sh -o "data/test2" -p "test outpath"
a workflow is generated, but I get this note:
and renku workflow show
looks like:
with the output directory defined as output instead of the output file itself.
Interestingly the same happens when running renku run
with --no-output-detection
and explicitly specifying the output file:
Is there a way to circumvent this problem and call renku run
while parsing the output path to the script to run, e.g., by being recognized as parameter instead of output?
Best,
Almut