The following JIP script demonstrates what you can do with dynamic options. It grew out of a practical problem where we had to run the Flux Capacitor on both the output of the TopHat and the GEMTools pipeline.
The issue was that GEMTools creates a .bam file with the name that should be used for the capacitor output, but TopHat writes its output into a .bam file in a sub-folder, and the folders name is the one that we want to use for the result file.
The solution to the problem was the JIP script below, where we set the output option dynamically in the validation phase:
#!/usr/bin/env jip
# Run the flux. NOTE this outputs stuff relative to CWD
#
# usage:
# flux.jip --input <bam> --annotation <annotation>
#%begin validate
if basename(input.get()) == 'accepted_hits.bam':
name('flux-FL-${input|parent|name}')
add_output('output', r('${input|parent|name}.gtf'))
else:
name('flux-FL-${input|name|ext}')
add_output('output', r('${input|name|ext}.gtf'))
#%end
flux-capacitor --threads $JIP_THREADS -o ${output} -a ${annotation} -i ${input}
In the validate block, we check the inputs base name. If its equal to accepted_hits.bam, the output of TopHat, we add an output option and render (note the r function call to render a template) its content to the name of the parent folder. In all other cases, we render the value of the output option to be the input files’ base name minus its extension.
In the execution part of the script, we can now access a valid and properly set output option.
Because we had a handful of datasets that needed to be processed, we could also leverage the multiplexing capabilities of all pipelines. This command submitted all the TopHat based runs:
$~tophat> flux.jip --input `find . -name "accepted_hits.bam"` -- submit
The gemtools results were all located in a single folder, so we could start all of them with:
$~gemtools> flux.jip --input *.bam -- submit