how to pass two parameters into bash script

[avilella]

Hi,

I have a plugin with two text parameters, and I want to run them in a bash script like so:

mybash.sh

echo $1 and $2

So far it works with $1, but $2 is ignored. Any ideas? Thanks.

[veidenberg]

Hi @avilella, sorry for the delay. Can you share the plugin JSON?

[avilella]

Hi,

I thought of two ways of doing this, explained below. Let me know what's the best solution:

One way in which I run many of my pipelines is to pass the name of the output file in the STDOUT of each step, and in bash, use xargs -n1 -I % to pass that file as the input for the next step in a UNIX pipe. This is different from, say, a bwa mem x y z | samtools sort a b c | samtools ... in that I am passing the file as input/output, and I was wondering if there could be an example of this file-based pipeline json configuration for pline, rather than the example in pipelines/read_mapping example json file, which is like the bwa-mem/samtools type. Would that be possible?

Another way in which I am running things sometimes is where I have two bash parameters, e.g. $1 and $2, interspersed in a UNIX pipe in the pipeline.sh file, e.g.:

#pipeline.sh

python3 ~/tool1/step1.py --input $1 --defaults xyz | xargs -n1 -I % python3 ~/tool2/step2.py --input % --parameter $2 | xargs -n1 -I % python3 ~/tool3/step3.py --input % --defaults abc

In the example above, $1 is a file parameter, and $2 is a string parameter, so my pline .json only needs to care about those 2 parameters, but because they are interspersed in different steps chained together by the UNIX pipe in pipeline.sh, $2 is not seen by pline . Maybe there is a way of doing this without having to transform the entire pipeline.sh into a multi-step pline pipeline.
E.g. if I have a json like below, I would like to somehow map out the first parameter to $1 and the second parameter to $2:

{
  program: "bash /home/user/pipeline1/pipeline.sh",
  name: "foo",
  desc: "Needs two parameters, one file one string",
  outFiles: "outfilename",
  jobName: "foo results",
  submit: ["Run tests"],
  prefix: "-",
  category: "XYZ",
  options: [
        {group: "Options", desc: "Hover option labels for description", options: [
                {file: "", desc: "this is the file parameter"},
                {text: "", desc: "this is the text parameter"},
        ]}
  ]
}

Thanks in advance.

[avilella]

I just found out that the 'Samtools view' step in the demo site shows a few maybe relevant options for ' * pipe' and others on the input side:

Is the json somewhere I could have a look?

[veidenberg]

The JSON files for tools and the pipeline used in the demo site are available in Pline homepage -> downloads.
The premise of Pline is that each command-line tool is described in a JSON and a pipeline JSON describes the order (and default inputs) for each tool. These tools will then be launched one by one, writing input/output files for each step to its subfolder.
Selecting the "pipe" option will change the default tools launch behavior in that the STDOUT of the preceding step will be piped to the next tool, so the intermediate output is not written to disk (only the input files of the preceding step and the output file of the following step are stored).
With that option (for file inputs, and you can set 'pipe' as default value in JSON), you can pipe multiple tools together while still being able to set the parameter values for each tool/step separately.

[avilella]

I think my equivalent that I do with xargs in the command-line is using the 'standard output' type of connection between one step and the next:

[veidenberg]

The standard ouput and pipe options differ in that the former writes stdout to file and then passes it as filename argument to the next step, whereas the latter does direct piping (merging two tools into one step without intermediate file dumps).