Arguments with spaces cause errors

[arisp99]

Bug Description

When arguments contain spaces, argparse is unable to parse the spaces and causes our code to crash. I encountered this error when attempting to run through the analysis test run. When running through the first section of the tutorial, the user is asked to specify the probe sets used before running the wrangler app. In the example, we write

probe_sets_used="DR1,VAR4"

However, if the user adds a space in between these two probe sets:

probe_sets_used="DR1, VAR4"

and tries to run the wrangler app using the following:

singularity run --app wrangler \
    -B $project_resources:/opt/project_resources \
    -B $fastq_dir:/opt/data \
    -B $wrangler_dir:/opt/analysis \
    $miptools -p $probe_sets_used -s $sample_sets_used -e $experiment_id \
    -l  $sample_list -c $cpu_number -m $min_capture_length

the app will fail with the following error: generate_wrangler_scripts.py: error: unrecognized arguments: VAR4.

Our code does aim to deal with these cases by cleaning up spaces in the following lines:

MIPTools/src/generate_wrangler_scripts.py

Lines 79 to 96 in 09f52ef

	# get sample sets and probe sets to be processed.
	# sample sets should be provided as a comma separated text
	# sometimes semicolon is used and spaces may be included by mistake
	# we'll check for these potential mistakes and create a list of sample sets
	# to process.
	sam_set = args["sample_sets"]
	sam_set = sam_set.split(",")
	sample_sets = []
	for s in sam_set:
	sample_sets.extend(s.strip().split(";"))
	sample_sets = set([s.strip() for s in sample_sets])
	# probe sets are provided and will be processed similarly to sample sets
	pr_set = args["probe_sets"]
	pr_set = pr_set.split(",")
	probe_sets = []
	for p in pr_set:
	probe_sets.extend(p.strip().split(";"))
	probe_sets = set([p.strip() for p in probe_sets])

However, the code fails before we are able to reach this stage. It seems that the error is triggered by argparse here:

MIPTools/src/generate_wrangler_scripts.py

Line 68 in 09f52ef

args = vars(parser.parse_args())

Suggested Implementation

Seeing as we are unable to deal with spaces within the python file, I suggest we deal with any spaces and misformatted arguments within the wrangler app before feeding in our arguments to the python script. We can do simple string manipulation in bash as outlined here. For instance, to remove whitespace:

probe_sets_used="DR1, VAR4"
echo bash ${probe_sets_used// /}
#> DR1,VAR4

[aydemiro]

yes we can change the app code to:

    # remove accidental spaces
    probe_sets=${probe_sets// /}
    sample_sets=${sample_sets// /}

    # Create wrangler bash scripts using python
    python /opt/src/generate_wrangler_scripts.py \
    -c ${cpu_count} -e ${experiment_id} ${keep_files} \
    -l /opt/analysis/${sample_list} -m ${min_capture_length} \
    -n ${server_number} -p ${probe_sets} -s ${sample_sets} \
    -w ${cluster_script} -x ${stitch_options}

Edit: probe_sets and sample_sets parameters are used in the app code instead of probe_sets_used.