<sequenceReadsPaired> input generates a separate invocation for each pair

[innovate-invent]

Describe the bug

input generates a separate invocation for each pair
IRIDA always creates a list:pair with a single pair in the list

Steps to reproduce the problem

What were you doing when you encountered the problem?

1. Add two samples to cart
2. Run RefSeqMasher Pipeline on both samples
3. View analysis list

Expected behaviour

What did you expect to happen?
A single analysis should be created, passing a list:pair to the workflow.
The list:pair should have two elements, a pair for each sample.

Additional context

Additional information to help solve the problem. Test files, screenshots, IRIDA version information can go here.

[apetkau]

Thank you for reporting @innovate-invent, I appreciate it.

I'm not quite sure what you mean for this bug though, since this sounds like it's working as expected. The RefSeqMasher pipeline should submit a separate invocation for each pair. This can be controlled with the <requiresSingleSample> element in a workflow XML file - https://phac-nml.github.io/irida-documentation/developer/tools/pipelines/workflow-description/#requiressinglesample

For example:

<!-- Used when you wish to invoke a single Galaxy workflow instance for all samples -->
<requiresSingleSample>false</requiresSingleSample>

<!-- Used when you wish to invoke a separate Galaxy workflow instance for every sample-->
<requiresSingleSample>true</requiresSingleSample>

[innovate-invent]

The RefSeqMasher workflow expects a list:pair implying that all samples should be handed to a single invocation.

irida/src/main/resources/ca/corefacility/bioinformatics/irida/model/workflow/analysis/type/workflows/RefSeqMasherOnPairedReads/0.1/irida_workflow_structure.ga

Line 27 in 181b427

"tool_state": "{\"collection_type\": \"list:paired\", \"name\": \"sequence_reads_paired\"}",

The workflow should expect a pair if the intent is a single invocation per sample. <requiresSingleSample>true</requiresSingleSample> should cause a pair to be created instead of list:pair.

irida/src/main/resources/ca/corefacility/bioinformatics/irida/model/workflow/analysis/type/workflows/RefSeqMasherOnPairedReads/0.1/irida_workflow.xml

Lines 8 to 9 in 181b427

	<sequenceReadsPaired>sequence_reads_paired</sequenceReadsPaired>
	<requiresSingleSample>true</requiresSingleSample>

[apetkau]

Ah, okay. I understand. Yes, I can see how that can be confusing since list:pair implies an arbitrarily sized list.

I will relabel this as a request though since this was how it was designed from the beginning. The idea being that you use a list of paired-end files with a single element for pipelines that create a single invocation per sample.