MultiQC fails when input datasets that have the same dataset name/label

[jennaj]

Problem: When two or more inputs have the same dataset name/label, the tool fails when creating symlinks for the job.

This is a common use case: Same inputs, different parameters = same output names but different content == want to compare results and cannot

More details plus workaround and examples at both ORG and EU servers here: galaxyproject/usegalaxy-playbook#315

[bernt-matthias]

I imagine that quite a few tools (probably all with an input with multiple="true" ) have this problem. But fortunately the use case might be not that common for most of them.

I guess using the index from the loop

tools-iuc/tools/multiqc/multiqc.xml

Line 72 in 6b9df5f

#for $i, $repeat in enumerate( $results )

Is the easiest fix. Question is if this needs to be removed from the output as well.

Maybe also this works..
https://multiqc.info/docs/#sample-names-prefixed-with-directories

But, maybe both changes do not make sense, since users will have no way to differentiate the samples in the output. In this case a check for equal sample names and a reasonable error message might be better.

[jennaj]

But, maybe both changes do not make sense, since users will have no way to differentiate the samples in the output. In this case a check for equal sample names and a reasonable error message might be better.

Agree -- better error message that a biologist can interpret that points to more help directly on the tool form itself (rename inputs, plus how/why).

Would this be a papercut?

[bernt-matthias]

Would this be a papercut?

Maybe .. alternatively I could try to get this in #3305 ..

in case you know MultiQC well I would appreciate a review.

[jennaj]

Sure, that would simplify the MTS update/deployment to bundle it all together. Plus, just one version bump for all changes.

Looks like Maria is going to review. I can review too if this is added in as well :)

Thanks!