Adding module sourmash/gather

[vmikk]

PR checklist

Closes https://github.com/nf-core/modules/issues/2214

• [x] This comment contains a description of changes (with reason).
• [x] If you've fixed a bug or added code that should be tested, add tests!
• [x] If you've added a new tool - have you followed the module conventions in the contribution docs
• [x] If necessary, include test data in your PR.
• [x] Remove all TODO statements.
• [x] Emit the versions.yml file.
• [x] Follow the naming conventions.
• [x] Follow the parameters requirements.
• [ ] Follow the input/output options guidelines.
• [x] Add a resource label
• [x] Use BioConda and BioContainers if possible to fulfil software requirements.
• Ensure that the test works with either Docker / Singularity. Conda CI tests can be quite flaky:
  • [ ] PROFILE=docker pytest --tag sourmash --symlink --keep-workflow-wd --git-aware
  • [ ] PROFILE=singularity pytest --tag sourmash --symlink --keep-workflow-wd --git-aware
  • [ ] PROFILE=conda pytest --tag sourmash --symlink --keep-workflow-wd --git-aware

There is an error in tests, associated with different module - nf-core/fastq_align_dna Considering output guidelines - output files (signatures and csv are not compressed).

[jfy133]

@nf-core-bot fix linting

[Midnighter]

Added some comments! I also have a general question -- since this module is reliant on a database, what is the recommended way for a workflow to supply that database and to make sure that the k-mer sizes match between the database and the signature?

I think the most common way would be simply as a path input to the process and it would be up to the pipeline developer/user to ensure that there parameters like k-mer size are consistent.

[vmikk]

Dear Taylor! Thank you so much for the valuable comments and corrections! I will modify my PR accordingly soon (probably after the Nextflow Summit).

Considering the database parameter, I think it should work even with multiple databases if we'll use a channel with multiple values (e.g., with collect operator) - I will check it.

Also, I've almost drafted sourmash/taxonomy module. If you would agree, I will show it to you before the submission.

[taylorreiter]

Thanks @vmikk! Yes I would love to review the sourmash taxonomy module as well!

[jfy133]

I think, the input parameters style relates to the discussion on slack, so having an extensive configuration via ext.args might be preferred to the multiple inputs. Otherwise the module looks good to me!

I think these are generally OK though, I think the discussion slack was related to non-file related mandatory options?

[Midnighter]

I must admit, I'm confused which style is preferred. I think, all of the following inputs could be dropped in favor of using task.ext.args:

    val save_unassigned
    val save_matches_sig
    val save_prefetch
    val save_prefetch_csv

[jfy133]

I must admit, I'm confused which style is preferred. I think, all of the following inputs could be dropped in favor of using task.ext.args:

    val save_unassigned
    val save_matches_sig
    val save_prefetch
    val save_prefetch_csv

Yeah I guess the discussion is still on going.

The reason why I don't mind these as channels is because they aren't just simply Boolean flags, but you have to supply the time file name with them.

Did consistency with the reason why we have the meta.id as a default information to construct the output file name, I personally like it to be explicit, as then a pipeline dev knows clearly what types of output files a module can generate - and how they are generated (like in fastp).

[jfy133]

I would argue we can merge this in, unless you strongly disagree about the val channels @Midnighter ?

[Midnighter]

Already approved, can definitely be merged 😃