Remove duplicate metagenomics withName blocks in modules.config (closes #1173)

[Golchic]

Closes #1173. Approach as agreed with @TCLamnidis in the issue.

Changes

An older set of metagenomics withName blocks lived in the AUTHENTICATION
section of conf/modules.config, using the outdated
metagenomics/profiling/… + metagenomics/postprocessing/… publishDir
layout. They duplicated the blocks in the // METAGENOMIC SCREENING
section, which use the canonical <section>/<tool>/{data,stats}/ layout.
Nextflow's last-definition-wins meant the screening-section blocks were
already the effective config, so the old ones were dead weight.

• Removed the 10 outdated duplicates: BBMAP_BBDUK, MALT_RUN,
  CAT_CAT_MALT, KRAKEN2_KRAKEN2, KRAKENUNIQ_PRELOADEDKRAKENUNIQ,
  METAPHLAN_METAPHLAN, MEGAN_RMA2INFO, AMPS, TAXPASTA_MERGE,
  TAXPASTA_STANDARDISE.
• Moved the directives that only existed on the old blocks onto the
  screening-section blocks, ordered per docs/development/code_conventions.md
  (tag → ext.args → ext.prefix → publishDir):
  • KRAKEN2_KRAKEN2: tag
  • KRAKENUNIQ_PRELOADEDKRAKENUNIQ: tag + ext.prefix
• Removed a byte-identical duplicate of
  .*MERGE_LIBRARIES:SAMTOOLS_MERGE_LIBRARIES in the LIBRARY MERGE section.

MALTEXTRACT is left untouched — it has no screening-section counterpart
(not a duplicate), so relocating/retargeting it is out of scope here. Happy
to do that as a follow-up if wanted.

Points for review

• The moved KRAKEN2_KRAKEN2 tag ({ "${meta.sample_id}|single_end_mode_${meta.single_end}" })
  does not include ${meta.reference}, which code_conventions.md asks for
  on reference-specific processes. I moved it verbatim per "retain and move"
  rather than redesign it (the single_end_mode grouping looks intentional
  for metagenomic screening). Happy to align it to the
  ${meta.reference}|${meta.sample_id}_* format if you'd prefer.
• No CHANGELOG.md entry — dev doesn't appear to track per-PR changelog
  entries during the DSL2 work (last touched 2023). Glad to add one.
• docs/output.md unchanged: effective publishDir paths are the same (the
  screening-section blocks that already won are untouched).
• Didn't run nf-core pipelines lint locally; verified structurally
  (no remaining duplicate withName selectors, braces balanced, directive
  order per code conventions).

PR checklist

• This comment contains a description of changes (with reason).
• If you've fixed a bug or added code that should be tested, add tests! — config-only cleanup, covered by existing pipeline tests
• Make sure your code lints (nf-core pipelines lint). — not run locally, see above
• Ensure the test suite passes — via CI
• CHANGELOG.md is updated. — see above

[ilight1542]

Looks totally fine to me, ran through some test datasets and the output directory structure is A-OK 👍

[ilight1542]

i double checked and multi-ref is not an issue here, since MALT is run in groups of single stranded/double stranded datasets.

totally fine to leave implementation as it is and delete duplicate :)

[ilight1542]

correct, remove old version of publish directories structure

[ilight1542]

I will also suggest the mentioned adjustment to get the maltextract postprocessing in line with current publish directory standards

[ilight1542]

Suggested change

	path: { "${params.outdir}/metagenomics/maltextract/stats/" },

[ilight1542]

Suggested change

	pattern: 'results/*',

[ilight1542]

minor changes for maltextract

[ilight1542]

Suggested change

[ilight1542]

this will save everything for maltextract (still within the results subdirectory, couldnt figure that out), in-line with current standards :)

[Golchic]

Thanks for testing this out and for the detailed suggestions! Fixed all three — publishDir now points to metagenomics/maltextract/stats/, pattern widened to results/*, and dropped the saveAs so everything under results/ actually gets published. Let me know if anything else needs tweaking 🙂