DSL2: Duplicate module configurations in modules.conf

[TCLamnidis]

Some modules have duplicate withName configurations in modules.conf. This is likely from a merge conflict at some point.

The duplicates should be removed, without altering final configuration of the different modules, where possible.

[Golchic]

I had a look at conf/modules.config on dev (2e46e89d). There are 11 duplicated withName: selectors, in two distinct categories.

1. Exact duplicate — safe to remove

Selector	Lines
'.*MERGE_LIBRARIES:SAMTOOLS_MERGE_LIBRARIES'	1442 & 1487

Both blocks are byte-for-byte identical (tag, ext.prefix, publishDir = [ enabled: false ]). One can be deleted with no behaviour change.

2. Metagenomics selectors — two different configs, needs a decision

Each appears once in the earlier scattered block (~L940–1075) and once in the later // METAGENOMIC SCREENING section (~L1287–1440):

Selector	Earlier	Later (METAGENOMIC SCREENING)
BBMAP_BBDUK	942	1287
MALT_RUN	954	1307
CAT_CAT_MALT	973	1333
KRAKEN2_KRAKEN2	982	1350
KRAKENUNIQ_PRELOADEDKRAKENUNIQ	995	1369
METAPHLAN_METAPHLAN	1005	1384
MEGAN_RMA2INFO	1037	1395
AMPS	1048	1408
TAXPASTA_MERGE	1057	1417
TAXPASTA_STANDARDISE	1066	1428

These are not identical — they encode two different output-directory layouts:

• Earlier: single flat publishDir, grouped under metagenomics/profiling/<tool>/ and metagenomics/postprocessing/<tool>/
• Later: split publishDir into metagenomics/<tool>/data/ + metagenomics/<tool>/stats/

A few of the earlier blocks also carry directives the later ones drop:

• KRAKEN2_KRAKEN2 (earlier) has tag = { "${meta.sample_id}|single_end_mode_${meta.single_end}" }
• KRAKENUNIQ_PRELOADEDKRAKENUNIQ (earlier) has tag and ext.prefix = { "${meta.single_end}" }

Since the later definition wins, the METAGENOMIC SCREENING section is what currently takes effect, but the duplication makes the file misleading and the dropped tag/ext.prefix on kraken2/krakenuniq may be unintended.

Question for maintainers: which layout is canonical — the flat profiling/ + postprocessing/ grouping, or the split <tool>/data/ + <tool>/stats/? Once that's decided I'm happy to open a PR removing the other set (and carrying over the tag/ext.prefix directives where they should be kept).

[TCLamnidis]

Hi @Golchic!

The canonical publishDir is to place the results within a <section>/<tool>/{data,stats}/ directories. Indeed the older blocks are outdated.

As for the two KRAKEN configuration blocks, the missing directives should be retained and moved to the later blocks.

Please note that the order of directives in the finalised blocks should follow the conventions outlined in https://github.com/nf-core/eager/blob/dev/docs/development/code_conventions.md

Thanks for taking a look! A PR would be greatly appreciated ^_^