bakta/bakta: gzip output files

Description of feature

Bakta annotates genomes and writes a set of text-format output files (.embl, .faa, .ffn, .fna, .gbff, .gff3, .hypotheticals.tsv, .hypotheticals.faa, .tsv, .txt, .json), none of which are currently compressed by the bakta/bakta module. Several of these (.embl, .gbff, .fna, .ffn) can get large for bigger genomes/collections, and per the nf-core component review checklist, large outputs should be gzipped where the downstream tooling allows it.

We carry a local patch in nf-core/magmap that gzips all of Bakta's outputs (writing them into a ${prefix}/ subdirectory and running gzip ${prefix}/* at the end of the script), used across a run that maps against potentially large genome collections where the annotation output volume adds up. Flagging this as a candidate for the upstream module rather than keeping it local-only, in case it's useful more broadly. Happy to open a PR with the change if that's welcome — wanted to check first since it changes the output paths/extensions. I suppose this could be controlled by a value argument to the module.

Locate the bakta/bakta module and its nf-test coverage, then read how output files and paths are currently declared. Check the module conventions for optional parameters and compressed outputs before deciding how the requested gzip behavior should be exposed. Done means the intended Bakta outputs are compressed without breaking declared paths, and the module tests cover the resulting outputs.