If I have a module that reads fna.gz file, I'm assuming I put FASTA (http://edamontology.org/format_1929 # FASTA) as the ontology in the meta.yml file. Is that correct? Or should I put the EDAM link for GZ?

Does the main process read the compressed file or you decompress it inside the module?

The main.nf file reads a compressed file, which is decompressed inside a python script. I can separate the decompression from the python script, but I'm not sure about the advantages and disadvantages of doing so or leave it inside the python script. The template seems to say - use the compressed version whenever possible. Is that not right?

Since there's already the GUNZIP module for that, I'd suggest to only include the actual process as it makes the code more reusable.

Okay. I'll put that on the TODO list

Of course, unless there's a very strong constraint that makes it unavoidable or unpractical

Don't think so? Maybe the FASTA files could be big if they're not compressed

You may have a look at modules for indexing or mapping as they usually work with large fasta files EDIT: I refer to things like STAR

Thank you. But that seems to imply that I should use the compressed version, and not the GUNZIP module. Would you agree, or am I missing something?

That's correct. If your tool either works with gz or you can pipe then it's ok to pass the gz as input. I'm not sure about splitFasta, it looks like it works with uncompressed data.