added new module: hackgap/count

[LeonHornich]

This PR adds a module supporting Hackgap, a k-mer counter supporting gapped k-mers.

Note: the tool outputs 2 types of files:

• .hash: hash table resulting from Bucketed Cuckoo. Since this is a probabilistic method, the files content (and with this the nf-test snapshots) are not reproducible between runs. That's why we only check for the files existence during testing.
• .info: metadata of hashing process.

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!
• If you've added a new tool - have you followed the module conventions in the contribution docs
• If necessary, include test data in your PR.
• Remove all TODO statements.
• Broadcast software version numbers to topic: versions - See version_topics
• Follow the naming conventions.
• Follow the parameters requirements.
• Follow the input/output options guidelines.
• Add a resource label
• 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:
  • For modules:
    • nf-core modules test <MODULE> --profile docker
    • nf-core modules test <MODULE> --profile singularity
    • nf-core modules test <MODULE> --profile conda

[piplus2]

better passing task.cpus only as this may become hidden from the user

[SPPearce]

What is the issue with this?

[piplus2]

this hides the cpu settings from the user. I don't get why one should set task.cpus and then it gets into 1/4th. Passing the right number makes it transparent.

[piplus2]

I get that they must sum to task.cpus, but 1/4 seems arbitrary. Any reason why that fraction?

[LeonHornich]

No particular reason. I used a similar logic on the last module i added and it wasn't commented on. Therefore I assumed it was good to go. But please if you have a suggestion for a better cpu thread handling let me know, I"ll gladly adapt it :)

[piplus2]

https://gitlab.com/rahmannlab/hackgap/-/blob/master/hackgap/hackgap/hackgap_count.py

def check_threads(args):
    adjust = (args.subtables is None) or (args.threads_split is None) or (args.threads_read is None)
    # 1. Define number of subtables
    cpus = os.cpu_count()
    subtables = args.subtables
    if subtables is None:
        subtables = max(min(cpus // 2 - 1, cpus - 3, 19), 1)
        if (subtables % 2) == 0:
            subtables += 1
    if subtables < 1:
        debugprint0(f"- Error: At least one subtable is required, but {subtables=}")
        sys.exit(1)
    if subtables % 2 == 0:
        debugprint0(f"- Error: Number of subtables must be odd, but {subtables=}")
        sys.exit(1)
    # 2. Define threads for reading files
    threads_read = args.threads_read
    if threads_read is None:
        threads_read = int(ceil(subtables / 10))  # who knows?
    # 3. Define threads for
    threads_split = args.threads_split
    if threads_split is None:
        threads_split = 2 * threads_read
    if adjust and (subtables + threads_read + threads_split >= cpus):
        threads_read = 1
        threads_split = 2
    # 4. Return results
    assert threads_read >= 1
    assert threads_split >= 1
    return (subtables, threads_read, threads_split)

from the source code it looks like that threads_split is suggested to be twice threads_read. I'd suggest to follow the same logic. Also please be sure that the sum subtables + threads_read + threads_split is never >= task.cpus, or it drops to 1 and 2 as from the source. Thanks!

[piplus2]

tiny stuff

[piplus2]

LGTM once threads are fixed