Hi all :slightly_smiling_face: Does anybody know ...
# modules
l
Hi all 🙂 Does anybody know how to handle very big containers for nf-core modules? I am working on the learnMSA module. The current container is not the complete one and will fail for the recommended command line (--use_language_model flag). I am now planning to use their official container - which is 13 GB and the github tests run out of space. Anybody has some reccomandation or had similar issues? Thanks!
shakefist 1
s
You can only use a stub test potentially
this 1
but without even using the container it becomes pretty pointless
(you could override the container with a config)
j
Would it be an option to increase the memory of self-hosted-runners and use those? I'm not sure if this is too much work or it's worth it for one module
@mashehu since you know about the self-hosted runners, what do you think?
l
But won't the stub run also use the container?
s
Yeah, normally the stub run will use the container; you could also overwrite that with a config inside the test. But that doesn't really then actually test anything at all
l
Ah I see!
Thanks! I will wait to see if the self-hosted runners are an option and i'll otherwise proceed with this suggestion 🙂
@Simon Pearce do u know how to overwrite the container with the config? I am trying with this but it does not overwrite the main container
Copy code
process {
    container = "nf-core/ubuntu:22.04"
}
It will still try to fetch the one specified in the main.nf
This is my branch so you can see all the code I am using if you want https://github.com/luisas/modules/tree/update_learnmsa/modules/nf-core/learnmsa/align
s
you might need to put it inside a
withName:
😍 1
l
worked - thanks a ton!
Here the PR if any of you has some time 🙂 https://github.com/nf-core/modules/pull/7413/files Thanks!