Yes, you can <@UPC8CHDKQ>, if I get one on: <githu...
# release-review-trading
d
Yes, you can @Rike Hanssen, if I get one on: github.com/nf-core/magmap/pull/228. Two template updates, three new features and some bug fixes. Only 169 files!
🤝 1
Some non-GPU tests are failing.
Looks a bit random, different tests for different container types, so maybe just temporary.
r
will check. I am about half way through yours, but have meetings now. I got a few notes, I think my biggest one is around the empty processes for the version collection. Why is this needed?
d
You mean
PROKKA_VERSION
and
BAKTA_VERSION
? They're there because we use
storeDir
to save files to be able to use directories as caches, so genomes doesn't have to go through annotation multiple times. When a user runs the pipeline with a new project they just point to the same directory and as many genomes might already be annotated they will just be skipped. This is not supported by versions neither in the old file format nor non-file topic channel. The problem is that we have to remember to update the container in these modules whenever the main tool container is updated. Now, this is recorded in
AGENTS.md
to help us with that.
r
how can you ensure that the genomes in that directory were downlaoded with the container version you are reporting?
d
I can't. That's a glitch, but well worth it. In one machine I have 80,000 prokaryotic genomes annotated now.
This is I believe not fully documented, so worth an issue. I'll create one right away.
This is documented now. I pushed an update directly to upstream/dev so it's part of what you see. Unfortunately, all tests were triggered.