Dsl2 first draft#162
Conversation
48cdf4c to
b8db1ee
Compare
|
Thanks a lot for reviewing @d4straub ! The optional output seems to work in both cases so far, at least with the current Nextflow versions. |
KevinMenden
left a comment
There was a problem hiding this comment.
Looks good to me!
I've been mainly just looking at the modules and DSL2 stuff. Most of the modules are local so you can basically write them however you want. Probably a good idea though to make them similar to nf-core modules, so it's easier to put them on nf-core/modules later.
Mostly I found that a process label is missing for I think basically all local modules, and sometimes emit statements are missing.
|
Hi @KevinMenden , I added the missing emit statements (apart from one exception where it doesn't make sense, see above). Regarding the labels, we didn't use them for mag because it doesn't make sense to squeeze all those processes with very different requirements in these few categories. Thus I would also prefer adding the labels to the modules only when porting them to nf-core later. |
9e90a82 to
6822d86
Compare
|
|
I additionally changed some defaults in |
|
Okay, but the modules are always the same processes, so requirements should be similar right? |
eb51f45 to
303a3c2
Compare
|
Thanks a lot for reviewing @KevinMenden ! Regarding labels: it is just that for mag there are for various processes error strategies applied for specific |
d4straub
left a comment
There was a problem hiding this comment.
Could't find any issues any more, looks good!
|
Parallelised GitHub CI test to address #170. |
OK, here is a first running DSL2 version:
nf-core/test-datasetsin new format (Add mag test data samplesheets for dsl2 branch test-datasets#226)fastpmodule by nf-core fastp modulenextflowVersion >= 20.10.0for now, assuming release, although singularity requires>=20.11.0-edge)fix non-reproducible centrifuge error (on CFC and locally), any changed dependencies?Use of local
Bowtie2modules instead of using existing nf-core modules:bowtie2_removal_build/align: different behaviour -> mapped reads are discarded for contamination removalbowtie2_assembly_build/align: since the index need to be build for all assemblies, I needed to hand over ametamap to the index building process as well to avoid complicated channel logic before and after calling the moduleUse of local
quastmodules instead of using nf-core modules:quast,quast_bins: metaQUAST is used (not QUAST), with different processing for binsI observed a non-reproducible centrifuge error (which disappears when using
-resume) on different systems, which I did not observe with previous pipeline versions. However, since I could not reproduce it recently I will postpone this.Since there are so many changes, I don't know what else to describe here. Let me know, if I should add something.
PR checklist
scrape_software_versions.pynf-core lint .).nextflow run . -profile test,docker).docs/usage.mdis updated.docs/output.mdis updated.CHANGELOG.mdis updated.README.mdis updated (including new tool citations and authors/contributors).