This message was deleted.
# shameless-promo
s
This message was deleted.
p
Nice, will take a look!
Do you think that there's scope for MultiQC to make these tables etc. itself directly from the inputs, without the need of any complexity on the pipeline side?
s
yes absolutely LOL but, in my mind, the more difficult part about that is deciding on how the data should be formatted inside, and coming out of, Nextflow; https://github.com/nextflow-io/nextflow/discussions/4925 IMO, the old style of using the embedded "version YAML" block in each Nextflow script is mostly deprecated with the new topic channels +
eval
output type however it seems to leave a void in deciding how that same data should now be formatted its on my To-Do list for that repo to also implement a method that can handle a plain Map interface, or maybe I could even expose the
ProcessVersion
class interface directly https://github.com/stevekm/nf-niceutils/blob/ca7fe6470123f5b46854f12ec55e001ae8c2d[…]0cb/src/main/groovy/com/nextflow/plugin/ExampleFunctions.groovy , so that users might be able to immediately bind their versions into a coherent object format inside their pipeline, which could then also be serialized for external use e.g. with MultiQC
s
ultimately, I believe the most elegant solution, would be for Nextflow to embed the entire "versions" functionality into Nextflow itself, so that something like this
Copy code
process FASTQ {
  container "docker/fastqc-1.2"

  input:
  path(files)

  output:
  path(report)

  version:
  val("FastQC"), eval('fastqc --version | sed -e \'s/FastQC v//g\'')

  script:
  """
  multiqc ...
  """

}

workflow QC_WORKFLOW{
  take:
  files_ch

  main:
  FASTQ(files_ch)
}

workflow {

  
QC_WORKFLOW(files_ch)
QC_WORKFLOW.version.view()

// [ ["software": "FastQC", 
//     "version": "1.2", 
//      "container": "docker/fastqc-1.2"
//      "process":"FASTQC", 
//       "id": "QC_WORKFLOW:FASTQC"] ]

}
This would make it easier for both MultiQC, and other methods, to access the versions
p
You mean like this issue that Paolo and I wrote together over a coffee in 2018? šŸ˜…
This issue was duplicated here and has had quite a bit of activity in the past year.
p
Ah, @Ben Sherman just closed them all because they are already fixed by the new
eval
directive. I thought that they had! But thought that I was going crazy as I couldn't remember what we called it and couldn't see it referenced there šŸ˜…
s
yea did not realize it was so heavily discussed already, its weird because when I searched for "versions" on the github issues, I did not see any of these. Well as long as its been reviewed before its all good šŸ™‚
p
Well we haven't actually shipped anything yet, this still needs to be brought over into nf-core style and guidelines. There was a discussion about how
eval
should be implemented but I don't think that there's been a serious effort to figure out how to update modules and the template yet.
b
I have two PRs against rnaseq for the channel topics and eval output: • https://github.com/nf-core/rnaseq/pull/1109 • https://github.com/nf-core/rnaseq/pull/1115 I think they're waiting for the next stable release (24.04) before incorporating these features into nf-core, and indeed we haven't figured out what to do with the modules
gratitude thank you 1
this 1