https://nf-co.re logo
Join Slack
Powered by
Final wrap up for me: Merged another modules updat...
# seqinspector
m
Final wrap up for me: Merged another modules update: https://github.com/nf-core/seqinspector/pull/215 For https://github.com/nf-core/seqinspector/pull/209 I think we just need to rethink the logic to get to the decision. We cannot access the results before the pipeline has ended. But there are ways to make this work, I'll work on it on Monday. For https://github.com/nf-core/seqinspector/pull/212 I think it's just a question of properly rerunning the tests. I'll work on it on Monday too.
👍 1
n
For https://github.com/nf-core/seqinspector/pull/209: An updated code has been pushed which now works. You can have a look if the logic to get the decision is now okay. Then we will continue with the tests later.
m
I'm on https://github.com/nf-core/seqinspector/pull/212 for now, I'll have a look at 209 before lunch, and hopefully we merge by end of day
thanks @NellyN
👍 1
https://github.com/nf-core/seqinspector/pull/212 is good to go. @Matilda Åslin I hope you don't mind me pushing some commits on your branch
m
Awesome thanks!
🎉 1
m
merged 212, I'm looking at 209 now
@NellyN so as far as I understand we can either run multiqc_global or sav depending on what we have in the input samplesheet
what are the modifications I need to make on this samplesheet to get the other option?
n
The rundir should be the runfolder itself not .tar.gz. It assumes the rundir is not zipped according to how it will run in production
m
oh I see
so if I want to test it properly I should have a untar step
I can make it work
It's not a problem
that's my personal motto
🙏 1
n
Yes, Thanks 🙏
m
and then to get the multiqcsav, it's one single rundir for the whole samplesheet, right?
n
Right...
Otherwise, if more than one rundir or if any rundir is missing for a sample, it will give a warning and run multiqc global instead
m
ok, I can do a test for that
🙏 1
I will take longer than planned, but no worries
n
Thank you so much.
We had these changes that makes the test to pass but it requires changing the samplesheets in the testdata repo to provide the rundir as a directory and not .zipped file ... and also edited the schemna_input.json file to NOT restrict the rundir to .tar.gz pattern I have set my pipelines_testdata_base_path variable in the test file below to my local test-data repo that provides the expected rundir (i.e not zipped)
NovaSeq6000.nf.test
Not sure if changing all test-data repo samplesheet csv to provide rundir folders instead of zipped .tar.gz files is an acceptable approach. Case the solution in the code above does an untar for the specified instrument runfolder So that our samplesheets matches what is expected in production.... Is that an okay solution?
m
yeah my idea would be to run untar beforehand, and have a dedicated samplesheet that has this new path in
n
For some reason the tests fail currently as its going through all the code and giving error 
The provided Path object is not a directory: /nf-core/test-datasets/raw/seqinspector/testdata/MiSeq/220422_M11111_0222_000000000-K9H97.tar.gz
Even if it's for a testcase that does not use multiqcsav
image.png
m
yeah I think the issue is that it's not a path, and that won't work with the channel manipulation you added to get the rundir info
👍 1
ok, I managed to get the extraction of the rundir in my test
🎉 1
🙏 1
Just editing the samplesheet now
🙏 1
And I managed to use it in the test
n
Awesome. Thanks
m
Now onto doing more cleanup
pushed the new logic for the test
hopefully that'll work on other people computer as well 🙏
🙏 1
ok, I think I almost got it
🎉 1
ok, this is a bigger time drain that I expected, but I am getting there, and I nicely refactored some duplicated code
🙏 1
2 Views
Open in Slack
PreviousNext
Powered by