https://nf-co.re logo
Join Slack
Powered by
Hi, as a side activity while migrating the pipelin...
# hackathon-oct-2026
d
Hi, as a side activity while migrating the pipeline to a new template, I've created a scripted migration tool: github.com/Skitionek/nf-migrate Validated it against handmade migrations performed in a bunch of repos: github.com/Skitionek/nf-migrate/blob/…/validation.md It seems that 80% of manual groundwork can be cut off - leaving the final 20% for AI to polish off! 🤖✨
p
Great! You were sat at the wrong table though eyebrows nf-co.re/hackathon-projects/hackathon-barcelona-2026/ai-template-updates
🫣 1
We've been doing the same thing all morning 🙂 Building a skill and scripts to do exactly this
But this is great - I'll point Claude at what you've done and see what we can merge
🎉 1
👍 1
d
p
I think your work will be really complementary actually
🎉 1
So my agent skill also had a bunch of deterministic scripts - looking now to see how much they overlap
We've just been discussing here and I think a good approach would be to move as much of this deterministic stuff into nf-core #CE5LG7WMB itself - so when the PR is opened it's already had these things applied and it has fewer merge conflicts to start with.
Then the skill we've been building here can just be the final touches which are different every time
d
Yes, that would be wonderful, greatly easing maintenance 🎉
I will open PR on tools with what I've got
p
Please don't!
We are doing that already 😅
@Dominik Maszczyk is there a reason you used Alembic approach instead of using git merges?
h
Reviewing the changes I found it highly educational to see the changes between templates!
p
You mean vs. this view? (which shows all of tools, not just the template?)
oh I see, you mean in the PR that each release is a commit
👍 1
but you have the same in the automated PRs already? https://github.com/nf-core/metaboigniter/pull/111
d
@Phil Ewels, some migrations involve more than text changes. Instead of patches, we can use 
nf-core modules update ...
or 
nf-core pipelines lint --fix
which gives us much more resilient results.
p
message has been deleted
@Dominik Maszczyk yup - my plan was to run those as well, but is that incompatible with git merges?
d
No, you can divide things into two mechanisms. As long as things work, it's all good- I would be curious to see your solution.
Small today's update As nf-migrator seemed to work well, as a side task, I've run it against a collection of modules. Successfully migrated and opened PRs in the following: | Pipeline | Template | PR | | --- | --- | --- | | circdna | 4.0.3 | [#105](github.com/nf-core/circdna/pull/105) | | demo | 4.0.3 | [#73](github.com/nf-core/demo/pull/73) | | hadge | 4.1.0 | [#121](github.com/nf-core/hadge/pull/121) | | lsmquant | 4.1.0 | [#90](github.com/nf-core/lsmquant/pull/90) | | multiplesequencealign | 4.0.3 | [#255](github.com/nf-core/multiplesequencealign/pull/255) | | proteinfold | 4.1.0 | [#660](github.com/nf-core/proteinfold/pull/660) | | rarevariantburden | 4.0.2 | [#16](github.com/nf-core/rarevariantburden/pull/16) | | rnadnavar | 4.1.0 | [#100](github.com/nf-core/rnadnavar/pull/100) | | sammyseq | 3.5.1 | [#133](github.com/nf-core/sammyseq/pull/133) |
👀 1
m
Hi @Dominik Maszczyk, thank you for including the sammyseq pipeline in your migration effort! I wanted to review your PRs but unfortunately I was not in Barcelona and I missed some of the development, can we have a chat to guide me in understanding the changes?
I probably need the help of @Phil Ewels to understand how to proceed
Open in Slack
PreviousNext
Powered by