How do pipelines deal with issues that are largely LLM output and where the person aiming the LLM may not fully grasp whats going on? I try to respond, but I wonder how much effort I should put into explaining whats going on, since the issue, by nature of coming from an LLM, looks very nice and sophisticated, but given that I assume it was generated, my motivation to write things that will then be turned into a prompt is very low?

Can you link the issue? I am curious.

Wait for the first human answer and you'll undersstand 😛

you can point them to https://github.com/nf-core/tools/pull/4211/changes but don't put too much effort in it, can be quite draining.

I was mainly wondering how others deal with this, I am not intending to put in a lot of time, and tbh I dont expect that the people who generate issues will care much about the guidelines 😕

add a

awaiting-feeback

label and just see what happens (these labels should be picked up during clean-ups)

I would ask a question, if that isnt answered personally I'd ignore the issue if it doesnt help others as well.

Yeah, ok that is also kind of my approach. Idk why, but replying to LLM output just feels like a waste of time

Should be alright if there is a legitimate issue to be fair

There is a "legitimate" issue, when trying to run the local module outside of the pipeline, because the module depends on a file in

bin

. This is however not the intention of this module.

we keep getting this in tools as well, part of open source development atm, I am afraid. But feel free to close if it is too much useless back and forth

Yeah, probably more suitable for #C07KND2JP54 🙂

the thing is how can you tell it's LLM output? It's seems to me it's only an issue well formulated (the different parts originating from the nf-core github issue template)

Thats a fair question. It is hard to explain, because it is very subjective and of course there is no proof. I am probably also quick to assume that things are LLM generated. In this case, it is the language / sound, how certain statements are phrased, the way the code is annotated.

Actually I think you're right, it does look like a LLM output. Anyway, it's worth investigating since there seems to be a real bug. At least the description in this issue is quite complete ^^

Well, thats the thing, there is no bug in the pipeline,

genomescope.R

is in pipeline

bin/