Pixi tasks that depend on files whose path is specified in a settings.toml

[concimuscb]

I normally define input paths of files in a separate settings.toml.

On the other hand, the idea of using pixi tasks to run entire pipelines and skip steps which can be skipped depending on inputs/outputs sounds great.

The issue is that the two seem incompatible. From what I could gather, when defining tasks, you have to hard code the names of the files they depend on. I can hardcode the names of the scripts and intermediate datasets but, many times, the initial dataset will change name.

This leaves me with the option of having to move all the settings.toml inside the code, which I hate; or not being able to use pixi tasks.

Is there any way I can reconciliate both workflows?

[ruben-arts]

Right now yes, you can be more flexible with the use of [activation.env] variables, but we are working towards actual arguments for the tasks soon.

[concimuscb]

Sounds amazing. Thank you for the great tool I am already telling all Python users I know about it!