Has anyone taken a look at how Snakemake does this...
# wg-meta-pipelines
e
m
In a way, isn't that what is being aimed for here? nf-core tools fetches a copy of the code to be used as a subworkflow. We then use
include
to use the subworkflow. It's a bit more difficult for us because we need the channel outputs accessible, whereas the snakemake users have their folder structure to work on. It's easier to map to, whereas we would need to define channels for all possible outputs. I think this is where my approach becomes similar but you still have to run the entire workflow (or disable parts with flags or ext.when), where as with snakemake you just use
use