I’d like to run a node function my_fun only after a recon-all node has finished processing. However my_fun does not require any of the recon-all outputs, so I’m not sure how to connect them in my workflow.
I thought maybe there is a way to check if the recon-all node has finished processing, and feed this in my_fun? Is this possible?
To ensure my_fun is run after recon-all, just have an input that receives one output from recon-all and then discard it in the function. Having that connection will force nipype to evaluate my_fun only after recon-all finished.
It is one the roadmap, but so far you have to connect some dummy data to your my_fun to trigger the function at the end of recon-all and to avoid parallel processing.
nipype is a dataflow framework, so data has to flow. as @oesteban suggested, you can trigger a node as a result of something else. typically the reason you want to run something afterwards is because you are waiting for “something” to finish. just let that “something” then trigger the node.