Running Multiple Instances of AutoscoperM Locally

[QYetman]

We’re trying to run an optimization on the filter settings using a particle swarm optimization implemented in Python. Using Microsoft MPI and mpi4py to parallelize it and PyAutoscoper to communicate between the Python script and AutoscoperM running through Slicer (I have been opening all the instances of AutoscoperM I need at the start of the run). The issue is that I can’t seem to find a way to communicate with multiple instances of AutoscoperM running locally. AutoscoperConnection(address, …) seems to be hardcoded to a specific port that AutoscoperM always uses.

Spoke with JC at the most recent Autoscoper meeting. It was suggested that an option to specify the port when starting AutoscoperM could be added. He also mentioned code that could be used to find a free port for AutoscoperM to use. The AutoscoperConnection() class could then have the option to specify the port as well.

Let me know if you require any further information.