This integration in VPM has been done via a generic Python server (PyVPMlauncher v1r0) using the Python3 based pyserver template and able to start/stop any executable via Popen()/Popen.kill() and periodically monitoring its execution state via Popen.poll().
After some tuning effort, mostly needed to make the cWB machinery more user independent (it is now running as virgorun), the two main top cWb processes run.py (on olserver145 machine) and run_ts.py (on submit1 machine) can now be managed from the Low Latency Analysis VMP instance (Fig. 1).
Further levels of integration could be studied for cWB but this fast solution has allowed to quickly integrate in VPM applications developed for a different environment without making changes into them.
More efforts are also needed in order to make the cWB software installation and configuration more "Cascina compliant" and this will be carried out in view of the September OPA challenge.
To be noted that the PyVPMlauncher application is completely generic and could be used for other similar needs of integrating external applications in VPM.
