Allow an algorithm to report back to TPL errors when running

[alastair]

We should have a clear way for algorithms to report back to the TPL when they fail to run so that the TPL can set the actionStatus of the ControlAction to Failed.

I see that we are using subprocess.run and catching a blanket exception in order to set the failed status:

tpc/tpl/application.py

Line 543 in 4624129

except:

subprocess.run can raise an exception when using check_call which I think is a good way of capturing this status.

We should be more careful about this exception handler, it's generally not a good idea to catch a blanket Exception, we should identify each potential error and handle these situations.