pyunload

1. The use of this procedure is optional: the Python interpreter is automatically released at the end of a model execution. However, you may prefer to release it sooner to free resources allocated by Python and to reset the interpreter state. Moreover, releasing the interpreter makes it available for initialization in another Mosel model which is executed in the same Mosel instance.

2. Some Python extensions, like NumPy and pandas, do not work if pyinit and pyunload are called multiple times inside of a single Mosel instance. This is due to a bug in Python's finalization function. See https://docs.python.org/3/c-api/init.html\#c.Py\_FinalizeEx and https://github.com/numpy/numpy/issues/8097 for more details.