public class PythonExecutioner
extends Object

Allows execution of python scripts managed by an internal interpreter. An end user may specify a python script to run via any of the execution methods available in this class. At static initialization time (when the class is first initialized) a number of components are setup: 1. The python path. A user may over ride this with the system property DEFAULT_PYTHON_PATH_PROPERTY 2. Since this executioner uses javacpp to manage and run python interpreters underneath the covers, a user may also over ride the system property JAVACPP_PYTHON_APPEND_TYPE with one of the PythonExecutioner.JavaCppPathType values. This will allow the user to determine whether the javacpp default python path is used at all, and if so whether it is appended, prepended, or not used. This behavior is useful when you need to use an external python distribution such as anaconda. 3. A proper numpy import for use with javacpp: We call numpy import ourselves to ensure proper loading of native libraries needed by numpy are allowed to load in the proper order. If we don't do this, it causes a variety of issues with running numpy. (User must still include "import numpy as np" in their scripts). 4. Various python scripts pre defined on the classpath included right with the java code. These are auxillary python scripts used for loading classes, pre defining certain kinds of behavior in order for us to manipulate values within the python memory, as well as pulling them out of memory for integration within the internal python executioner. For more information on how this works, please take a look at the init() method. Generally, a user defining a python script for use by the python executioner will have a set of defined target input values and output values. These values should not be present when actually running the script, but just referenced. In order to test your python script for execution outside the engine, we recommend commenting out a few default values as dummy input values. This will allow an end user to test their script before trying to use the server. In order to get output values out of a python script, all a user has to do is define the output variables they want being used in the final output in the actual pipeline. For example, if a user wants to return a dictionary, they just have to create a dictionary with that name and based on the configured PythonVariables passed as outputs to one of the execution methods, we can pull the values out automatically. For input definitions, it is similar. You just define the values you want used in PythonVariables and we will automatically generate code for defining those values as desired for running. This allows the user to customize values dynamically at runtime but reference them by name in a python script.