Feature/python inline task

[inv-ajin]

Pull Request type

• Bugfix
• Feature
• Refactoring (no functional changes, no api changes)
• Build related changes
• WHOSUSING.md
• Other (please describe):

NOTE: Please remember to run ./gradlew spotlessApply to fix any format violations.

Changes in this PR

This PR implements support for Python as a scripting language for inline tasks in Netflix Conductor, based on the roadmap outlined here.
The feature extends Conductor's usability by enabling lightweight Python code execution as inline tasks.
The Python environment used for executing scripts is Jython.
Key changes include:

• Added a new inline task type that supports executing Python scripts (previously only JavaScript was supported).
  

• Introduced a new dependency for Jython in the build.gradle file inside conductor/core.
  

• Implemented a new Evaluator to execute Python scripts located in conductor/core/java/core/executions/evaluators.
  

• Added a new task type Python in the Java SDK.
  

Issue #

Alternatives considered

One alternative approach is to use GraalVM's Python engine instead of Jython, which could offer improved performance and compatibility.

Describe alternative implementation you have considered

[v1r3n]

Thanks for the PR. I would suggest take a look at the GraalVM implementation that is supposed to offer better performance overall.

https://www.graalvm.org/python/

It will also be useful to add support have pre-defined set of libraries added to the engine, which will allow working with 3P libraries that are statically defined and can be added at runtime.

[v1r3n]

How about using something like https://www.graalvm.org/python/

[inv-ajin]

@v1r3n
Thank you for the feedback! We are planning to implement the Python inline task using GraalVM as suggested,
I'll keep you updated on the progress.

[inv-ajin]

Hi,
@v1r3n I've made the requested changes and implemented using GralVm. Could you please review it again?

[dcore94]

Sorry for jumping in but I'd like to ask that one think that I'm missing is the possibility that expression may be an array of strings where each array element represents a line of the expression. This would make writing the expression in JSON documents much much easier and more maintainable than using \n. Applies in general to INLINE Tasks...

[v1r3n]

I am going to merge this for now @dcore94 I think the suggestion to support multiple lines is a good one. I also wonder if there is a way to also reference external files etc.

@inv-ajin I suggest if you can followup with a PR to add additional modules (could be part of the build process or env variables etc at the startup) that will really improve the usability.
e.g. I would personally like to import langchain (https://github.com/langchain-ai/langchain) libraries so I can write LLM apps with python system task worker.