Comparison vs DL4J / java-cpp

[Robbie-Palmer]

FLIP-206 compares PemJa to Jython, GraalVM, JPype and Jep

deeplearning4j's Python4J framework seems more comparable to PemJa than any of these
It depends on javacpp-presets which bundles CPython into a Jar and depends on java-cpp to abstract from the JNI

API example:

try(PythonGIL gil = PythonGIL.lock()){
        try(PythonGC gc = PythonGC.watch()){
            List<PythonVariable> inputs = new ArrayList<>();
            inputs.add(new PythonVariable<>("x", PythonTypes.STR, "Hello "));
            inputs.add(new PythonVariable<>("y", PythonTypes.STR, "World"));
            PythonVariable out = new PythonVariable<>("z", PythonTypes.STR);
            String code = "z = x + y";
            PythonExecutioner.exec(code, inputs, Collections.singletonList(out));
            System.out.println(out.getValue());
        }
    }catch (Throwable e){
        e.printStackTrace();
    }

A comparison between this approach and that of PemJa would be very useful

[HuangXingBo]

@Robbie-Palmer Thanks lot for bringing up this interesting topic. I haven't looked into python4j in depth before. Just from the API example you provided, I have two questions about python4j

1. Does python4j support the invocation of python objects or methods(invoke and invokeMethod in pemja), not just the execution of entire code fragments exec.
2. Does python4j support calling back java in python code?

[Robbie-Palmer]

@HuangXingBo I believe the answer to both of those is no! So those are definitely pros versus python4j

Another key difference (with pros and cons) is python4j comes with the Python VM bundled inside the Jar, while PemJa depends on interfacing with an external Python VM

Pros:

• Can deploy as a simple Jar without environment changes

Cons:

• Python version is fixed by the version of the library
• Non standard library packages still need to be installed and the bundled Python VM pointed to them

[HuangXingBo]

@Robbie-Palmer Thanks a lot for the patient explanation. I am not sure whether python4j bundled pythonvm is a customized CPython, which will bring a lot of optimization possibilities. But from my side, dependency management is very important. It seems that the dependency management of python4j is very inconvenient.