YAQ about reloading python modules in Matlab after python changes

I understand that there have been a number of questions regarding issues with Matlab not reloading python modules after they have been changed. Based off other answers, I have an acceptable solution, but I think I am hitting a bug.

My python:

class myTestClass:
    def __init__(self, somestr):
        self.loc = 1234
        print("Init called")

    def __del__(self):
        print("Destructor")

    def myTestFunc(self):
        a = 456
        print(f'My Test Func called loc={self.loc} a={a}')

My Matlab:

%clear classes

myMatlabTestClass = py.importlib.import_module('myMatlabTestClass');
py.importlib.reload(myMatlabTestClass);

pyclass = myMatlabTestClass.myTestClass("hello");
pyclass.myTestFunc()

The problem:

The above is sensitive to python code changes if I change the value of self.loc, but changing the value of a in myTestFunc requires the inclusion of clear classes.

I fear that clear classes will be too expensive and may impact other code, depending on where it is called.

I'd like to understand why a and self.loc would be treated differently here and if there is a workaround.

[Alexander]

Not quite what I wanted, but this code works flawlessly:

myMatlabTestClass = py.importlib.import_module('myMatlabTestClass');
py.importlib.reload(myMatlabTestClass);

pyrun("import myMatlabTestClass")
pyrun("myClass = myMatlabTestClass.myTestClass('somestr')")
pyrun("myClass.myTestFunc()")

Answer 1

[SHAQLAIN]

class myTestClass:
    def __init__(self, somestr):
        self.loc = 1234
        print("Init called")

    def __del__(self):
        print("Destructor")

    def myTestFunc(self):
        a = 456
        print(f'My Test Func called loc={self.loc} a={a}')

[Walter Roberson]

Could you point out the difference between this and what was originally posted?

[Alexander]

The problem is not the python code, it is in matlab reloading the python code when changes are made to it.