How pickle changes according to how the interpreter gets called

pickle_deal.py

"""Pickle dumps test.

When pickle dumps a class, it will include the FQDN for the module. In this example,
we see how an imported class gets dumped including its module information, whereas
a local one has the current module name, which could be `__main__` or other depending
on how the interpreter was called!

If you save this file as `pickle_deal.py` and then in a python terminal you import it
`import pickle_deal`, you will see the following printed:
b'\x80\x04\x95)\x00\x00\x00\x00\x00\x00\x00\x8c\x0bpickle_deal\x94\x8c\x07Decimal\x94\x93\x94\x8c\x07123.456\x94\x85\x94R\x94.'
b'\x80\x04\x95%\x00\x00\x00\x00\x00\x00\x00\x8c\x07decimal\x94\x8c\x07Decimal\x94\x93\x94\x8c\x07123.456\x94\x85\x94R\x94.'

If you otherwise run the file with `python3 pickle_deal.py`, you will see the same as
running this code in a python interpreter:
b'\x80\x04\x95&\x00\x00\x00\x00\x00\x00\x00\x8c\x08__main__\x94\x8c\x07Decimal\x94\x93\x94\x8c\x07123.456\x94\x85\x94R\x94.'
b'\x80\x04\x95%\x00\x00\x00\x00\x00\x00\x00\x8c\x07decimal\x94\x8c\x07Decimal\x94\x93\x94\x8c\x07123.456\x94\x85\x94R\x94.'

This implies that when pickling local classes, one must be careful on the way the
module gets called, otherwise the result of the pickling could differ.
"""
import pickle
from decimal import Decimal as NonLocalDecimal


class Decimal(NonLocalDecimal):
    pass


value = '123.456'

print(pickle.dumps(Decimal(value)))
print(pickle.dumps(NonLocalDecimal(value)))