It basically just uses the default EncodedNumber.BASE, irrespective of the actual base that was used to encode that particular encrypted number.
See also the problem raised in issue #75: If you run the alternative_base.py example with numbers that have a non-zero exponent, then the addition will produce a wrong result.
decrease_exponent_to
contains this line: https://github.com/n1analytics/python-paillier/blob/955f8c0bfa9623be15b75462b121d28acf70f04b/phe/paillier.py#L592It basically just uses the default EncodedNumber.BASE, irrespective of the actual base that was used to encode that particular encrypted number.
See also the problem raised in issue #75: If you run the
alternative_base.py
example with numbers that have a non-zero exponent, then the addition will produce a wrong result.