[core] Setting loglevel does not work for quantization

Quantization doesn't seems to be impacted by logLevel.

Further investigation show that current logLevel is defined as a static member. However each python module are built separatly and thus does not share the same symbol table. Thus setting the consoleLevel with the aidge_core module does not impact the quantization module (or any other C++ module, pure python module work has they use the aidge_core binding).

To remediate to this, I propose to use python capsule system to share data across modules.