What is the expected behavior?

From a calibration point of view, pulse amplitude is better to be represented in (float |amp|, float phase) rather than complex amp. Usually I tune pulse amplitude by amplitude sweep followed by error amplification, and then calibrate the phase separately. I often write wrapper function of one in the pulse library to convert (amp, phase) into complex value as an input, but changing the argument of those functions is more straightforward. Since they are public function, is it worth changing the interface?