spectroscopy experiments: set central frequency and frequency range as input parameters

[RoyStegeman]

The current implementations of resonator spectroscopy, resonator punchout and qubit spectroscopy perform a scan around a central frequency. This frequency is determined from stored calibration parameters:

• spectroscopy: bare_frequency if power_level='high' or dressed_frequency if power_level='low' from calibration.json
• punchout: probe frequency from parameters.json.

The documentation just mentions this as being the "readout frequency", so understanding what happens even requires looking at the code.

This introduces some issues:

1. experiments used to determine frequencies depend on those same frequencies
2. reduced flexibility: for example if the user is iterating and wants the center frequency to be in the middle between bare and dressed frequencies in a punchout experiment
3. lack of transparency, see e.g. how the central frequency is determined for punchout.

proposed solution: take the central value or min/max range as input parameters.

[alecandido]

This is part of the discussion on suitable defaults, as described in #1381 (comment).

The choice is not necessarily bad, but I share your critics about lack of transparency and flexibility.
The general principle should be the one I exposed in the cited comment, and I also proposed a different uniform interface for sweeps in #1300.

However, maintaining the current set of defaults for frequencies is not necessarily bad. But only as defaults, and for sure they should be documented.