[WIP] Spectroscopy and calibrations integration V2

[eggerdj]

Summary

This is a WIP PR designed to show and discuss how calibrations and experiments could be integrated through the analysis class. An alternative option is shown in #80.

Details and comments

The user would run the calibration experiment as follows:

calibrations = BackendCalibrations(backend)
...
qubit = 5
qubit_freq = calibrations.get_qubit_frequencies()[qubit]
frequences = np.linspace(qubit_freq - 10e6, qubit_freq + 10e6)
spec = QubitSpectroscopy(qubit, frequencies)
spec.set_analysis_options(calibrations=calibrations, force_update=True, calibration_group="my_group")
spec.run(backend)

The pros of this integration are:

• The BackendCalibrations class will remain lightweight.

The cons of this integration are:

• The analysis class may not be reusable for other types of peak fits unless extra logic will be needed to support other peak-fit experiments such as spectroscopy on 1<->2.
• The user needs to write a lot of code.

[yaelbh]

When you compare the number of lines that the user has to write in #79 and #80, note that there is something "unfair" for #79: #79 contains two lines for calculating frequencies, while frequencies is just None in #80. So in the code snippet above, these two lines should be removed, for a fair comparison.

By the way, In both #79 and #80, if we want to calculate the frequencies from the calibrations, we can do it internally for the user.

[eggerdj]

Closed by #88