Fix ramsey update

[andrea-pasquale]

The drive frequency is now updated only when a detuning is provided, which is one of the assumption of the fit. When detuning is 0 we update T2 since we expect it to be more reliable.

Now Ramsey can be used to monitor T2 in a more stable way given that the fit will include both oscillations and an exponential decay.

Checklist:

• Reviewers confirm new code works as expected.
• Tests are passing.
• Coverage does not decrease.
• Documentation is updated.
• Compatibility with Qibo modules (Please edit this section if the current pull request is not compatible with the following branches).
  • Qibo: master
  • Qibolab: main
  • Qibolab_platforms_qrc: main

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 97.50%. Comparing base (c4fdce8) to head (4902a18).
Report is 6 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #976   +/-   ##
=======================================
  Coverage   97.50%   97.50%           
=======================================
  Files         123      123           
  Lines        9701     9701           
=======================================
  Hits         9459     9459           
  Misses        242      242           

Flag	Coverage Δ
unittests	97.50% <100.00%> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
src/qibocal/protocols/ramsey/ramsey.py	99.00% <100.00%> (-0.09%)	⬇️
src/qibocal/protocols/ramsey/ramsey_signal.py	97.54% <100.00%> (-0.10%)	⬇️
src/qibocal/protocols/ramsey/utils.py	100.00% <100.00%> (ø)

[alecandido]

In general, it would be better to use None as default (and type is already Optional which is the union of None with the thing being parametrized, not there to signal it has a default).

In this way, it would be smooth, and working in the same way for a 0. or very small detuning.

All the behavior currently attributed to 0. should be moved to None, including the update (with float is also possible to use nan, but I wouldn't recommend it: nan is not always an intentionally missing value, but possibly the result of an undefined float operation - it's also sometimes non-trivial to specify, since it lacks a literal, and identify, because of "weird" comparison).

[GabrielePalazzo]

Shall we log what we are updating in the runcard in case of conditional updates (that depend on fitting)?

[andrea-pasquale]

This is on the roadmap #910
For the time being I would keep those manipulations directly in the python file where we are handling scripts.

[andrea-pasquale]

In general, it would be better to use None as default (and type is already Optional which is the union of None with the thing being parametrized, not there to signal it has a default).

In this way, it would be smooth, and working in the same way for a 0. or very small detuning.

All the behavior currently attributed to 0. should be moved to None, including the update (with float is also possible to use nan, but I wouldn't recommend it: nan is not always an intentionally missing value, but possibly the result of an undefined float operation - it's also sometimes non-trivial to specify, since it lacks a literal, and identify, because of "weird" comparison).

Thanks @alecandido in the last commits I've changed the default behavior to None