Classifiers

[Edoardo-Pedicillo]

This PR ports the classifiers' module to the autocalibration framework.

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

Merging #447 (93f721c) into main (11e9198) will increase coverage by 0.11%.
Report is 2 commits behind head on main.
The diff coverage is 98.03%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #447      +/-   ##
==========================================
+ Coverage   95.83%   95.95%   +0.11%     
==========================================
  Files          55       61       +6     
  Lines        4106     4425     +319     
==========================================
+ Hits         3935     4246     +311     
- Misses        171      179       +8     

Flag	Coverage Δ
unittests	95.95% <98.03%> (+0.11%)	⬆️

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

Files Changed	Coverage Δ
src/qibocal/fitting/classifier/scikit_utils.py	83.33% <83.33%> (ø)
src/qibocal/fitting/classifier/qubit_fit.py	96.42% <96.42%> (ø)
src/qibocal/fitting/classifier/run.py	97.52% <97.52%> (ø)
src/qibocal/fitting/classifier/data.py	100.00% <100.00%> (ø)
src/qibocal/fitting/classifier/linear_svm.py	100.00% <100.00%> (ø)
src/qibocal/fitting/classifier/naive_bayes.py	100.00% <100.00%> (ø)
...bocal/protocols/characterization/classification.py	100.00% <100.00%> (ø)
...zation/readout_optimization/resonator_frequency.py	100.00% <100.00%> (ø)
src/qibocal/protocols/characterization/utils.py	95.04% <100.00%> (+0.21%)	⬆️

[rodolfocarobene]

Thanks @Edoardo-Pedicillo, the code looks good to me (and finally it's fast!), I have just a couple of general comments:

• we should add a detailed explanation of the routine in the documentation, with a list of the classifiers and a general explanation. For example, I did not really know which classifiers were available. Also, I have no idea what a ROC curve is or what AUC stands for... Maybe the names on the axis could help, but in any case I would also go for a documentation page
• Maybe the plotting of the ROC curves and of the benchmarks should be plotted only if more than one classifiers is used
• I got some warnings, I don't know if you can fix them / suppress them in some way. otherwise good anyway

The keyword argument 'parallel=True' was specified but no transformation for parallel execution was p
ossible.

To find out why, try turning on parallel diagnostics, see https://numba.readthedocs.io/en/stable/user
/parallel.html#diagnostics for help.

File "../qibocal/src/qibocal/protocols/characterization/utils.py", line 175:
@njit(["float64[:] (float64[:], float64[:])"], parallel=True, cache=True)
def cumulative(input_data, points):
^

  warnings.warn(errors.NumbaPerformanceWarning(msg,

/home/rodolfoc/venv/qibo/lib/python3.10/site-packages/skl2onnx/algebra/onnx_ops.py:159: UserWarning
: OpSchema.FormalParameter.typeStr is deprecated and will be removed in 1.16. Use OpSchema.FormalPa
rameter.type_str instead.
  tys = obj.typeStr or ''
/home/rodolfoc/venv/qibo/lib/python3.10/site-packages/skl2onnx/algebra/automation.py:154: UserWarni
ng: OpSchema.FormalParameter.isHomogeneous is deprecated and will be removed in 1.16. Use OpSchema.
FormalParameter.is_homogeneous instead.
  if getattr(obj, 'isHomogeneous', False):
/home/rodolfoc/venv/qibo/lib/python3.10/site-packages/jinja2/environment.py:485: UserWarning: OpSch
ema.FormalParameter.typeStr is deprecated and will be removed in 1.16. Use OpSchema.FormalParameter
.type_str instead.
  return getattr(obj, attribute)

[rodolfocarobene]

There are some conflicts, but in general looks good to me

[andrea-pasquale]

Thanks @Edoardo-Pedicillo!
Generally it looks good to me, I tested it with both RFSoC and qblox and it works.
A few minor comments.
Could you please also edit the example runcard by adding the new version of this protocol?

[andrea-pasquale]

Thanks for all the updates @Edoardo-Pedicillo.

I have added the show_contour parameter.

I think that maybe it would be better to show or not show directly the contour by toggling it on and off in the legend. By going to the previous I meant to show just the classification and the raw data without anything else (possibly by increasing the size of the plot).
The problem with having the show_contour parameter is that in order to show it or not show it you will need to rerun the experiment again.

I also have a small comment down below.

[andrea-pasquale]

Thanks @Edoardo-Pedicillo