Add entangling_layer to models.encodings

[renatomello]

Checklist:

• Reviewers confirm new code works as expected.
• Tests are passing.
• Coverage does not decrease.
• Documentation is updated.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 99.95%. Comparing base (a603ddd) to head (0b9355f).

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #1260   +/-   ##
=======================================
  Coverage   99.95%   99.95%           
=======================================
  Files          72       72           
  Lines       10454    10490   +36     
=======================================
+ Hits        10449    10485   +36     
  Misses          5        5           

Flag	Coverage Δ
unittests	99.95% <100.00%> (+<0.01%)	⬆️

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

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[MatteoRobbiati]

Super fine with this kind of feature!

I think it would be nicer to start thinking how to move these features (or, better, if we want this) to the qiboml package.

Personally, I think this is exactly that kind of feature (I mean the encodings in general) which would fit perfectly with it.

I would say this because it could be crucial to understand how to organize the code structure and because I am currently working on the variational model and, consequently, on the encoding and decoding strategy.

[renatomello]

Super fine with this kind of feature!

I think it would be nicer to start thinking how to move these features (or, better, if we want this) to the qiboml package.

Personally, I think this is exactly that kind of feature (I mean the encodings in general) which would fit perfectly with it.

I would say this because it could be crucial to understand how to organize the code structure and because I am currently working on the variational model and, consequently, on the encoding and decoding strategy.

Data encoders are very important for QML, but not exclusive to QML. For instance, it's important for "regular" QFT and / or Quantum Phase Estimation applications (e.g. the example of unary encoding application to option pricing included in qibo's documentation, or the paper(s) in which the current implementation of the unary encoder was based on.). After all, a data encoder is nothing but a state preparation, so the applications are broader than QML.

[MatteoRobbiati]

Data encoders are very important for QML, but not exclusive to QML. For instance, it's important for "regular" QFT and / or Quantum Phase Estimation applications (e.g. the example of unary encoding application to option pricing included in qibo's documentation, or the paper(s) in which the current implementation of the unary encoder was based on.). After all, a data encoder is nothing but a state preparation, so the applications are broader than QML.

Thanks for this reply! I see, this perfectly makes sense. So the best way will be to integrate the encodings into qiboml, picking them up from Qibo. My only concern is not to make redundancy between the two packages (but at this point this has to be done in qiboml).

[MatteoRobbiati]

A final review, with a small suggestion follows.

[MatteoRobbiati]

Since you use this to define a Circuit, I think a similar error will be raised directly from line 332.