Use scalable variant in templates

[nelimee]

This PR changes how scaling is performed in tqec.

This fix #241.

Before

Before that PR, the Dimension class was storing a LinearFunction describing how the Dimension was supposed to scale and a value representing the current value of the Dimension.

The Template instance were taking the Dimension instances as input, and were using them to scale internally, but were only outputting integers resulting from the "instantiation" of such Dimension instances.

In this PR

This PR completely removes the Dimension class. Template instances are now responsible for storing the scaling parameter k along with the LinearFunction instances needed to scale.

The Template class now requires the implementation of a new scalable_shape property that returns the shape of the Template instance as a fully scalable structure. It basically returns the full information (the full function) instead of returning its value on one specific point (before). For that, ComposedTemplate needed a PiecewiseLinearFunction class due to the use of min/max in the computation of the methods needed for scalable_shape.

As side (but important) notes:

• the ShiftedTemplate and StackedTemplate classes are removed from the code base, because they are not currently used and are more classes that would need maintenance.
• all the JSON-related methods are also removed, due to their replacement in #240.

[review-notebook-app[bot]]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[nelimee]

This is ready for review.