Plugging new parametrization inside instrumentation and optimization

[jrapin]

Types of changes

• Docs change / refactoring / dependency upgrade
• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)

Motivation and Context / Related issue

How Has This Been Tested (if it applies)

FYI: this is expected to land after January 8th

This plugs parametrization as created in #323
To do so, the new Instrumentation is extended with old Variable methods, and the old Variable inherits from Instrumentation  and overrides the new methods. 

This is not expected to change anything for you, but some printed information may change, and deprecation warnings will appear for some methods (Eg.: set_cheap_constraint_checker -> register_cheap_constraint with a slightly different API). Some probably unused functions will disappear (get_description) making the changes breaking anyway. At this point you will be able to use new Parameter classes, but their initialization may be inaccurate so it will not be advised in some cases (you would get a user warning). Also, because of the compatibility, we expect optimizers to be about 25% slower in this version. The optimizer will still only take an Instrumentation (new version) as input for its initialization.

Checklist

• The documentation is up-to-date with the changes I made.
• I have read the CONTRIBUTING document and completed the CLA (see CONTRIBUTING).
• All tests passed, and additional code has been covered with new tests.

[jrapin]

For the record:
Locally pytest nevergrad/optimization has gone up from 45s to 70s. Next PR will hopefully revert the change, which is due to the need of keeping both compatibilities.