If sample_shape is used, mark Random Variable as is_datamodel

[jcozar87]

The context of datamodel is used to expand the RandomVariable when required. Optionally, the user might find easier to use sample_shape instead of using the context datamodel.

Multiple tests should be done to test this behaviour. Moreover, at least one change must be done in order to make this working correctly: if sample_shape is used, the Random Variable should change the attribute is_datamodel=True.

[jcozar87]

This is discouraged, because we need to know which variables are in datamodel. As one variable, which is child of another whose sample_shape has been provided, is a datamodel too, we need to label it as datamodel=True.

In this way, it is easier for the user to use the datamodel context, and, if he wants to specify a datamodel size, the argument for the context can be provided (or not).