@history endpoint normalization

@history endpoint returns a list of changes on the object which may be different depending on the action. Right now the specific attributes for each kind of change are normalized at the same level of the history information which makes difficult to validate as a json schema and scale to more kind of possible actions.

My proposal that we implemented on guillotina is:

[{
    'actor': ID_USER,
    'comments': COMMENTS,
    'time': TIME_OF_THE_ACTION,
    'title': TITLE,
    'type': 'workflow',
    'data': {
         'action': ID_ACTION,
         'review_state': REVIEW_STATE,
    }
},
{
    'actor': ID_USER,
    'comments': COMMENTS,
    'time': TIME_OF_THE_ACTION,
    'title': TITLE,
    'type': 'version',
    'data': {
         'version': ID_VERSION,
         'may_revert': IS_POSSIBLE_TO_REVERT,
    }
}]