push_valuesets should take an instance of ValueSest

[bdrillard]

The push_valuesets function is defined as taking the name of an ontologies database for the purpose of discovering the valuesets [1]. This may be heavy-handed in instances where datascientists want to work with experimental instances of ValueSets, as it would require them to save tables having new versions (which might not make sense of "working" instances of ValueSets).

We should open the Python API to allow the user to provide an instance of ValueSets directly. This change could be performed passively by still defaulting the database parameter to "ontologies" but performing an instance check against the class ValueSets or loading the valuesets from the source database if a the argument is a string.

This issue may make sense to include in the work of #25.

[1] -- https://github.com/cerner/bunsen/blob/master/python/bunsen/stu3/valuesets.py#L57

[rbrush]

Note that users of the Python API can already pass code values explicitly for a ValueSet, as seen in the push_valuesets usage at [1]. Users could just have an array of the code values they're interested in in the notebook and edit that inline for experimental values.

That said, I think does make sense to further overload push_valuesets to allow users to pass in ValueSets directly as you describe if it's more convenient for that workflow.

[1]
https://github.com/cerner/bunsen-tutorial/blob/master/data_engineering_tutorial.ipynb