Make inputs + outputs consistent, provide general serializer

[jsstevenson]

Feature description

• Different fetchers handle things like text normalization slightly differently. I think we should always just be forcing our expected enums.
• Make sure that inputs for things like search terms always look the same (e.g. as types)
• Make a type util kind of method to recursively dump something like nested namedtuples and enums into a more easily jsonable type.

Use case

End users may want simpler datatypes when eg merging outputs into a dataframe

Acceptance Criteria

Open ended, but

• do a review of all top level fetch methods and identify + resolve inconsistencies
• 'normalization' should always happen (no optional)
• make a jsonify() or dictify() method that strips namedtuple and enum types

Proposed solution

No response

Alternatives considered

No response

Implementation details

No response

Potential Impact

No response

Additional context

No response

Contribution

None