Could we have this be the actual schema we use in the search endpoint? Just to make sure it's valid.

Also want to make sure the JSON serialization works with a full search request object for things like enums and dates.

I added something to the test using get_search_request(). Something like this you have in mind?

I was thinking we would do search_query = fields.Nested(OpportunitySearchRequestV1Schema) - that way we can only take in a schema that is valid for search. Anything that isn't a valid search request should be rejected (eg. a user couldn't just pass {"some_field": "whatever"}.)

I checked in the changes. This makes me wonder if we should also store the version of the search request schema for validation later. If, for example, we want the response to correspond to the schema that was inputted. It would be able to handle schema changes (and we could just output a dict response type validated against the stored schema version in a new db column. Otherwise we would not be able to validate response schema if this OpportunitySearchRequestV1Schema changes.

I was thinking we would do search_query = fields.Nested(OpportunitySearchRequestV1Schema) - that way we can only take in a schema that is valid for search. Anything that isn't a valid search request should be rejected (eg. a user couldn't just pass {"some_field": "whatever"}.)

We can exclude the begin here since we already began it in the route. Actually, in a lot of cases this would error with an "already started transaction error", guessing that didn't happen because you hadn't done any DB operations yet.