Schema mismatch between 'regular' and querying endpoints

[jl-wynen]

When posting to and getting from datasets, we can use the *ObsoleteDTO models from the OpenAPI schema. But datasets/fullquery uses the database schema for datasets. This means that with the ongoing migration, we cannot assign the result of a query to the same model as the result of GET datasets.

For example, the output DTO contains proposalId (singular) but the schema contains proposalIds (plural). So if we want to query by proposal, we need to pass proposalIds = ["..."] and get back an object that contains a field proposalIds. This cannot be assigned to a model generated from the DTO. Or with the changes proposed in SciCatProject/scitacean#265, it can be assigned but leads to the returned model to have proposalId = None.

Does it make sense to change datasets/fullquery to use the same DTOs as datasets?