get_dataset_collection should return an iteratable

[koenvo]

When having lots of datasets (like 100.000+) the get_dataset_collection shouldn't return all at once. This will give memory issues. An approach with a cursor should work.

TODO:

• make sure we have a incremental id (like uuid7) so we can sort on dataset_id to maintain order while slicing
• Add parameter to get_dataset_collection to use pagination

[koenvo]

Sorted on created_at.

Solved in #37.

Also added iter_dataset_collections