Move pagination logic to database level in get_assertions_by_badge

[Daniel-1600]

The get_assertions_by_badge endpoint in assertions.py currently fetches all assertions
from the database and then slices the list in Python to achieve pagination. This is inefficient
as it loads unnecessary data into memory.

Now that tahrir-api supports limit and offset parameters in get_assertions_by_badge
(see fedora-infra/tahrir-api#265), we can pass pagination parameters directly to the database
query instead.

Depends on: fedora-infra/tahrir-api#265