This PR addresses an N+1 query performance issue when fetching event lists via the GraphQL API, specifically impacting pEvent related fields such as organisation, nextOccurrenceDatetime, lastOccurrenceDatetime, and hasSpaceForEnrolments.

Problem:
Previously, when querying a list of events, each event's pEvent resolver would trigger individual database queries for its associated PalvelutarjotinEvent and its related organisation and occurrence data. This resulted in a high number of database queries (N+1 pattern), significantly degrading performance for event list requests.

Solution:

1. Bulk Fetch organisation: Added .select_related("organisation") to the PalvelutarjotinEvent queryset in graphene_linked_events/schema.py to fetch related organisation data in a single JOIN.
2. Optimize Occurrence Resolvers: Modified resolve_next_occurrence_datetime and resolve_last_occurrence_datetime in occurrences/schema.py to perform filtering on already prefetched occurrences in Python, rather than issuing new database queries for each event.
3. Optimize has_space_for_enrolments: Updated the PalvelutarjotinEvent.has_space_for_enrolments() method in occurrences/models.py to also utilize Python-side filtering on prefetched occurrences.

These changes ensure that related data is fetched efficiently, drastically reducing the number of database queries for event list operations.

Fixes KULTUS-API-5