[Proposal] Unique species tests #7
Description
Problem
Ordinary tests are based on iNaturalist's observations API, and use random sorting to produce unique experiences for each quiz. In some respects, this is useful because it mirrors real-life expectations. The downside is that some species appear with high relative frequency, and if someone is aiming to learn all species, this is not always ideal.
Proposal
A new type of test that fetches any species that are present (under the appropriate filters) equally often. This probably requires a different API strategy. First, it should use the species_counts API to fetch all species observed under the given filters (location, taxa). Then it should pick randomly from this list and fetch a single observation per test question.
Notes
The species_counts API probably supports more than just species. It could probably be used to fetch genuses, families, etc too and thus support genus or family name tests