Add tests for expected model results

[pleary]

• These would be different from the automated tests that Github would run
• These tests would use real models and associated files and real sample images and test that the results align with expected results
• The tests would evaluate all major API parameters and considerations (for example iconic taxon filtering, input lat/lng) and that results are modified accordingly
• These tests could be replicated on other implementations of similar model-based APIs to confirm they are implementing the same features and get the same or similar results