Since the prior code for v3 uses list comprehension to build the tracks returned, we can check if we're using v4 and if the data[data_type] is a dictionary, and wrap it in a list to ensure the same behavior.