Test data updates

[rpkilby]

The test suite should rely on endpoints/fantasy-database for the models & fixtures used in testing. If the canonical JSON-API test suite will be based on the endpoints/endpoints test suite, then using the fantasy-database will make integration of that test suite just a little bit easier in the long run.

• Wait for Python/Django implementation endpoints/fantasy-database#1 and package relase to PyPI
• Add fantasy-database to tests_require
• Strip out old models where necessary, and replace with fantasy book models (Maybe? Are there cases that the fantasy models don't cover?)