unit tests

[BurntSushi]

This must be done before porting to Python 3. #78

Strategy to come shortly.

[leighlondon]

What's the strategy you're thinking of for these tests?

[BurntSushi]

@leighlondon I've start some testing on nfldb here: https://github.com/BurntSushi/nfldb/tree/master/tests

For the most part, I want to test the API to make sure it is giving the results I expect it to give. There should be similar tests for nflgame.

(I don't consider unit tests as something which should test the validity of the data. That's another can of worms.)

[leighlondon]

I'd be interested in adding some tests to this if it'd help out.

# tests/test_game.py
from nflgame.game import FieldPosition, PossessionTime

class TestPossessionTime:
    def test_minutes_to_total_seconds(self):
        clock = PossessionTime("5:00")
        assert clock.total_seconds() == 300
    # ...

class TestFieldPosition:
    def test_offset_redzone_yards_correct(self):
        fp = FieldPosition(yardline="NYJ 2", pos_team="NE")
        assert fp.offset == 48
    # ...

I take it you mean something like that for the API tests, so that it's data-agnostic but testing the base classes and functions?

[BurntSushi]

Yes, that looks reasonable. You can certainly rely on expecting certain data values for your tests.

Although, please use py.test. (It's really easy to use. Check out nfldb's tests.)

Thanks. :D