class APITest(FlaskTestCase, FickleTestCase):
def create_app(self):
self.backend = Backend()
app = API(__name__, backend=self.backend)
app.config['TESTING'] = True
return app
def get(self, *args, **kwargs):
return self.client.get(*args, **kwargs)
def post(self, *args, **kwargs):
kwargs['content_type'] = 'application/json'
if kwargs.get('data'):
kwargs['data'] = json.dumps(kwargs['data'])
return self.client.post(*args, **kwargs)
def load(self, dataset):
return self.post('/load', data={
'data': dataset.data.tolist(),
'target': dataset.target.tolist()
})
def assert_success(self, response, status=200):
self.assertEqual(response.status_code, status)
def assert_error(self, response, status=400):
self.assertEqual(response.status_code, status)
def test_root(self):
response = self.get('/')
self.assert_success(response)
def test_load(self):
dataset = datasets.load_iris()
response = self.load(dataset)
self.assert_success(response, status=201)
self.assertTrue(self.backend.isloaded())
def test_fit_when_not_loaded(self):
response = self.post('/fit')
self.assert_error(response, status=501)
def test_fit_when_loaded(self):
dataset = datasets.load_iris()
self.load(dataset)
response = self.post('/fit')
self.assert_success(response)
self.assertTrue(self.backend.istrained())
def test_predict_when_trained(self):
dataset = datasets.load_iris()
self.load(dataset)
self.post('/fit')
sample = dataset['data'][:10]
response = self.post('/predict', data=sample.tolist())
self.assert200(response)
self.assertEqual(len(response.json), 10)
self.assertEqual(response.json, self.backend.predict(sample))
def test_istrained_without_load(self):
backend = Backend()
dataset = datasets.load_iris()
self.assertFalse(backend.istrained())
backend.load(dataset)
self.assertFalse(backend.istrained())
backend.fit()
self.assertTrue(backend.istrained())
def test_isloaded(self):
backend = Backend()
dataset = datasets.load_iris()
self.assertFalse(backend.isloaded())
backend.load(dataset)
self.assertTrue(backend.isloaded())
backend.load(dataset)
self.assertTrue(backend.isloaded())
def test_istrained_with_load(self):
old_backend = Backend()
dataset = datasets.load_iris()
old_backend.load(dataset)
old_backend.fit()
new_backend = Backend()
self.assertTrue(new_backend.istrained(read=True))
def test_predict_when_trained(self):
backend = Backend()
dataset = datasets.load_iris()
backend.load(dataset)
backend.fit()
sample = dataset["data"][:10]
predictions = backend.predict(sample)
self.assertEqual(len(predictions), 10)
def test_load(self):
backend = Backend()
dataset = datasets.load_iris()
self.assertTrue(backend.load(dataset))
def test_fit_when_loaded(self):
backend = Backend()
dataset = datasets.load_iris()
backend.load(dataset)
self.assertTrue(backend.fit())
def test_fit_when_not_loaded(self):
backend = Backend()
with self.assertRaises(RuntimeError):
backend.fit()
def create_app(self):
self.backend = Backend()
app = API(__name__, backend=self.backend)
app.config['TESTING'] = True
return app
