def test_predict_nomodel(self, mock_modelstatus, mock_conn):
mock_conn.get.return_value = None
mock_modelstatus.return_value = "Model Doesn't Exist Flask response"
returned = predict('model_doesntexist')
self.assertEqual(mock_conn.get.called, True)
self.assertEqual(returned, "Model Doesn't Exist Flask response")
def test_predict_returnspred(self, mock_preprocess, mock_bytesdf,
mock_getreqdata, mock_pickle, mock_conn):
mock_conn.get.return_value = 'good_model'
mock_getreqdata.return_value = b'good data'
mock_bytesdf.return_value = 'good_data_df'
mock_preprocess.return_value = 'processed_data_df'
mock_pickle.loads.return_value.predict.return_value.tolist.return_value = 'class predictions list'
mock_pickle.loads.return_value.predict_proba.return_value.tolist.return_value = 'class probability predictions list'
predict('good_model')
mock_pickle.loads.assert_called_with('good_model')
self.assertTrue(mock_getreqdata.called, True)
mock_bytesdf.assert_called_with(b'good data')
mock_preprocess.assert_called_with('good_data_df')
mock_pickle.loads.return_value.predict.assert_called_with(
'processed_data_df')
mock_pickle.loads.return_value.predict_proba.assert_called_with(
'processed_data_df')
def getisbndetails(isbn):
api = 'https://www.googleapis.com/books/v1/volumes?q=isbn:'
isbnapi = api + isbn
r = requests.get(isbnapi)
data = r.json()
bookdetails = {}
bookdetails['title'] = data['items'][0]['volumeInfo']['title']
bookdetails['authors'] = data['items'][0]['volumeInfo']['authors']
bookdetails['pageCount'] = data['items'][0]['volumeInfo']['pageCount']
bookdetails['genre'] = predict(data['items'][0]['volumeInfo']['title'])
bookdetails['isbn'] = isbn
bookdetails['image'] = data['items'][0]['volumeInfo']['imageLinks'][
'thumbnail']
language = data['items'][0]['volumeInfo']['language']
if language.lower() == 'en':
bookdetails['language'] = 'English'
else:
bookdetails['language'] = language
return bookdetails
def prediction(message):
result = app.predict(message)
return result
def test_show(s):
run_error(s, 'unknown contract', app.show, 'test-contract1')
run(s, app.create, 'test-contract1', 'terms', '1 hour', '.5')
out = run(s, app.show, 'test-contract1')
assert 'terms (Unresolved)' in out
assert 'Closes' in out
user1 = app.lookup_or_create_user(s, 'user1')
user2 = app.lookup_or_create_user(s, 'user2')
app.predict(s, user1, 'test-contract1', '.6')
app.predict(s, user2, 'test-contract1', '.8')
app.predict(s, user1, 'test-contract1', '.4')
app.predict(s, user1, 'test-contract1', '.2')
app.predict(s, user2, 'test-contract1', '.1')
app.predict(s, user1, 'test-contract1', '.0001')
run(s, app.resolve, 'test-contract1', 'false')
out = run(s, app.show, 'test-contract1')
assert 'terms (Resolved False)' in out
assert 'Was to close' in out
assert '50.00% test (' in out
assert '60.00% user1 (' in out
assert '80.00% user2 (' in out
assert '40.00% user1 (' in out
assert '20.00% user1 (' in out
assert '10.00% user2 (' in out
assert '0.01% user1 (' in out
assert 'scores:' in out
assert 'user1: 126.84' in out
assert 'user2: -57.54' in out
run(s, app.cancel, 'test-contract1')
out = run(s, app.show, 'test-contract1')
assert 'scores' not in out
assert 'Was to close' in out
assert 'Cancelled' in out
# User 'test' doesn't get lots of points for setting the house odds at 1%
# and then immediately predicting 99%.
run(s, app.create, 'test-contract2', 'terms', '1 hour', '.01')
run(s, app.predict, 'test-contract2', '.99')
app.predict(s, user1, 'test-contract2', '.9')
run(s, app.predict, 'test-contract2', '.99')
run(s, app.resolve, 'test-contract2', 'true')
out = run(s, app.show, 'test-contract2')
assert 'test: 9.53' in out
assert 'user1: -9.53' in out
def get_prediction():
data = request.get_json()
output = predict(convert_to_df(dict(data)))
with open('output.csv', 'a+') as f:
f.write(','.join(str(v) for v in output[0]) + "\n")
return str(','.join(str(v) for v in output[0]))
def test_predict(self):
self.assertIsInstance(app_.predict(self._test_object()), list)