WIP: Unittests

blobquiet · Mar 25, 2022 · fa1ccbe · fa1ccbe
1 parent 9a91807
commit fa1ccbe
Show file tree

Hide file tree

Showing 4 changed files with 21 additions and 13 deletions.
diff --git a/fedbiomed/common/training_plans/_fedbiosklearn.py b/fedbiomed/common/training_plans/_fedbiosklearn.py
@@ -100,6 +100,8 @@ def __init__(self, model_args: dict = {}):
         if not isinstance(model_args, dict):
             model_args = {}
 
+        print('###', model_args)
+
         if 'model' not in model_args:
             msg = ErrorNumbers.FB303.value + ": SKLEARN model not provided"
             logger.critical(msg)

diff --git a/tests/test_fedbiosklearn.py b/tests/test_fedbiosklearn.py
@@ -6,55 +6,59 @@
 from sklearn.linear_model import SGDRegressor
 
 from fedbiomed.common.training_plans import SGDSkLearnModel
-from fedbiomed.common.exceptions     import FedbiomedTrainingPlanError
+from fedbiomed.common.exceptions import FedbiomedTrainingPlanError
 
 
 class TestModel(SGDSkLearnModel):
     """
     What is it ?
     """
+
+    def __init__(self, model_args: dict = {}):
+        super(TestModel).__init__()
+
     def adhoc(self):
         print('adhoc')
 
+
 class TestFedbiosklearn(unittest.TestCase):
 
     def setUp(self):
         self.model = SGDRegressor(max_iter=1000, tol=1e-3)
 
-
     def tearDown(self):
         pass
 
     def test_init(self):
-        kw = {'toto': 'le' , 'lelo':'la', 'max_iter':7000, 'tol': 0.3456, 'n_features': 10, 'model': 'SGDRegressor' }
+        kw = {'toto': 'le', 'lelo': 'la', 'max_iter': 7000, 'tol': 0.3456, 'n_features': 10, 'model': 'SGDRegressor'}
         fbsk = SGDSkLearnModel(kw)
         m = fbsk.get_model()
         p = m.get_params()
-        self.assertEqual(p['max_iter'] , 7000)
-        self.assertEqual(p['tol'],  0.3456 )
-        self.assertTrue( np.allclose(m.coef_, np.zeros(10)) )
+        self.assertEqual(p['max_iter'], 7000)
+        self.assertEqual(p['tol'], 0.3456)
+        self.assertTrue(np.allclose(m.coef_, np.zeros(10)))
         self.assertIsNone(p.get('lelo'))
         self.assertIsNone(p.get('toto'))
         self.assertIsNone(p.get('model'))
 
     def test_not_implemented_method(self):
         kw = {'toto': 'le', 'lelo': 'la', 'max_iter': 7000, 'tol': 0.3456, 'n_features': 10, 'model': 'SGDRegressor'}
         t = TestModel(kw)
-        self.assertRaises(FedbiomedTrainingPlanError,lambda: t.training_data())
+        self.assertRaises(FedbiomedTrainingPlanError, lambda: t.training_data())
 
     def test_save_and_load(self):
         randomfile = tempfile.NamedTemporaryFile()
 
-        skm = SGDSkLearnModel({'max_iter': 1000, 'tol':1e-3, 'n_features': 5, 'model': 'SGDRegressor'})
+        skm = SGDSkLearnModel({'max_iter': 1000, 'tol': 1e-3, 'n_features': 5, 'model': 'SGDRegressor'})
         skm.save(randomfile.name)
 
-        self.assertTrue(os.path.exists(randomfile.name) and os.path.getsize(randomfile.name) > 0  )
+        self.assertTrue(os.path.exists(randomfile.name) and os.path.getsize(randomfile.name) > 0)
 
         m = skm.load(randomfile.name)
 
-        self.assertEqual(m.max_iter,1000)
+        self.assertEqual(m.max_iter, 1000)
         self.assertEqual(m.tol, 0.001)
 
 
 if __name__ == '__main__':  # pragma: no cover
-    unittest.main()
+    unittest.main()
diff --git a/tests/test_json.py b/tests/test_json.py
@@ -27,8 +27,6 @@ def test_basic_json(self):
         loop1 = js.serialize_msg(js.deserialize_msg(msg))
         self.assertEqual(loop1, msg)
 
-        loop2 = js.deserialize_msg(js.serialize_msg(msg))
-        self.assertEqual(loop1, msg)
         pass
 
     def test_errnum_json(self):

diff --git a/tests/testsupport/fake_training_plan.py b/tests/testsupport/fake_training_plan.py
@@ -86,6 +86,10 @@ def training_routine(self, **kwargs):
         """
         time.sleep(FakeModel.SLEEPING_TIME)
 
+    def testing_routine(self, **kwargs):
+        pass
+
+
     def after_training_params(self) -> List[int]:
         """Fakes `after_training_params` method of TrainingPlan classes.
         Originally used to get the parameters after training is performed.