added assertion to runner

ssm-lab · Jun 18, 2024 · 16867a0 · 16867a0
1 parent d9da52f
commit 16867a0
Show file tree

Hide file tree

Showing 3 changed files with 12 additions and 5 deletions.
diff --git a/input/opinions-test.txt b/input/opinions-test.txt
@@ -1,2 +1,2 @@
 3
-[1,1], -2
+[1,1], 2
diff --git a/src/runner.py b/src/runner.py
@@ -171,8 +171,12 @@ def discrete_policy_grad(self, max_episodes, advice=None, is_random=False):
             logging.debug('Generating default policy')
             policy = self.get_default_policy(environment)
             if advice:
+                original_policy = policy
                 logging.info(f'\t\t\t Shaping policy with human input at u={advice.u}')
                 policy = self.shape_policy(policy, advice)
+                if advice.u==1.0:
+                    assert np.array_equal(original_policy, policy)
+                    print(np.array_equal(original_policy, policy))
 
             #logging.debug('Initial policy:')
             #logging.debug(policy)

diff --git a/tests/runner_tests.py b/tests/runner_tests.py
@@ -10,24 +10,27 @@ class RunnerTests(unittest.TestCase):
     def testShaping(self):
         #3x3 grid
 
-        default_policy = default_policy = np.full((9, 4), 0.25)
-        print(default_policy)
+        policy = np.full((9, 4), 0.25)
+        p1 = policy
+
+        print(p1)
 
         #advice about the middle cell
         file = os.path.abspath(f'input/opinions-test.txt')
         opinion_parser = OpinionParser()
 
         human_input = opinion_parser.parse(file)
         #advice = Advice(human_input, 0)
-        advice = Advice(human_input, 1)
+        advice = Advice(human_input, 0.99)
 
         print(f'{advice} @u={advice.u}.')
 
         r = Runner(12, 63, 2, 10)
-        policy = r.shape_policy(default_policy, advice)
+        policy = r.shape_policy(policy, advice)
 
         print(policy)
 
+        print(np.array_equal(p1, policy))
 
 if __name__ == "__main__":
     unittest.main()