#5 unit test rendering + clean-up random test

tests/flexible/test_fexible_random.py

@@ -8,18 +8,19 @@ def test_random(self):
         env = gym.make('JSSEnv:flexible-jss-v1', env_config={'instance_path': '../../JSSEnv/envs/instances/flexible/mt10c1.fjs'})
 
         average = 0
-        for _ in range(100):
-            state = env.reset()
-            self.assertEqual(env.current_time_step, 0)
-            self.assertNotEqual(env.todo_time_step_job.all(), env.nb_op_job.all())
-            self.assertEqual(env.todo_time_step_job.all(), 0)
+        for _ in range(1):
+            env.reset()
+            self.assertEqual(env.current_time_step, 0, msg="We start at timestep 0")
+            self.assertNotEqual(env.todo_time_step_job.all(), env.nb_op_job.all(), msg="We don't modify nb op job")
+            self.assertEqual(env.todo_time_step_job.all(), 0, msg="We reset todo step job at each reset")
             done = False
             while not done:
                 legal_actions = env.get_legal_actions()
                 actions = np.random.choice(len(legal_actions), 1, p=(legal_actions / legal_actions.sum()))[0]
                 state, reward, done, _ = env.step(actions)
             print(env.current_time_step)
-            self.assertEqual(env.todo_time_step_job.all(), env.nb_op_job.all())
+            self.assertEqual(env.todo_time_step_job.all(), env.nb_op_job.all(), msg="We need to perform all job's op")
+            env.render()
         '''
         average = 0
         for _ in range(100):

tests/flexible/test_flexible_rendering_random.py

@@ -0,0 +1,25 @@
+import unittest
+
+import gym
+import imageio
+import numpy as np
+
+
+class TestRendering(unittest.TestCase):
+
+    def test_random_mt10c1_rendering(self):
+        env = gym.make('JSSEnv:flexible-jss-v1', env_config={'instance_path': '../../JSSEnv/envs/instances/flexible/mt10c1.fjs'})
+        env.reset()
+        self.assertEqual(env.current_time_step, 0)
+        done = False
+        step_nb = 0
+        images = []
+        while not done:
+            legal_actions = env.get_legal_actions()
+            actions = np.random.choice(len(legal_actions), 1, p=(legal_actions / legal_actions.sum()))[0]
+            state, reward, done, _ = env.step(actions)
+            temp_image = env.render().to_image()
+            images.append(imageio.imread(temp_image))
+        imageio.mimsave("mt10c1.gif", images)
+        env.reset()
+        self.assertEqual(env.current_time_step, 0)