tests: Removed mxnet from functional tests + minor fix on rewards (#362)

* ci: change workflow

* changed timeout

* fix function reach reward

* print logs

* removing mxnet

* res'

.circleci/config.yml

@@ -731,7 +731,6 @@ workflows:
       - functional_tests:
           requires:
             - build_base
-            - integration_tests
       - functional_test_doom:
           requires:
             - build_doom_env

rl_coach/tests/test_checkpoint.py

@@ -20,12 +20,12 @@ import time
 import pytest
 import signal
 import tempfile
-import numpy as np
 import pandas as pd
 import rl_coach.tests.utils.args_utils as a_utils
 import rl_coach.tests.utils.test_utils as test_utils
 import rl_coach.tests.utils.presets_utils as p_utils
 from rl_coach import checkpoint
+from rl_coach.logger import screen
 from rl_coach.tests.utils.definitions import Definitions as Def
 
 
@@ -55,16 +55,14 @@ def test_get_checkpoint_state():
 
 
 @pytest.mark.functional_test
-@pytest.mark.parametrize("framework", ["mxnet", "tensorflow"])
+@pytest.mark.parametrize("framework", ["tensorflow"])
 def test_restore_checkpoint(preset_args, clres, framework,
-                            start_time=time.time(),
                             timeout=Def.TimeOuts.test_time_limit):
     """
     Create checkpoints and restore them in second run.
     :param preset_args: all preset that can be tested for argument tests
     :param clres: logs and csv files
     :param framework: name of the test framework
-    :param start_time: test started time
     :param timeout: max time for test
     """
 
@@ -79,7 +77,7 @@ def test_restore_checkpoint(preset_args, clres, framework,
             'python3', 'rl_coach/coach.py',
             '-p', '{}'.format(preset_args),
             '-e', '{}'.format("ExpName_" + preset_args),
-            '--seed', '{}'.format(42),
+            '--seed', '{}'.format(4),
             '-f', '{}'.format(framework),
         ]
 
@@ -90,6 +88,8 @@ def test_restore_checkpoint(preset_args, clres, framework,
 
         return p
 
+    start_time=time.time()
+
     if framework == "mxnet":
         # update preset name - for mxnet framework we are using *_DQN
         preset_args = Def.Presets.mxnet_args_test[0]
@@ -113,9 +113,12 @@ def test_restore_checkpoint(preset_args, clres, framework,
     if os.path.exists(checkpoint_test_dir):
         shutil.rmtree(checkpoint_test_dir)
 
-    assert a_utils.is_reward_reached(csv_path=csv_list[0],
-                                     p_valid_params=p_valid_params,
-                                     start_time=start_time, time_limit=timeout)
+    res = a_utils.is_reward_reached(csv_path=csv_list[0],
+                                    p_valid_params=p_valid_params,
+                                    start_time=start_time, time_limit=timeout)
+    if not res:
+        screen.error(open(clres.stdout.name).read(), crash=False)
+        assert False
 
     entities = a_utils.get_files_from_dir(checkpoint_dir)
 

rl_coach/tests/utils/args_utils.py

@@ -117,32 +117,33 @@ def is_reward_reached(csv_path, p_valid_params, start_time, time_limit):
     last_num_episodes = 0
     csv = None
     reward_reached = False
+    reward_str = 'Evaluation Reward'
 
-    while csv is None or (csv['Episode #'].values[-1]
-           < p_valid_params.max_episodes_to_achieve_reward and
-           time.time() - start_time < time_limit):
-
-        csv = pd.read_csv(csv_path)
-
-        if 'Evaluation Reward' not in csv.keys():
+    while csv is None or (csv[csv.columns[0]].values[
+                              -1] < p_valid_params.max_episodes_to_achieve_reward and time.time() - start_time < time_limit):
+        try:
+            csv = pd.read_csv(csv_path)
+        except:
+            # sometimes the csv is being written at the same time we are
+            # trying to read it. no problem -> try again
             continue
 
-        rewards = csv['Evaluation Reward'].values
+        if reward_str not in csv.keys():
+            continue
 
+        rewards = csv[reward_str].values
         rewards = rewards[~np.isnan(rewards)]
-        if len(rewards) >= 1:
-            averaged_rewards = np.convolve(rewards, np.ones(
-                min(len(rewards), win_size)) / win_size, mode='valid')
 
+        if len(rewards) >= 1:
+            averaged_rewards = np.convolve(rewards, np.ones(min(len(rewards), win_size)) / win_size, mode='valid')
         else:
-            # May be in heat-up steps
             time.sleep(1)
             continue
 
-        if csv['Episode #'].shape[0] - last_num_episodes <= 0:
+        if csv[csv.columns[0]].shape[0] - last_num_episodes <= 0:
             continue
 
-        last_num_episodes = csv['Episode #'].values[-1]
+        last_num_episodes = csv[csv.columns[0]].values[-1]
 
         # check if reward is enough
         if np.any(averaged_rewards >= p_valid_params.min_reward_threshold):
@@ -408,6 +409,7 @@ def validate_arg_result(flag, p_valid_params, clres=None, process=None,
         csv_path = get_csv_path(clres=clres)
         assert len(csv_path) > 0, \
             Def.Consts.ASSERT_MSG.format("path not found", csv_path)
+        time.sleep(5)
 
         get_reward = is_reward_reached(csv_path=csv_path[0],
                                        p_valid_params=p_valid_params,

rl_coach/tests/utils/definitions.py

@@ -139,7 +139,7 @@ class Definitions:
         N_csv_lines = 100  # number of lines to validate on csv file
 
     class TimeOuts:
-        test_time_limit = 60 * 60
+        test_time_limit = 30 * 60
         wait_for_files = 20
         wait_for_csv = 240
         test_run = 60