Cloudpickle for experiment API method_call (#821)

- use cloudpickle for method_call dump/loads
- add integration test for examples/

setup.py

@@ -18,6 +18,7 @@
     'akro==0.0.6',
     'cached_property',
     'click',
+    'cloudpickle',
     'cma==1.1.06',
     # dm_control throws an error during install about not being able to
     # find a build dependency (absl-py). Later pip executes the `install`

src/garage/experiment/experiment.py

@@ -10,6 +10,7 @@
 import subprocess
 import sys
 
+import cloudpickle
 import dateutil.tz
 import numpy as np
 
@@ -276,7 +277,7 @@ def run_experiment(method_call=None,
 
     for task in batch_tasks:
         call = task.pop('method_call')
-        data = base64.b64encode(pickle.dumps(call)).decode('utf-8')
+        data = base64.b64encode(cloudpickle.dumps(call)).decode('utf-8')
         task['args_data'] = data
         exp_count += 1
 

src/garage/experiment/experiment_wrapper.py

@@ -12,6 +12,7 @@
 import sys
 import uuid
 
+import cloudpickle
 import dateutil.tz
 import dowel
 from dowel import logger
@@ -182,7 +183,7 @@ def run_experiment(argv):
         snapshot_mode=args.snapshot_mode,
         snapshot_gap=args.snapshot_gap)
 
-    method_call = pickle.loads(base64.b64decode(args.args_data))
+    method_call = cloudpickle.loads(base64.b64decode(args.args_data))
     try:
         method_call(snapshot_config, variant_data, args.resume_from_dir,
                     args.resume_from_epoch)

tests/integration_tests/test_examples.py

@@ -0,0 +1,38 @@
+"""
+This is an integration test to make sure scripts from examples/
+work when running `python examples/xx/xxx.py`.
+"""
+from garage.experiment import LocalRunner, run_experiment
+from garage.np.baselines import LinearFeatureBaseline
+from garage.tf.algos import VPG
+from garage.tf.envs import TfEnv
+from garage.tf.policies import CategoricalMLPPolicy
+
+
+def _run_task(snapshot_config, *_):
+    with LocalRunner(snapshot_config=snapshot_config) as runner:
+        env = TfEnv(env_name='CartPole-v1')
+
+        policy = CategoricalMLPPolicy(
+            name='policy', env_spec=env.spec, hidden_sizes=(32, 32))
+
+        baseline = LinearFeatureBaseline(env_spec=env.spec)
+
+        algo = VPG(
+            env_spec=env.spec,
+            policy=policy,
+            baseline=baseline,
+            max_path_length=100,
+            discount=0.99,
+            optimizer_args=dict(tf_optimizer_args=dict(learning_rate=0.01, )))
+
+        runner.setup(algo, env)
+        runner.train(n_epochs=3, batch_size=100)
+
+
+if __name__ == '__main__':
+    run_experiment(
+        _run_task,
+        snapshot_mode='last',
+        seed=1,
+    )