Merge pull request #4385 from Unity-Technologies/release_2_verified-barracuda-1.0.2

update verified brach with barracuda 1.0.2

.pre-commit-config.yaml

         # flake8-tidy-imports is used for banned-modules, not actually tidying
         additional_dependencies: [flake8-comprehensions==3.2.2, flake8-tidy-imports==4.1.0, flake8-bugbear==20.1.4]
 
+-   repo: https://github.com/asottile/pyupgrade
+    rev: v2.7.0
+    hooks:
+    -   id: pyupgrade
+        args: [--py3-plus, --py36-plus]
+        exclude: >
+            (?x)^(
+                .*barracuda.py|
+                .*_pb2.py|
+                .*_pb2_grpc.py
+            )$
+
 -   repo: https://github.com/pre-commit/pre-commit-hooks
     rev: v2.5.0
     hooks:

.yamato/com.unity.ml-agents-test.yml

     image: {{ platform.image }}
     flavor: {{ platform.flavor}}
   commands:
-    - python -m pip install unity-downloader-cli --extra-index-url https://artifactory.eu-cph-1.unityops.net/api/pypi/common-python/simple
+    - python -m pip install unity-downloader-cli --index-url https://artifactory.prd.it.unity3d.com/artifactory/api/pypi/unity-pypi-local/simple --upgrade
     - unity-downloader-cli -u trunk -c editor --wait --fast
     - npm install upm-ci-utils@stable -g --registry https://artifactory.prd.cds.internal.unity3d.com/artifactory/api/npm/upm-npm
     - upm-ci package test -u {{ editor.version }} --package-path com.unity.ml-agents {{ editor.coverageOptions }}

.yamato/gym-interface-test.yml

   variables:
     UNITY_VERSION: {{ editor.version }}
   commands:
-    - pip install pyyaml
+    - pip install pyyaml --index-url https://artifactory.prd.it.unity3d.com/artifactory/api/pypi/pypi/simple
     - python -u -m ml-agents.tests.yamato.setup_venv
     - ./venv/bin/python ml-agents/tests/yamato/scripts/run_gym.py --env=artifacts/testPlayer-Basic
   dependencies:
-    changes:
-      only:
-        - "com.unity.ml-agents/**"
-        - "Project/**"
-        - "ml-agents/**"
-        - "ml-agents-envs/**"
-        - ".yamato/gym-interface-test.yml"
-      except:
-        - "*.md"
-        - "com.unity.ml-agents/*.md"
-        - "com.unity.ml-agents/**/*.md"
+    expression: |
+      (pull_request.target eq "master" OR
+      pull_request.target match "release.+") AND
+      NOT pull_request.draft AND
+      (pull_request.changes.any match "com.unity.ml-agents/**" OR
+      pull_request.changes.any match "Project/**" OR
+      pull_request.changes.any match "ml-agents/**" OR
+      pull_request.changes.any match "ml-agents-envs/**" OR
+      pull_request.changes.any match "gym-unity/**" OR
+      pull_request.changes.any match ".yamato/gym-interface-test.yml") AND
+      NOT pull_request.changes.all match "**/*.md"
 {% endfor %}

.yamato/protobuf-generation-test.yml

       nuget install Grpc.Tools -Version $GRPC_VERSION -OutputDirectory protobuf-definitions/
       python3 -m venv venv
       . venv/bin/activate
-      pip install --upgrade pip
-      pip install grpcio-tools==1.13.0  --progress-bar=off
-      pip install mypy-protobuf==1.16.0 --progress-bar=off
+      pip install --upgrade pip --index-url https://artifactory.prd.it.unity3d.com/artifactory/api/pypi/pypi/simple
+      pip install grpcio==1.28.1 grpcio-tools==1.13.0 protobuf==3.11.3 six==1.14.0 mypy-protobuf==1.16.0  --progress-bar=off --index-url https://artifactory.prd.it.unity3d.com/artifactory/api/pypi/pypi/simple
       cd protobuf-definitions
       chmod +x Grpc.Tools.$GRPC_VERSION/tools/macosx_x64/protoc
       chmod +x Grpc.Tools.$GRPC_VERSION/tools/macosx_x64/grpc_csharp_plugin

.yamato/python-ll-api-test.yml

   variables:
     UNITY_VERSION: {{ editor.version }}
   commands:
-    - pip install pyyaml
+    - pip install pyyaml --index-url https://artifactory.prd.it.unity3d.com/artifactory/api/pypi/pypi/simple
-    - ./venv/bin/python ml-agents/tests/yamato/scripts/run_llapi.py 
+    - ./venv/bin/python ml-agents/tests/yamato/scripts/run_llapi.py
     - ./venv/bin/python ml-agents/tests/yamato/scripts/run_llapi.py --env=artifacts/testPlayer-Basic
     - ./venv/bin/python ml-agents/tests/yamato/scripts/run_llapi.py --env=artifacts/testPlayer-WallJump
     - ./venv/bin/python ml-agents/tests/yamato/scripts/run_llapi.py --env=artifacts/testPlayer-Bouncer
     cancel_old_ci: true
-    changes:
-      only:
-        - "com.unity.ml-agents/**"
-        - "Project/**"
-        - "ml-agents/**"
-        - "ml-agents-envs/**"
-        - ".yamato/python-ll-api-test.yml"
-      except:
-        - "*.md"
-        - "com.unity.ml-agents/*.md"
-        - "com.unity.ml-agents/**/*.md"
+    expression: |
+      (pull_request.target eq "master" OR
+      pull_request.target match "release.+") AND
+      NOT pull_request.draft AND
+      (pull_request.changes.any match "com.unity.ml-agents/**" OR
+      pull_request.changes.any match "Project/**" OR
+      pull_request.changes.any match "ml-agents/**" OR
+      pull_request.changes.any match "ml-agents-envs/**" OR
+      pull_request.changes.any match ".yamato/python-ll-api-test.yml") AND
+      NOT pull_request.changes.all match "**/*.md"
 {% endfor %}

.yamato/standalone-build-test.yml

   variables:
     UNITY_VERSION: {{ editor.version }}
   commands:
-    - pip install pyyaml
+    - pip install pyyaml --index-url https://artifactory.prd.it.unity3d.com/artifactory/api/pypi/pypi/simple
     - python -u -m ml-agents.tests.yamato.standalone_build_tests
     - python -u -m ml-agents.tests.yamato.standalone_build_tests --scene=Assets/ML-Agents/Examples/Basic/Scenes/Basic.unity
     - python -u -m ml-agents.tests.yamato.standalone_build_tests --scene=Assets/ML-Agents/Examples/Bouncer/Scenes/Bouncer.unity 

.yamato/training-int-tests.yml

   variables:
     UNITY_VERSION: {{ editor.version }}
   commands:
-    - pip install pyyaml
+    - pip install pyyaml --index-url https://artifactory.prd.it.unity3d.com/artifactory/api/pypi/pypi/simple
     - python -u -m ml-agents.tests.yamato.training_int_tests
     # Backwards-compatibility tests.
     # If we make a breaking change to the communication protocol, these will need

com.unity.ml-agents/CHANGELOG.md

 and this project adheres to
 [Semantic Versioning](http://semver.org/spec/v2.0.0.html).
 
+## [1.0.4] - 2020-08-19
+### Minor Changes
+#### com.unity.ml-agents (C#)
+- Update Barracuda to 1.0.2. (#4385)
+- Explicitly call out dependencies in package.json.
+### Bug Fixes
+#### com.unity.ml-agents (C#)
+#### ml-agents / ml-agents-envs / gym-unity (Python)
+
 ## [1.0.3] - 2020-07-07
 ### Minor Changes
 #### com.unity.ml-agents (C#)

com.unity.ml-agents/Runtime/Academy.cs

         /// Unity package version of com.unity.ml-agents.
         /// This must match the version string in package.json and is checked in a unit test.
         /// </summary>
-        internal const string k_PackageVersion = "1.0.3";
+        internal const string k_PackageVersion = "1.0.4";
 
         const int k_EditorTrainingPort = 5004;
 

com.unity.ml-agents/package.json

 {
   "name": "com.unity.ml-agents",
   "displayName": "ML Agents",
-  "version": "1.0.3",
+  "version": "1.0.4",
-    "com.unity.barracuda": "1.0.1"
+    "com.unity.barracuda": "1.0.2",
+    "com.unity.modules.imageconversion": "1.0.0",
+    "com.unity.modules.jsonserialize": "1.0.0",
+    "com.unity.modules.physics": "1.0.0",
+    "com.unity.modules.physics2d": "1.0.0"
   }
 }

gym-unity/setup.py

         tag = os.getenv("CIRCLE_TAG")
 
         if tag != EXPECTED_TAG:
-            info = "Git tag: {0} does not match the expected tag of this app: {1}".format(
+            info = "Git tag: {} does not match the expected tag of this app: {}".format(
                 tag, EXPECTED_TAG
             )
             sys.exit(info)
     author_email="ML-Agents@unity3d.com",
     url="https://github.com/Unity-Technologies/ml-agents",
     packages=find_packages(),
-    install_requires=["gym", "mlagents_envs=={}".format(VERSION)],
+    install_requires=["gym", f"mlagents_envs=={VERSION}"],
     cmdclass={"verify": VerifyVersionCommand},
 )

ml-agents-envs/mlagents_envs/base_env.py

         :returns: The DecisionStep
         """
         if agent_id not in self.agent_id_to_index:
-            raise KeyError(
-                "agent_id {} is not present in the DecisionSteps".format(agent_id)
-            )
+            raise KeyError(f"agent_id {agent_id} is not present in the DecisionSteps")
         agent_index = self._agent_id_to_index[agent_id]  # type: ignore
         agent_obs = []
         for batched_obs in self.obs:
         specific agent
         """
         if agent_id not in self.agent_id_to_index:
-            raise KeyError(
-                "agent_id {} is not present in the TerminalSteps".format(agent_id)
-            )
+            raise KeyError(f"agent_id {agent_id} is not present in the TerminalSteps")
         agent_index = self._agent_id_to_index[agent_id]  # type: ignore
         agent_obs = []
         for batched_obs in self.obs:

ml-agents-envs/mlagents_envs/communicator.py

 from mlagents_envs.communicator_objects.unity_input_pb2 import UnityInputProto
 
 
-class Communicator(object):
+class Communicator:
     def __init__(self, worker_id=0, base_port=5005):
         """
         Python side of the communication. Must be used in pair with the right Unity Communicator equivalent.

ml-agents-envs/mlagents_envs/environment.py

             for _sc in side_channels:
                 if _sc.channel_id in self.side_channels:
                     raise UnityEnvironmentException(
-                        "There cannot be two side channels with the same channel id {0}.".format(
+                        "There cannot be two side channels with the same channel id {}.".format(
                             _sc.channel_id
                         )
                     )
             .replace(".x86", "")
         )
         true_filename = os.path.basename(os.path.normpath(env_path))
-        logger.debug("The true file name is {}".format(true_filename))
+        logger.debug(f"The true file name is {true_filename}")
 
         if not (glob.glob(env_path) or glob.glob(env_path + ".*")):
             return None
                 f"Couldn't launch the {file_name} environment. Provided filename does not match any environments."
             )
         else:
-            logger.debug("This is the launch string {}".format(launch_string))
+            logger.debug(f"This is the launch string {launch_string}")
             # Launch Unity environment
             subprocess_args = [launch_string]
             if no_graphics:
     def _assert_behavior_exists(self, behavior_name: str) -> None:
         if behavior_name not in self._env_specs:
             raise UnityActionException(
-                "The group {0} does not correspond to an existing agent group "
+                "The group {} does not correspond to an existing agent group "
                 "in the environment".format(behavior_name)
             )
 
         expected_shape = (len(self._env_state[behavior_name][0]), spec.action_size)
         if action.shape != expected_shape:
             raise UnityActionException(
-                "The behavior {0} needs an input of dimension {1} for "
+                "The behavior {} needs an input of dimension {} for "
-                "dimension {2}".format(behavior_name, expected_shape, action.shape)
+                "dimension {}".format(behavior_name, expected_shape, action.shape)
             )
         if action.dtype != expected_type:
             action = action.astype(expected_type)
         expected_shape = (spec.action_size,)
         if action.shape != expected_shape:
             raise UnityActionException(
-                f"The Agent {0} with BehaviorName {1} needs an input of dimension "
-                f"{2} but received input of dimension {3}".format(
-                    agent_id, behavior_name, expected_shape, action.shape
-                )
+                f"The Agent {agent_id} with BehaviorName {behavior_name} needs an input of dimension "
+                f"{expected_shape} but received input of dimension {action.shape}"
             )
         expected_type = np.float32 if spec.is_action_continuous() else np.int32
         if action.dtype != expected_type:
                 )
             if len(message_data) != message_len:
                 raise UnityEnvironmentException(
-                    "The message received by the side channel {0} was "
+                    "The message received by the side channel {} was "
                     "unexpectedly short. Make sure your Unity Environment "
                     "sending side channel data properly.".format(channel_id)
                 )
             else:
                 logger.warning(
                     "Unknown side channel data received. Channel type "
-                    ": {0}.".format(channel_id)
+                    ": {}.".format(channel_id)
                 )
 
     @staticmethod

ml-agents-envs/mlagents_envs/exception.py

 
     def __init__(self, worker_id):
         message = self.MESSAGE_TEMPLATE.format(str(worker_id))
-        super(UnityWorkerInUseException, self).__init__(message)
+        super().__init__(message)

ml-agents-envs/mlagents_envs/rpc_communicator.py

         s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
         try:
             s.bind(("localhost", port))
-        except socket.error:
+        except OSError:
             raise UnityWorkerInUseException(self.worker_id)
         finally:
             s.close()

ml-agents-envs/mlagents_envs/side_channel/environment_parameters_channel.py

         FLOAT = 0
 
     def __init__(self) -> None:
-        channel_id = uuid.UUID(("534c891e-810f-11ea-a9d0-822485860400"))
+        channel_id = uuid.UUID("534c891e-810f-11ea-a9d0-822485860400")
         super().__init__(channel_id)
 
     def on_message_received(self, msg: IncomingMessage) -> None:

ml-agents-envs/mlagents_envs/side_channel/float_properties_channel.py

     def __init__(self, channel_id: uuid.UUID = None) -> None:
         self._float_properties: Dict[str, float] = {}
         if channel_id is None:
-            channel_id = uuid.UUID(("60ccf7d0-4f7e-11ea-b238-784f4387d1f7"))
+            channel_id = uuid.UUID("60ccf7d0-4f7e-11ea-b238-784f4387d1f7")
         super().__init__(channel_id)
 
     def on_message_received(self, msg: IncomingMessage) -> None:

ml-agents-envs/mlagents_envs/tests/test_side_channel.py

     sender = RawBytesChannel(guid)
     receiver = RawBytesChannel(guid)
 
-    sender.send_raw_data("foo".encode("ascii"))
-    sender.send_raw_data("bar".encode("ascii"))
+    sender.send_raw_data(b"foo")
+    sender.send_raw_data(b"bar")
 
     data = UnityEnvironment._generate_side_channel_data({sender.channel_id: sender})
     UnityEnvironment._parse_side_channel_message({receiver.channel_id: receiver}, data)

ml-agents-envs/setup.py

         tag = os.getenv("CIRCLE_TAG")
 
         if tag != EXPECTED_TAG:
-            info = "Git tag: {0} does not match the expected tag of this app: {1}".format(
+            info = "Git tag: {} does not match the expected tag of this app: {}".format(
                 tag, EXPECTED_TAG
             )
             sys.exit(info)

ml-agents/mlagents/trainers/buffer.py

         super().__init__()
 
     def __str__(self):
-        return ", ".join(["'{0}' : {1}".format(k, str(self[k])) for k in self.keys()])
+        return ", ".join(["'{}' : {}".format(k, str(self[k])) for k in self.keys()])
 
     def reset_agent(self) -> None:
         """
             key_list = list(self.keys())
         if not self.check_length(key_list):
             raise BufferException(
-                "The length of the fields {0} were not of same length".format(key_list)
+                f"The length of the fields {key_list} were not of same length"
             )
         for field_key in key_list:
             target_buffer[field_key].extend(

ml-agents/mlagents/trainers/components/bc/model.py

 from mlagents.trainers.policy.tf_policy import TFPolicy
 
 
-class BCModel(object):
+class BCModel:
     def __init__(
         self, policy: TFPolicy, learning_rate: float = 3e-4, anneal_steps: int = 0
     ):

ml-agents/mlagents/trainers/components/bc/module.py

         for k in param_keys:
             if k not in config_dict:
                 raise UnityTrainerException(
-                    "The required pre-training hyper-parameter {0} was not defined. Please check your \
+                    "The required pre-training hyper-parameter {} was not defined. Please check your \
                     trainer YAML file.".format(
                         k
                     )

ml-agents/mlagents/trainers/components/reward_signals/__init__.py

         for k in param_keys:
             if k not in config_dict:
                 raise UnityTrainerException(
-                    "The hyper-parameter {0} could not be found for {1}.".format(
+                    "The hyper-parameter {} could not be found for {}.".format(
                         k, cls.__name__
                     )
                 )

ml-agents/mlagents/trainers/components/reward_signals/curiosity/model.py

 from mlagents.trainers.policy.tf_policy import TFPolicy
 
 
-class CuriosityModel(object):
+class CuriosityModel:
     def __init__(
         self, policy: TFPolicy, encoding_size: int = 128, learning_rate: float = 3e-4
     ):
                     self.encoding_size,
                     ModelUtils.swish,
                     1,
-                    "curiosity_stream_{}_visual_obs_encoder".format(i),
+                    f"curiosity_stream_{i}_visual_obs_encoder",
                     False,
                 )
 
                     ModelUtils.swish,
                     1,
-                    "curiosity_stream_{}_visual_obs_encoder".format(i),
+                    f"curiosity_stream_{i}_visual_obs_encoder",
                     True,
                 )
                 visual_encoders.append(encoded_visual)

ml-agents/mlagents/trainers/components/reward_signals/gail/model.py

 EPSILON = 1e-7
 
 
-class GAILModel(object):
+class GAILModel:
     def __init__(
         self,
         policy: TFPolicy,
                     self.encoding_size,
                     ModelUtils.swish,
                     1,
-                    "gail_stream_{}_visual_obs_encoder".format(i),
+                    f"gail_stream_{i}_visual_obs_encoder",
                     False,
                 )
 
                     ModelUtils.swish,
                     1,
-                    "gail_stream_{}_visual_obs_encoder".format(i),
+                    f"gail_stream_{i}_visual_obs_encoder",
                     True,
                 )
                 visual_policy_encoders.append(encoded_policy_visual)

ml-agents/mlagents/trainers/components/reward_signals/reward_signal_factory.py

     """
     rcls = NAME_TO_CLASS.get(name)
     if not rcls:
-        raise UnityTrainerException("Unknown reward signal type {0}".format(name))
+        raise UnityTrainerException(f"Unknown reward signal type {name}")
-            "Unknown parameters given for reward signal {0}".format(name)
+            f"Unknown parameters given for reward signal {name}"
         )
     return class_inst

ml-agents/mlagents/trainers/curriculum.py

                 for key in parameters:
                     config[key] = parameters[key][self.lesson_num]
                 logger.info(
-                    "{0} lesson changed. Now in lesson {1}: {2}".format(
+                    "{} lesson changed. Now in lesson {}: {}".format(
                         self.brain_name,
                         self.lesson_num,
                         ", ".join([str(x) + " -> " + str(config[x]) for x in config]),
         try:
             with open(config_path) as data_file:
                 return Curriculum._load_curriculum(data_file)
-        except IOError:
-            raise CurriculumLoadingError(
-                "The file {0} could not be found.".format(config_path)
-            )
+        except OSError:
+            raise CurriculumLoadingError(f"The file {config_path} could not be found.")
-            raise CurriculumLoadingError(
-                "There was an error decoding {}".format(config_path)
-            )
+            raise CurriculumLoadingError(f"There was an error decoding {config_path}")
 
     @staticmethod
     def _load_curriculum(fp: TextIO) -> Dict:

ml-agents/mlagents/trainers/ghost/controller.py

         """
         self._queue.append(self._learning_team)
         self._learning_team = self._queue.popleft()
-        logger.debug(
-            "Learning team {} swapped on step {}".format(self._learning_team, step)
-        )
+        logger.debug(f"Learning team {self._learning_team} swapped on step {step}")
         self._changed_training_team = True
 
     # Adapted from https://github.com/Unity-Technologies/ml-agents/pull/1975 and

ml-agents/mlagents/trainers/ghost/trainer.py

         :param run_id: The identifier of the current run
         """
 
-        super(GhostTrainer, self).__init__(
+        super().__init__(
             brain_name, trainer_parameters, training, run_id, reward_buff_cap
         )
 

ml-agents/mlagents/trainers/models.py

             )
         else:
             raise UnityTrainerException(
-                "The learning rate schedule {} is invalid.".format(lr_schedule)
+                f"The learning rate schedule {lr_schedule} is invalid."
             )
         return learning_rate
 
                     h_size,
                     activation=activation,
                     reuse=reuse,
-                    name="hidden_{}".format(i),
+                    name=f"hidden_{i}",
                     kernel_initializer=tf.initializers.variance_scaling(1.0),
                 )
         return hidden
         """
         value_heads = {}
         for name in stream_names:
-            value = tf.layers.dense(hidden_input, 1, name="{}_value".format(name))
+            value = tf.layers.dense(hidden_input, 1, name=f"{name}_value")
             value_heads[name] = value
         value = tf.reduce_mean(list(value_heads.values()), 0)
         return value_heads, value

ml-agents/mlagents/trainers/policy/tf_policy.py

             self.sequence_length = trainer_parameters["sequence_length"]
             if self.m_size == 0:
                 raise UnityPolicyException(
-                    "The memory size for brain {0} is 0 even "
+                    "The memory size for brain {} is 0 even "
-                    "The memory size for brain {0} is {1} "
+                    "The memory size for brain {} is {} "
                     "but it must be divisible by 2.".format(
                         brain.brain_name, self.m_size
                     )
             ckpt = tf.train.get_checkpoint_state(model_path)
             if ckpt is None:
                 raise UnityPolicyException(
-                    "The model {0} could not be loaded. Make "
+                    "The model {} could not be loaded. Make "
                     "sure you specified the right "
                     "--run-id and that the previous run you are loading from had the same "
                     "behavior names.".format(model_path)
             except tf.errors.NotFoundError:
                 raise UnityPolicyException(
-                    "The model {0} was found but could not be loaded. Make "
+                    "The model {} was found but could not be loaded. Make "
                     "sure the model is from the same version of ML-Agents, has the same behavior parameters, "
                     "and is using the same trainer configuration as the current run.".format(
                         model_path
                     )
                 )
             else:
-                logger.info(
-                    "Resuming training from step {}.".format(self.get_current_step())
-                )
+                logger.info(f"Resuming training from step {self.get_current_step()}.")
 
     def initialize_or_load(self):
         # If there is an initialize path, load from that. Else, load from the set model path.

ml-agents/mlagents/trainers/ppo/optimizer.py

         self.old_values = {}
         for name in value_heads.keys():
             returns_holder = tf.placeholder(
-                shape=[None], dtype=tf.float32, name="{}_returns".format(name)
+                shape=[None], dtype=tf.float32, name=f"{name}_returns"
-                shape=[None], dtype=tf.float32, name="{}_value_estimate".format(name)
+                shape=[None], dtype=tf.float32, name=f"{name}_value_estimate"
             )
             self.returns_holders[name] = returns_holder
             self.old_values[name] = old_value
             self.all_old_log_probs: mini_batch["action_probs"],
         }
         for name in self.reward_signals:
-            feed_dict[self.returns_holders[name]] = mini_batch[
-                "{}_returns".format(name)
-            ]
-            feed_dict[self.old_values[name]] = mini_batch[
-                "{}_value_estimates".format(name)
-            ]
+            feed_dict[self.returns_holders[name]] = mini_batch[f"{name}_returns"]
+            feed_dict[self.old_values[name]] = mini_batch[f"{name}_value_estimates"]
 
         if self.policy.output_pre is not None and "actions_pre" in mini_batch:
             feed_dict[self.policy.output_pre] = mini_batch["actions_pre"]

ml-agents/mlagents/trainers/ppo/trainer.py

         :param seed: The seed the model will be initialized with
         :param run_id: The identifier of the current run
         """
-        super(PPOTrainer, self).__init__(
+        super().__init__(
             brain_name, trainer_parameters, training, run_id, reward_buff_cap
         )
         self.param_keys = [
             trajectory.done_reached and not trajectory.max_step_reached,
         )
         for name, v in value_estimates.items():
-            agent_buffer_trajectory["{}_value_estimates".format(name)].extend(v)
+            agent_buffer_trajectory[f"{name}_value_estimates"].extend(v)
             self._stats_reporter.add_stat(
                 self.optimizer.reward_signals[name].value_name, np.mean(v)
             )
             evaluate_result = reward_signal.evaluate_batch(
                 agent_buffer_trajectory
             ).scaled_reward
-            agent_buffer_trajectory["{}_rewards".format(name)].extend(evaluate_result)
+            agent_buffer_trajectory[f"{name}_rewards"].extend(evaluate_result)
             # Report the reward signals
             self.collected_rewards[name][agent_id] += np.sum(evaluate_result)
 
         for name in self.optimizer.reward_signals:
             bootstrap_value = value_next[name]
 
-            local_rewards = agent_buffer_trajectory[
-                "{}_rewards".format(name)
-            ].get_batch()
+            local_rewards = agent_buffer_trajectory[f"{name}_rewards"].get_batch()
-                "{}_value_estimates".format(name)
+                f"{name}_value_estimates"
             ].get_batch()
             local_advantage = get_gae(
                 rewards=local_rewards,
             )
             local_return = local_advantage + local_value_estimates
             # This is later use as target for the different value estimates
-            agent_buffer_trajectory["{}_returns".format(name)].set(local_return)
-            agent_buffer_trajectory["{}_advantage".format(name)].set(local_advantage)
+            agent_buffer_trajectory[f"{name}_returns"].set(local_return)
+            agent_buffer_trajectory[f"{name}_advantage"].set(local_advantage)
             tmp_advantages.append(local_advantage)
             tmp_returns.append(local_return)
 

ml-agents/mlagents/trainers/sac/network.py

         """
         self.value_heads = {}
         for name in stream_names:
-            value = tf.layers.dense(hidden_input, 1, name="{}_value".format(name))
+            value = tf.layers.dense(hidden_input, 1, name=f"{name}_value")
             self.value_heads[name] = value
         self.value = tf.reduce_mean(list(self.value_heads.values()), 0)
 
 
             q1_heads = {}
             for name in stream_names:
-                _q1 = tf.layers.dense(q1_hidden, num_outputs, name="{}_q1".format(name))
+                _q1 = tf.layers.dense(q1_hidden, num_outputs, name=f"{name}_q1")
                 q1_heads[name] = _q1
 
             q1 = tf.reduce_mean(list(q1_heads.values()), axis=0)
 
             q2_heads = {}
             for name in stream_names:
-                _q2 = tf.layers.dense(q2_hidden, num_outputs, name="{}_q2".format(name))
+                _q2 = tf.layers.dense(q2_hidden, num_outputs, name=f"{name}_q2")
                 q2_heads[name] = _q2
 
             q2 = tf.reduce_mean(list(q2_heads.values()), axis=0)

ml-agents/mlagents/trainers/sac/optimizer.py

                 )
 
             rewards_holder = tf.placeholder(
-                shape=[None], dtype=tf.float32, name="{}_rewards".format(name)
+                shape=[None], dtype=tf.float32, name=f"{name}_rewards"
             )
             self.rewards_holders[name] = rewards_holder
 
             self.policy.mask_input: batch["masks"] * burn_in_mask,
         }
         for name in self.reward_signals:
-            feed_dict[self.rewards_holders[name]] = batch["{}_rewards".format(name)]
+            feed_dict[self.rewards_holders[name]] = batch[f"{name}_rewards"]
 
         if self.policy.use_continuous_act:
             feed_dict[self.policy_network.external_action_in] = batch["actions"]

ml-agents/mlagents/trainers/sac/trainer.py

         filename = os.path.join(
             self.trainer_parameters["model_path"], "last_replay_buffer.hdf5"
         )
-        logger.info("Saving Experience Replay Buffer to {}".format(filename))
+        logger.info(f"Saving Experience Replay Buffer to {filename}")
         with open(filename, "wb") as file_object:
             self.update_buffer.save_to_file(file_object)
 
         filename = os.path.join(
             self.trainer_parameters["model_path"], "last_replay_buffer.hdf5"
         )
-        logger.info("Loading Experience Replay Buffer from {}".format(filename))
+        logger.info(f"Loading Experience Replay Buffer from {filename}")
         with open(filename, "rb+") as file_object:
             self.update_buffer.load_from_file(file_object)
         logger.info(
 
         batch_update_stats: Dict[str, list] = defaultdict(list)
         while self.step / self.update_steps > self.steps_per_update:
-            logger.debug("Updating SAC policy at step {}".format(self.step))
+            logger.debug(f"Updating SAC policy at step {self.step}")
             buffer = self.update_buffer
             if (
                 self.update_buffer.num_experiences
                 )
                 # Get rewards for each reward
                 for name, signal in self.optimizer.reward_signals.items():
-                    sampled_minibatch[
-                        "{}_rewards".format(name)
-                    ] = signal.evaluate_batch(sampled_minibatch).scaled_reward
+                    sampled_minibatch[f"{name}_rewards"] = signal.evaluate_batch(
+                        sampled_minibatch
+                    ).scaled_reward
 
                 update_stats = self.optimizer.update(sampled_minibatch, n_sequences)
                 for stat_name, value in update_stats.items():
             # Get minibatches for reward signal update if needed
             reward_signal_minibatches = {}
             for name, signal in self.optimizer.reward_signals.items():
-                logger.debug("Updating {} at step {}".format(name, self.step))
+                logger.debug(f"Updating {name} at step {self.step}")
                 # Some signals don't need a minibatch to be sampled - so we don't!
                 if signal.update_dict:
                     reward_signal_minibatches[name] = buffer.sample_mini_batch(

ml-agents/mlagents/trainers/sampler_class.py

         for param_name, cur_param_dict in self.reset_param_dict.items():
             if "sampler-type" not in cur_param_dict:
                 raise SamplerException(
-                    "'sampler_type' argument hasn't been supplied for the {0} parameter".format(
+                    "'sampler_type' argument hasn't been supplied for the {} parameter".format(
                         param_name
                     )
                 )

ml-agents/mlagents/trainers/stats.py

             )
             if self.self_play and "Self-play/ELO" in values:
                 elo_stats = values["Self-play/ELO"]
-                logger.info("{} ELO: {:0.3f}. ".format(category, elo_stats.mean))
+                logger.info(f"{category} ELO: {elo_stats.mean:0.3f}. ")
         else:
             logger.info(
                 "{}: Step: {}. No episode was completed since last summary. {}".format(
     ) -> None:
         if property_type == StatsPropertyType.HYPERPARAMETERS:
             logger.info(
-                """Hyperparameters for behavior name {0}: \n{1}""".format(
+                """Hyperparameters for behavior name {}: \n{}""".format(
                     category, self._dict_to_str(value, 0)
                 )
             )
                 [
                     "\t"
                     + "  " * num_tabs
-                    + "{0}:\t{1}".format(
+                    + "{}:\t{}".format(
                         x, self._dict_to_str(param_dict[x], num_tabs + 1)
                     )
                     for x in param_dict
         self._maybe_create_summary_writer(category)
         for key, value in values.items():
             summary = tf.Summary()
-            summary.value.add(tag="{}".format(key), simple_value=value.mean)
+            summary.value.add(tag=f"{key}", simple_value=value.mean)
             self.summary_writers[category].add_summary(summary, step)
             self.summary_writers[category].flush()
 
         for file_name in os.listdir(directory_name):
             if file_name.startswith("events.out"):
                 logger.warning(
-                    "{} was left over from a previous run. Deleting.".format(file_name)
+                    f"{file_name} was left over from a previous run. Deleting."
                 )
                 full_fname = os.path.join(directory_name, file_name)
                 try:
                 s_op = tf.summary.text(
                     name,
                     tf.convert_to_tensor(
-                        ([[str(x), str(input_dict[x])] for x in input_dict])
+                        [[str(x), str(input_dict[x])] for x in input_dict]
                     ),
                 )
                 s = sess.run(s_op)

ml-agents/mlagents/trainers/tests/test_simple_rl.py

 def default_reward_processor(rewards, last_n_rewards=5):
     rewards_to_use = rewards[-last_n_rewards:]
     # For debugging tests
-    print("Last {} rewards:".format(last_n_rewards), rewards_to_use)
+    print(f"Last {last_n_rewards} rewards:", rewards_to_use)
     return np.array(rewards[-last_n_rewards:], dtype=np.float32).mean()
 
 

ml-agents/mlagents/trainers/trainer/rl_trainer.py

     """
 
     def __init__(self, *args, **kwargs):
-        super(RLTrainer, self).__init__(*args, **kwargs)
+        super().__init__(*args, **kwargs)
         # Make sure we have at least one reward_signal
         if not self.trainer_parameters["reward_signals"]:
             raise UnityTrainerException(

ml-agents/mlagents/trainers/trainer/trainer.py

         for k in self.param_keys:
             if k not in self.trainer_parameters:
                 raise UnityTrainerException(
-                    "The hyper-parameter {0} could not be found for the {1} trainer of "
-                    "brain {2}.".format(k, self.__class__, self.brain_name)
+                    "The hyper-parameter {} could not be found for the {} trainer of "
+                    "brain {}.".format(k, self.__class__, self.brain_name)
                 )
 
     @property

ml-agents/mlagents/trainers/trainer_controller.py

 from mlagents.trainers.agent_processor import AgentManager
 
 
-class TrainerController(object):
+class TrainerController:
     def __init__(
         self,
         trainer_factory: TrainerFactory,

ml-agents/mlagents/trainers/trainer_util.py

     try:
         with open(config_path) as data_file:
             return _load_config(data_file)
-    except IOError:
+    except OSError:
         abs_path = os.path.abspath(config_path)
         raise TrainerConfigError(f"Config file could not be found at {abs_path}.")
     except UnicodeDecodeError:

ml-agents/setup.py

-from io import open
 import os
 import sys
 
         tag = os.getenv("CIRCLE_TAG")
 
         if tag != EXPECTED_TAG:
-            info = "Git tag: {0} does not match the expected tag of this app: {1}".format(
+            info = "Git tag: {} does not match the expected tag of this app: {}".format(
                 tag, EXPECTED_TAG
             )
             sys.exit(info)
         # Test-only dependencies should go in test_requirements.txt, not here.
         "grpcio>=1.11.0",
         "h5py>=2.9.0",
-        "mlagents_envs=={}".format(VERSION),
+        f"mlagents_envs=={VERSION}",
         "numpy>=1.13.3,<2.0",
         "Pillow>=4.2.1",
         "protobuf>=3.6",

ml-agents/tests/yamato/check_coverage_percent.py

-from __future__ import print_function
 import sys
 import os
 
             summary_xml = os.path.join(dirpath, SUMMARY_XML_FILENAME)
             break
     if not summary_xml:
-        print("Couldn't find {} in root directory".format(SUMMARY_XML_FILENAME))
+        print(f"Couldn't find {SUMMARY_XML_FILENAME} in root directory")
         sys.exit(1)
 
     with open(summary_xml) as f:

ml-agents/tests/yamato/scripts/run_gym.py

 
         if len(env.observation_space.shape) == 1:
             # Examine the initial vector observation
-            print("Agent observations look like: \n{}".format(initial_observations))
+            print(f"Agent observations look like: \n{initial_observations}")
 
         for _episode in range(10):
             env.reset()
                 actions = env.action_space.sample()
                 obs, reward, done, _ = env.step(actions)
                 episode_rewards += reward
-            print("Total reward this episode: {}".format(episode_rewards))
+            print(f"Total reward this episode: {episode_rewards}")
     finally:
         env.close()
 

ml-agents/tests/yamato/scripts/run_llapi.py

                 if tracked_agent in terminal_steps:
                     episode_rewards += terminal_steps[tracked_agent].reward
                     done = True
-            print("Total reward this episode: {}".format(episode_rewards))
+            print(f"Total reward this episode: {episode_rewards}")
     finally:
         env.close()
 

ml-agents/tests/yamato/yamato_utils.py

     if extra_packages:
         pip_commands += extra_packages
     for cmd in pip_commands:
+        pip_index_url = "--index-url https://artifactory.prd.it.unity3d.com/artifactory/api/pypi/pypi/simple"
-            f"source {venv_path}/bin/activate; python -m pip install -q {cmd}",
+            f"source {venv_path}/bin/activate; python -m pip install -q {cmd} {pip_index_url}",
             shell=True,
         )
     return venv_path

test_requirements.txt

 # Test-only dependencies should go here, not in setup.py
 pytest>4.0.0,<6.0.0
 pytest-cov==2.6.1
-pytest-xdist
+pytest-xdist==1.34.0
 
 # onnx doesn't currently have a wheel for 3.8
 tf2onnx>=1.5.5;python_version<'3.8'

utils/validate_versions.py

 
 
 def set_package_version(new_version: str) -> None:
-    with open(UNITY_PACKAGE_JSON_PATH, "r") as f:
+    with open(UNITY_PACKAGE_JSON_PATH) as f:
         package_json = json.load(f)
     if "version" in package_json:
         package_json["version"] = new_version