Format code on master (black + isort) (#538)

* Config files

* Add autoflake

* Update isort exclude; add pre-commit to requirements

* Manually fix a few bad cases

Author: lihuoran

.github/linters/pyproject.toml

@@ -0,0 +1,7 @@
+[tool.black]
+line-length = 120
+
+[tool.isort]
+profile = "black"
+line_length = 120
+known_first_party = "maro"

.github/linters/tox.ini

@@ -5,7 +5,9 @@ ignore =
     # line break after binary operator
     W504,
     # line break before binary operator
-    W503
+    W503,
+    # whitespace before ':'
+    E203
 
 exclude =
     .git,
@@ -27,14 +29,5 @@ max-line-length = 120
 per-file-ignores =
     # import not used: ignore in __init__.py files
     __init__.py:F401
-    # igore invalid escape sequence in cli main script to show banner
+    # ignore invalid escape sequence in cli main script to show banner
     maro.py:W605
-
-[isort]
-indent = "    "
-line_length = 120
-use_parentheses = True
-multi_line_output = 6
-known_first_party = maro
-filter_files = True
-skip_glob = maro/__init__.py, tests/*, examples/*, setup.py

.github/workflows/lint.yml

@@ -45,12 +45,12 @@ jobs:
         uses: github/super-linter@latest
         env:
           VALIDATE_ALL_CODEBASE: false
-          VALIDATE_PYTHON_PYLINT: false # disable pylint, as we have not configure it
-          VALIDATE_PYTHON_BLACK: false # same as above
+          VALIDATE_PYTHON_PYLINT: false # disable pylint, as we have not configured it
           VALIDATE_PYTHON_MYPY: false # same as above
           VALIDATE_JSCPD: false # Can not exclude specific file: https://github.com/kucherenko/jscpd/issues/215
           PYTHON_FLAKE8_CONFIG_FILE: tox.ini
-          PYTHON_ISORT_CONFIG_FILE: tox.ini
+          PYTHON_BLACK_CONFIG_FILE: pyproject.toml
+          PYTHON_ISORT_CONFIG_FILE: pyproject.toml
           EDITORCONFIG_FILE_NAME: ../../.editorconfig
           FILTER_REGEX_INCLUDE: maro/.*
           FILTER_REGEX_EXCLUDE: tests/.*

.pre-commit-config.yaml

@@ -0,0 +1,51 @@
+# Copyright (c) Microsoft Corporation.
+# Licensed under the MIT license.
+
+repos:
+- repo: https://github.com/myint/autoflake
+  rev: v1.4
+  hooks:
+    - id: autoflake
+      args:
+        - --in-place
+        - --remove-unused-variables
+        - --remove-all-unused-imports
+      exclude: .*/__init__\.py|setup\.py
+- repo: https://github.com/pycqa/isort
+  rev: 5.10.1
+  hooks:
+    - id: isort
+      args:
+        - --settings-path=.github/linters/pyproject.toml
+        - --check
+- repo: https://github.com/asottile/add-trailing-comma
+  rev: v2.2.3
+  hooks:
+    - id: add-trailing-comma
+      name: add-trailing-comma (1st round)
+- repo: https://github.com/psf/black
+  rev: 22.3.0
+  hooks:
+    - id: black
+      name: black (1st round)
+      args:
+        - --config=.github/linters/pyproject.toml
+- repo: https://github.com/asottile/add-trailing-comma
+  rev: v2.2.3
+  hooks:
+    - id: add-trailing-comma
+      name: add-trailing-comma (2nd round)
+- repo: https://github.com/psf/black
+  rev: 22.3.0
+  hooks:
+    - id: black
+      name: black (2nd round)
+      args:
+        - --config=.github/linters/pyproject.toml
+- repo: https://gitlab.com/pycqa/flake8
+  rev: 3.7.9
+  hooks:
+    - id: flake8
+      args:
+        - --config=.github/linters/tox.ini
+      exclude: \.git|__pycache__|docs|build|dist|.*\.egg-info|docker_files|\.vscode|\.github|scripts|tests|maro\/backends\/.*.cp|setup.py

docs/source/conf.py

@@ -65,7 +65,7 @@
 
 
 # The name of the Pygments (syntax highlighting) style to use.
-pygments_style = 'sphinx'
+pygments_style = "sphinx"
 
 # If true, `todo` and `todoList` produce output, else they produce nothing.
 todo_include_todos = False

examples/cim/rl/algorithms/ac.py

@@ -1,14 +1,12 @@
 # Copyright (c) Microsoft Corporation.
 # Licensed under the MIT license.
 
-from typing import Dict
-
 import torch
 from torch.optim import Adam, RMSprop
 
 from maro.rl.model import DiscreteACBasedNet, FullyConnected, VNet
 from maro.rl.policy import DiscretePolicyGradient
-from maro.rl.training.algorithms import ActorCriticTrainer, ActorCriticParams
+from maro.rl.training.algorithms import ActorCriticParams, ActorCriticTrainer
 
 actor_net_conf = {
     "hidden_dims": [256, 128, 64],
@@ -58,10 +56,10 @@ def get_ac(state_dim: int, name: str) -> ActorCriticTrainer:
         name=name,
         params=ActorCriticParams(
             get_v_critic_net_func=lambda: MyCriticNet(state_dim),
-            reward_discount=.0,
+            reward_discount=0.0,
             grad_iters=10,
             critic_loss_cls=torch.nn.SmoothL1Loss,
             min_logp=None,
-            lam=.0,
+            lam=0.0,
         ),
     )

examples/cim/rl/algorithms/dqn.py

@@ -1,15 +1,13 @@
 # Copyright (c) Microsoft Corporation.
 # Licensed under the MIT license.
 
-from typing import Dict
-
 import torch
 from torch.optim import RMSprop
 
 from maro.rl.exploration import MultiLinearExplorationScheduler, epsilon_greedy
 from maro.rl.model import DiscreteQNet, FullyConnected
 from maro.rl.policy import ValueBasedPolicy
-from maro.rl.training.algorithms import DQNTrainer, DQNParams
+from maro.rl.training.algorithms import DQNParams, DQNTrainer
 
 q_net_conf = {
     "hidden_dims": [256, 128, 64, 32],
@@ -38,14 +36,18 @@ def get_dqn_policy(state_dim: int, action_num: int, name: str) -> ValueBasedPoli
         name=name,
         q_net=MyQNet(state_dim, action_num),
         exploration_strategy=(epsilon_greedy, {"epsilon": 0.4}),
-        exploration_scheduling_options=[(
-            "epsilon", MultiLinearExplorationScheduler, {
-                "splits": [(2, 0.32)],
-                "initial_value": 0.4,
-                "last_ep": 5,
-                "final_value": 0.0,
-            }
-        )],
+        exploration_scheduling_options=[
+            (
+                "epsilon",
+                MultiLinearExplorationScheduler,
+                {
+                    "splits": [(2, 0.32)],
+                    "initial_value": 0.4,
+                    "last_ep": 5,
+                    "final_value": 0.0,
+                },
+            ),
+        ],
         warmup=100,
     )
 
@@ -54,7 +56,7 @@ def get_dqn(name: str) -> DQNTrainer:
     return DQNTrainer(
         name=name,
         params=DQNParams(
-            reward_discount=.0,
+            reward_discount=0.0,
             update_target_every=5,
             num_epochs=10,
             soft_update_coef=0.1,

examples/cim/rl/algorithms/maddpg.py

@@ -2,30 +2,29 @@
 # Licensed under the MIT license.
 
 from functools import partial
-from typing import Dict, List
+from typing import List
 
 import torch
 from torch.optim import Adam, RMSprop
 
 from maro.rl.model import DiscreteACBasedNet, FullyConnected, MultiQNet
 from maro.rl.policy import DiscretePolicyGradient
-from maro.rl.training.algorithms import DiscreteMADDPGTrainer, DiscreteMADDPGParams
-
+from maro.rl.training.algorithms import DiscreteMADDPGParams, DiscreteMADDPGTrainer
 
 actor_net_conf = {
     "hidden_dims": [256, 128, 64],
     "activation": torch.nn.Tanh,
     "softmax": True,
     "batch_norm": False,
-    "head": True
+    "head": True,
 }
 critic_net_conf = {
     "hidden_dims": [256, 128, 64],
     "output_dim": 1,
     "activation": torch.nn.LeakyReLU,
     "softmax": False,
     "batch_norm": True,
-    "head": True
+    "head": True,
 }
 actor_learning_rate = 0.001
 critic_learning_rate = 0.001
@@ -64,9 +63,9 @@ def get_maddpg(state_dim: int, action_dims: List[int], name: str) -> DiscreteMAD
     return DiscreteMADDPGTrainer(
         name=name,
         params=DiscreteMADDPGParams(
-            reward_discount=.0,
+            reward_discount=0.0,
             num_epoch=10,
             get_q_critic_net_func=partial(get_multi_critic_net, state_dim, action_dims),
-            shared_critic=False
-        )
+            shared_critic=False,
+        ),
     )

examples/cim/rl/algorithms/ppo.py

@@ -15,11 +15,11 @@ def get_ppo(state_dim: int, name: str) -> PPOTrainer:
         name=name,
         params=PPOParams(
             get_v_critic_net_func=lambda: MyCriticNet(state_dim),
-            reward_discount=.0,
+            reward_discount=0.0,
             grad_iters=10,
             critic_loss_cls=torch.nn.SmoothL1Loss,
             min_logp=None,
-            lam=.0,
+            lam=0.0,
             clip_ratio=0.1,
         ),
     )

examples/cim/rl/config.py

@@ -4,7 +4,7 @@
 env_conf = {
     "scenario": "cim",
     "topology": "toy.4p_ssdd_l0.0",
-    "durations": 560
+    "durations": 560,
 }
 
 if env_conf["topology"].startswith("toy"):
@@ -17,27 +17,26 @@
 
 state_shaping_conf = {
     "look_back": 7,
-    "max_ports_downstream": 2
+    "max_ports_downstream": 2,
 }
 
 action_shaping_conf = {
     "action_space": [(i - 10) / 10 for i in range(21)],
     "finite_vessel_space": True,
-    "has_early_discharge": True
+    "has_early_discharge": True,
 }
 
 reward_shaping_conf = {
     "time_window": 99,
     "fulfillment_factor": 1.0,
     "shortage_factor": 1.0,
-    "time_decay": 0.97
+    "time_decay": 0.97,
 }
 
 # obtain state dimension from a temporary env_wrapper instance
-state_dim = (
-    (state_shaping_conf["look_back"] + 1) * (state_shaping_conf["max_ports_downstream"] + 1) * len(port_attributes)
-    + len(vessel_attributes)
-)
+state_dim = (state_shaping_conf["look_back"] + 1) * (state_shaping_conf["max_ports_downstream"] + 1) * len(
+    port_attributes,
+) + len(vessel_attributes)
 
 action_num = len(action_shaping_conf["action_space"])