Allow python script for hyperparameter configuration (#318)

* Allow loading configuration from python package in addition to yaml files.

* Add an example python configuration file.

* Add a test for training with the example python config file.

* Formatting fixes.

* Flake8 6 doesn't support inline comments

* Allow python scripts

* Remove debug print

* Add python config file usage example to README.md

* Update CHANGELOG.md

* Fix paths in python config example and add variant where we specify the python file instead of package.

* Update README

Co-authored-by: Antonin Raffin <antonin.raffin@dlr.de>
Co-authored-by: Antonin Raffin <antonin.raffin@ensta.org>

Author: ernestum

CHANGELOG.md

@@ -1,11 +1,13 @@
 ## Release 1.7.0a2 (WIP)
 
 ### Breaking Changes
+- `--yaml-file` argument was renamed to `-conf` (`--conf-file`) as now python file are supported too
 
 ### New Features
 - Specifying custom policies in yaml file is now supported (@Rick-v-E)
 - Added ``monitor_kwargs`` parameter
 - Handle the `env_kwargs` of `render:True` under the hood for panda-gym v1 envs in `enjoy` replay to match visualzation behavior of other envs
+- Added support for python config file
 
 ### Bug fixes
 - Allow `python -m rl_zoo3.cli` to be called directly

Makefile

@@ -1,4 +1,4 @@
-LINT_PATHS = *.py tests/ scripts/ rl_zoo3/
+LINT_PATHS = *.py tests/ scripts/ rl_zoo3/ hyperparams/python/*.py
 
 # Run pytest and coverage report
 pytest:

README.md

@@ -61,11 +61,23 @@ python train.py --algo algo_name --env env_id
 ```
 You can use `-P` (`--progress`) option to display a progress bar.
 
-Using a custom yaml file (which contains a `env_id` entry):
+Using a custom config file when it is a yaml file with a which contains a `env_id` entry:
 ```
-python train.py --algo algo_name --env env_id --yaml-file my_yaml.yml
+python train.py --algo algo_name --env env_id --conf-file my_yaml.yml
 ```
 
+You can also use a python file that contains a dictionary called `hyperparams` with an entry for each `env_id`.
+(see `hyperparams/python/ppo_config_example.py` for an example)
+```bash
+# You can pass a path to a python file
+python train.py --algo ppo --env MountainCarContinuous-v0 --conf-file hyperparams/python/ppo_config_example.py
+# Or pass a path to a file from a module (for instance my_package.my_file
+python train.py --algo ppo --env MountainCarContinuous-v0 --conf-file hyperparams.python.ppo_config_example
+```
+The advantage of this approach is that you can specify arbitrary python dictionaries
+and ensure that all their dependencies are imported in the config file itself.
+
+
 For example (with tensorboard support):
 ```
 python train.py --algo ppo --env CartPole-v1 --tensorboard-log /tmp/stable-baselines/
@@ -139,7 +151,7 @@ Remark: plotting with the `--rliable` option is usually slow as confidence inter
 
 ## Custom Environment
 
-The easiest way to add support for a custom environment is to edit `rl_zoo3/import_envs.py` and register your environment here. Then, you need to add a section for it in the hyperparameters file (`hyperparams/algo.yml` or a custom yaml file that you can specify using `--yaml-file` argument).
+The easiest way to add support for a custom environment is to edit `rl_zoo3/import_envs.py` and register your environment here. Then, you need to add a section for it in the hyperparameters file (`hyperparams/algo.yml` or a custom yaml file that you can specify using `--conf-file` argument).
 
 ## Enjoy a Trained Agent
 

hyperparams/python/ppo_config_example.py

@@ -0,0 +1,29 @@
+"""This file just serves as an example on how to configure the zoo
+using python scripts instead of yaml files."""
+import torch
+
+hyperparams = {
+    "MountainCarContinuous-v0": dict(
+        env_wrapper=[{"gym.wrappers.TimeLimit": {"max_episode_steps": 100}}],
+        normalize=True,
+        n_envs=1,
+        n_timesteps=20000.0,
+        policy="MlpPolicy",
+        batch_size=8,
+        n_steps=8,
+        gamma=0.9999,
+        learning_rate=7.77e-05,
+        ent_coef=0.00429,
+        clip_range=0.1,
+        n_epochs=2,
+        gae_lambda=0.9,
+        max_grad_norm=5,
+        vf_coef=0.19,
+        use_sde=True,
+        policy_kwargs=dict(
+            log_std_init=-3.29,
+            ortho_init=False,
+            activation_fn=torch.nn.ReLU,
+        ),
+    )
+}

rl_zoo3/exp_manager.py

@@ -1,4 +1,5 @@
 import argparse
+import importlib
 import os
 import pickle as pkl
 import time
@@ -93,7 +94,7 @@ def __init__(
         n_eval_envs: int = 1,
         no_optim_plots: bool = False,
         device: Union[th.device, str] = "auto",
-        yaml_file: Optional[str] = None,
+        config: Optional[str] = None,
         show_progress: bool = False,
     ):
         super().__init__()
@@ -108,7 +109,7 @@ def __init__(
             # Take the root folder
             default_path = Path(__file__).parent.parent
 
-        self.yaml_file = yaml_file or str(default_path / f"hyperparams/{self.algo}.yml")
+        self.config = config or str(default_path / f"hyperparams/{self.algo}.yml")
         self.env_kwargs = {} if env_kwargs is None else env_kwargs
         self.n_timesteps = n_timesteps
         self.normalize = False
@@ -281,16 +282,28 @@ def _save_config(self, saved_hyperparams: Dict[str, Any]) -> None:
         print(f"Log path: {self.save_path}")
 
     def read_hyperparameters(self) -> Tuple[Dict[str, Any], Dict[str, Any]]:
-        # Load hyperparameters from yaml file
-        print(f"Loading hyperparameters from: {self.yaml_file}")
-        with open(self.yaml_file) as f:
-            hyperparams_dict = yaml.safe_load(f)
-            if self.env_name.gym_id in list(hyperparams_dict.keys()):
-                hyperparams = hyperparams_dict[self.env_name.gym_id]
-            elif self._is_atari:
-                hyperparams = hyperparams_dict["atari"]
-            else:
-                raise ValueError(f"Hyperparameters not found for {self.algo}-{self.env_name.gym_id}")
+        print(f"Loading hyperparameters from: {self.config}")
+
+        if self.config.endswith(".yml") or self.config.endswith(".yaml"):
+            # Load hyperparameters from yaml file
+            with open(self.config) as f:
+                hyperparams_dict = yaml.safe_load(f)
+        elif self.config.endswith(".py"):
+            global_variables = {}
+            # Load hyperparameters from python file
+            exec(Path(self.config).read_text(), global_variables)
+            hyperparams_dict = global_variables["hyperparams"]
+        else:
+            # Load hyperparameters from python package
+            hyperparams_dict = importlib.import_module(self.config).hyperparams
+            # raise ValueError(f"Unsupported config file format: {self.config}")
+
+        if self.env_name.gym_id in list(hyperparams_dict.keys()):
+            hyperparams = hyperparams_dict[self.env_name.gym_id]
+        elif self._is_atari:
+            hyperparams = hyperparams_dict["atari"]
+        else:
+            raise ValueError(f"Hyperparameters not found for {self.algo}-{self.env_name.gym_id} in {self.config}")
 
         if self.custom_hyperparams is not None:
             # Overwrite hyperparams if needed
@@ -336,6 +349,10 @@ def _preprocess_normalization(self, hyperparams: Dict[str, Any]) -> Dict[str, An
                 self.normalize_kwargs = eval(self.normalize)
                 self.normalize = True
 
+            if isinstance(self.normalize, dict):
+                self.normalize_kwargs = self.normalize
+                self.normalize = True
+
             # Use the same discount factor as for the algorithm
             if "gamma" in hyperparams:
                 self.normalize_kwargs["gamma"] = hyperparams["gamma"]

rl_zoo3/train.py

@@ -122,7 +122,19 @@ def train():
         help="Overwrite hyperparameter (e.g. learning_rate:0.01 train_freq:10)",
     )
     parser.add_argument(
-        "-yaml", "--yaml-file", type=str, default=None, help="Custom yaml file from which the hyperparameters will be loaded"
+        "-conf",
+        "--conf-file",
+        type=str,
+        default=None,
+        help="Custom yaml file or python package from which the hyperparameters will be loaded."
+        "We expect that python packages contain a dictionary called 'hyperparams' which contains a key for each environment.",
+    )
+    parser.add_argument(
+        "-yaml",
+        "--yaml-file",
+        type=str,
+        default=None,
+        help="This parameter is deprecated, please use `--conf-file` instead",
     )
     parser.add_argument("-uuid", "--uuid", action="store_true", default=False, help="Ensure that the run has a unique ID")
     parser.add_argument(
@@ -150,6 +162,11 @@ def train():
     env_id = args.env
     registered_envs = set(gym.envs.registry.env_specs.keys())  # pytype: disable=module-attr
 
+    if args.yaml_file is not None:
+        raise ValueError(
+            "The`--yaml-file` parameter is deprecated and will be removed in RL Zoo3 v1.8, please use `--conf-file` instead",
+        )
+
     # If the environment is not found, suggest the closest match
     if env_id not in registered_envs:
         try:
@@ -234,7 +251,7 @@ def train():
         n_eval_envs=args.n_eval_envs,
         no_optim_plots=args.no_optim_plots,
         device=args.device,
-        yaml_file=args.yaml_file,
+        config=args.conf_file,
         show_progress=args.progress,
     )
 

rl_zoo3/version.txt

@@ -1 +1 @@
-1.7.0a2
+1.7.0a3

setup.cfg

@@ -16,7 +16,8 @@ markers =
 inputs = .
 
 [flake8]
-ignore = W503,W504,E203,E231  # line breaks before and after binary operators
+# line breaks before and after binary operators
+ignore = W503,W504,E203,E231
 # Ignore import not used when aliases are defined
 per-file-ignores =
   ./rl_zoo3/__init__.py:F401

tests/test_train.py

@@ -116,7 +116,7 @@ def test_custom_yaml(tmp_path):
         "CartPole-v1",
         "--log-folder",
         tmp_path,
-        "-yaml",
+        "-conf",
         "hyperparams/a2c.yml",
         "-params",
         "n_envs:2",
@@ -129,3 +129,23 @@ def test_custom_yaml(tmp_path):
 
     return_code = subprocess.call(["python", "train.py"] + args)
     _assert_eq(return_code, 0)
+
+
+@pytest.mark.parametrize("config_file", ["hyperparams.python.ppo_config_example", "hyperparams/python/ppo_config_example.py"])
+def test_python_config_file(tmp_path, config_file):
+    # Use the example python config file for training
+    args = [
+        "-n",
+        str(N_STEPS),
+        "--algo",
+        "ppo",
+        "--env",
+        "MountainCarContinuous-v0",
+        "--log-folder",
+        tmp_path,
+        "-conf",
+        config_file,
+    ]
+
+    return_code = subprocess.call(["python", "train.py"] + args)
+    _assert_eq(return_code, 0)