Rename to RL-Zoo3 and better packaging (#291)

* Rename and better packaging

* Move plot scripts inside package

Author: araffin

.coveragerc

@@ -2,7 +2,7 @@
 branch = False
 omit =
     tests/*
-    rl_zoo/utils/plot.py
+    rl_zoo3/utils/plot.py
 
 [report]
 exclude_lines =

CHANGELOG.md

@@ -5,7 +5,7 @@
 - low pass filter was removed
 
 ### New Features
-- RL Zoo cli: `rl_zoo train` and `rl_zoo enjoy`
+- RL Zoo cli: `rl_zoo3 train` and `rl_zoo3 enjoy`
 
 ### Bug fixes
 

Makefile

@@ -1,4 +1,4 @@
-LINT_PATHS = *.py tests/ scripts/ rl_zoo/
+LINT_PATHS = *.py tests/ scripts/ rl_zoo3/
 
 # Run pytest and coverage report
 pytest:
@@ -10,7 +10,7 @@ check-trained-agents:
 
 # Type check
 type:
-	pytype -j auto rl_zoo/ tests/ scripts/ -d import-error
+	pytype -j auto rl_zoo3/ tests/ scripts/ -d import-error
 
 lint:
 	# stop the build if there are Python syntax errors or undefined names
@@ -44,12 +44,14 @@ docker-gpu:
 
 # PyPi package release
 release:
+	# rm -r build/* dist/*
 	python setup.py sdist
 	python setup.py bdist_wheel
 	twine upload dist/*
 
 # Test PyPi package release
 test-release:
+	# rm -r build/* dist/*
 	python setup.py sdist
 	python setup.py bdist_wheel
 	twine upload --repository-url https://test.pypi.org/legacy/ dist/*

README.md

@@ -154,13 +154,13 @@ python enjoy.py --algo algo_name --env env_id -f logs/ --exp-id 1 --load-last-ch
 
 Upload model to hub (same syntax as for `enjoy.py`):
 ```
-python -m rl_zoo.push_to_hub --algo ppo --env CartPole-v1 -f logs/ -orga sb3 -m "Initial commit"
+python -m rl_zoo3.push_to_hub --algo ppo --env CartPole-v1 -f logs/ -orga sb3 -m "Initial commit"
 ```
 you can choose custom `repo-name` (default: `{algo}-{env_id}`) by passing a `--repo-name` argument.
 
 Download model from hub:
 ```
-python -m rl_zoo.load_from_hub --algo ppo --env CartPole-v1 -f logs/ -orga sb3
+python -m rl_zoo3.load_from_hub --algo ppo --env CartPole-v1 -f logs/ -orga sb3
 ```
 
 ## Hyperparameter yaml syntax
@@ -255,7 +255,7 @@ for multiple, specify a list:
 
 ```yaml
 env_wrapper:
-    - rl_zoo.wrappers.DoneOnSuccessWrapper:
+    - rl_zoo3.wrappers.DoneOnSuccessWrapper:
         reward_offset: 1.0
     - sb3_contrib.common.wrappers.TimeFeatureWrapper
 ```
@@ -279,7 +279,7 @@ Following the same syntax as env wrappers, you can also add custom callbacks to
 
 ```yaml
 callback:
-  - rl_zoo.callbacks.ParallelTrainCallback:
+  - rl_zoo3.callbacks.ParallelTrainCallback:
       gradient_steps: 256
 ```
 
@@ -306,19 +306,19 @@ Note: if you want to pass a string, you need to escape it like that: `my_string:
 Record 1000 steps with the latest saved model:
 
 ```
-python -m rl_zoo.record_video --algo ppo --env BipedalWalkerHardcore-v3 -n 1000
+python -m rl_zoo3.record_video --algo ppo --env BipedalWalkerHardcore-v3 -n 1000
 ```
 
 Use the best saved model instead:
 
 ```
-python -m rl_zoo.record_video --algo ppo --env BipedalWalkerHardcore-v3 -n 1000 --load-best
+python -m rl_zoo3.record_video --algo ppo --env BipedalWalkerHardcore-v3 -n 1000 --load-best
 ```
 
 Record a video of a checkpoint saved during training (here the checkpoint name is `rl_model_10000_steps.zip`):
 
 ```
-python -m rl_zoo.record_video --algo ppo --env BipedalWalkerHardcore-v3 -n 1000 --load-checkpoint 10000
+python -m rl_zoo3.record_video --algo ppo --env BipedalWalkerHardcore-v3 -n 1000 --load-checkpoint 10000
 ```
 
 ## Record a Video of a Training Experiment
@@ -328,18 +328,18 @@ Apart from recording videos of specific saved models, it is also possible to rec
 Record 1000 steps for each checkpoint, latest and best saved models:
 
 ```
-python -m rl_zoo.record_training --algo ppo --env CartPole-v1 -n 1000 -f logs --deterministic
+python -m rl_zoo3.record_training --algo ppo --env CartPole-v1 -n 1000 -f logs --deterministic
 ```
 
 The previous command will create a `mp4` file. To convert this file to `gif` format as well:
 
 ```
-python -m rl_zoo.record_training --algo ppo --env CartPole-v1 -n 1000 -f logs --deterministic --gif
+python -m rl_zoo3.record_training --algo ppo --env CartPole-v1 -n 1000 -f logs --deterministic --gif
 ```
 
 ## Current Collection: 195+ Trained Agents!
 
-Final performance of the trained agents can be found in [`benchmark.md`](./benchmark.md). To compute them, simply run `python -m rl_zoo.benchmark`.
+Final performance of the trained agents can be found in [`benchmark.md`](./benchmark.md). To compute them, simply run `python -m rl_zoo3.benchmark`.
 
 List and videos of trained agents can be found on our Huggingface page: https://huggingface.co/sb3
 

docker/Dockerfile

@@ -18,7 +18,7 @@ COPY requirements.txt /tmp/
 
 
 RUN \
-    mkdir -p ${CODE_DIR}/rl_zoo && \
+    mkdir -p ${CODE_DIR}/rl_zoo3 && \
     pip uninstall -y stable-baselines3 && \
     pip install -r /tmp/requirements.txt && \
     pip install pip install highway-env==1.5.0 && \

enjoy.py

@@ -1,4 +1,4 @@
-from rl_zoo.enjoy import enjoy
+from rl_zoo3.enjoy import enjoy
 
 if __name__ == "__main__":
     enjoy()

hyperparams/her.yml

@@ -59,7 +59,7 @@ FetchSlide-v1:
 FetchPickAndPlace-v1:
   env_wrapper:
     - sb3_contrib.common.wrappers.TimeFeatureWrapper
-    # - rl_zoo.wrappers.DoneOnSuccessWrapper:
+    # - rl_zoo3.wrappers.DoneOnSuccessWrapper:
     #     reward_offset: 0
     #     n_successes: 4
     # - stable_baselines3.common.monitor.Monitor
@@ -96,7 +96,7 @@ FetchReach-v1:
 NeckGoalEnvRelativeSparse-v2:
   model_class: 'sac'
   # env_wrapper:
-  #   - rl_zoo.wrappers.HistoryWrapper:
+  #   - rl_zoo3.wrappers.HistoryWrapper:
   #       horizon: 2
   #   - sb3_contrib.common.wrappers.TimeFeatureWrapper
   n_timesteps: !!float 1e6
@@ -122,7 +122,7 @@ NeckGoalEnvRelativeSparse-v2:
 NeckGoalEnvRelativeDense-v2:
   model_class: 'sac'
   env_wrapper:
-    - rl_zoo.wrappers.HistoryWrapperObsDict:
+    - rl_zoo3.wrappers.HistoryWrapperObsDict:
         horizon: 2
   #   - sb3_contrib.common.wrappers.TimeFeatureWrapper
   n_timesteps: !!float 1e6

hyperparams/ppo.yml

@@ -319,7 +319,7 @@ MiniGrid-FourRooms-v0:
 
 CarRacing-v0:
   env_wrapper:
-    - rl_zoo.wrappers.FrameSkip:
+    - rl_zoo3.wrappers.FrameSkip:
         skip: 2
     - gym.wrappers.resize_observation.ResizeObservation:
         shape: 64

hyperparams/ppo_lstm.yml

@@ -132,7 +132,7 @@ BipedalWalker-v3:
 # TO BE TUNED
 BipedalWalkerHardcore-v3:
   # env_wrapper:
-  #   - rl_zoo.wrappers.FrameSkip:
+  #   - rl_zoo3.wrappers.FrameSkip:
   #       skip: 2
   normalize: true
   n_envs: 32
@@ -285,7 +285,7 @@ InvertedPendulumSwingupBulletEnv-v0:
 
 CarRacing-v0:
   env_wrapper:
-    # - rl_zoo.wrappers.FrameSkip:
+    # - rl_zoo3.wrappers.FrameSkip:
     #     skip: 2
     - gym.wrappers.resize_observation.ResizeObservation:
         shape: 64

hyperparams/sac.yml

@@ -16,7 +16,7 @@ MountainCarContinuous-v0:
 
 Pendulum-v1:
   # callback:
-  #   - rl_zoo.callbacks.ParallelTrainCallback
+  #   - rl_zoo3.callbacks.ParallelTrainCallback
   n_timesteps: 20000
   policy: 'MlpPolicy'
   learning_rate: !!float 1e-3
@@ -74,9 +74,9 @@ BipedalWalkerHardcore-v3:
 HalfCheetahBulletEnv-v0: &pybullet-defaults
   # env_wrapper:
   #   - sb3_contrib.common.wrappers.TimeFeatureWrapper
-  #   - rl_zoo.wrappers.DelayedRewardWrapper:
+  #   - rl_zoo3.wrappers.DelayedRewardWrapper:
   #       delay: 10
-  #   - rl_zoo.wrappers.HistoryWrapper:
+  #   - rl_zoo3.wrappers.HistoryWrapper:
   #       horizon: 10
   n_timesteps: !!float 1e6
   policy: 'MlpPolicy'
@@ -163,12 +163,12 @@ MinitaurBulletDuckEnv-v0:
 # To be tuned
 CarRacing-v0:
   env_wrapper:
-    - rl_zoo.wrappers.FrameSkip:
+    - rl_zoo3.wrappers.FrameSkip:
         skip: 2
     # wrapper from https://github.com/araffin/aae-train-donkeycar
     - ae.wrapper.AutoencoderWrapper:
         ae_path: "logs/car_racing_rgb_160.pkl"
-    - rl_zoo.wrappers.HistoryWrapper:
+    - rl_zoo3.wrappers.HistoryWrapper:
         horizon: 2
   # frame_stack: 4
   normalize: True
@@ -238,7 +238,7 @@ donkey-generated-track-v0:
   env_wrapper:
     - gym.wrappers.time_limit.TimeLimit:
         max_episode_steps: 500
-    - rl_zoo.wrappers.HistoryWrapper:
+    - rl_zoo3.wrappers.HistoryWrapper:
         horizon: 5
   n_timesteps: !!float 1e6
   policy: 'MlpPolicy'
@@ -262,9 +262,9 @@ donkey-generated-track-v0:
 NeckEnvRelative-v2:
   <<: *pybullet-defaults
   env_wrapper:
-    - rl_zoo.wrappers.HistoryWrapper:
+    - rl_zoo3.wrappers.HistoryWrapper:
         horizon: 2
-    # - rl_zoo.wrappers.LowPassFilterWrapper:
+    # - rl_zoo3.wrappers.LowPassFilterWrapper:
     #     freq: 2.0
     #     df: 25.0
   n_timesteps: !!float 1e6

hyperparams/tqc.yml

@@ -258,12 +258,12 @@ parking-v0:
 # Tuned
 CarRacing-v0:
   env_wrapper:
-    - rl_zoo.wrappers.FrameSkip:
+    - rl_zoo3.wrappers.FrameSkip:
         skip: 2
     # wrapper from https://github.com/araffin/aae-train-donkeycar
     - ae.wrapper.AutoencoderWrapper:
         ae_path: "logs/car_racing_rgb_160.pkl"
-    - rl_zoo.wrappers.HistoryWrapper:
+    - rl_zoo3.wrappers.HistoryWrapper:
         horizon: 2
   # frame_stack: 4
   normalize: True
@@ -280,7 +280,7 @@ RocketLander-v0:
   n_timesteps: !!float 3e6
   policy: 'MlpPolicy'
   env_wrapper:
-    - rl_zoo.wrappers.FrameSkip:
+    - rl_zoo3.wrappers.FrameSkip:
         skip: 4
-    - rl_zoo.wrappers.HistoryWrapper:
+    - rl_zoo3.wrappers.HistoryWrapper:
         horizon: 2

requirements.txt

@@ -7,7 +7,6 @@ gym-minigrid
 scikit-optimize
 optuna
 pytablewriter~=0.64
-seaborn
 pyyaml>=5.1
 cloudpickle>=1.5.0
 plotly

rl_zoo/cli.py

@@ -1,6 +1,6 @@
 import os
 
-from rl_zoo.utils import (
+from rl_zoo3.utils import (
     ALGOS,
     create_test_env,
     get_latest_run_id,

rl_zoo/version.txt

@@ -9,7 +9,7 @@
 import pytablewriter
 from stable_baselines3.common.results_plotter import load_results, ts2xy
 
-from rl_zoo.utils import get_hf_trained_models, get_latest_run_id, get_saved_hyperparams, get_trained_models
+from rl_zoo3.utils import get_hf_trained_models, get_latest_run_id, get_saved_hyperparams, get_trained_models
 
 parser = argparse.ArgumentParser()
 parser.add_argument("--log-dir", help="Root log folder", default="rl-trained-agents/", type=str)

rl_zoo/__init__.py rl_zoo3/__init__.py

@@ -0,0 +1,22 @@
+import sys
+
+from rl_zoo3.enjoy import enjoy
+from rl_zoo3.plots import all_plots, plot_from_file, plot_train
+from rl_zoo3.train import train
+
+
+def main():
+    script_name = sys.argv[1]
+    # Remove script name
+    del sys.argv[1]
+    # Execute known script
+    known_scripts = {
+        "train": train,
+        "enjoy": enjoy,
+        "plot_train": plot_train,
+        "plot_from_file": plot_from_file,
+        "all_plots": all_plots,
+    }
+    if script_name not in known_scripts.keys():
+        raise ValueError(f"The script {script_name} is unknown, please use one of {known_scripts.keys()}")
+    known_scripts[script_name]()

rl_zoo/benchmark.py rl_zoo3/benchmark.py

@@ -9,12 +9,12 @@
 from huggingface_sb3 import EnvironmentName
 from stable_baselines3.common.utils import set_random_seed
 
-import rl_zoo.import_envs  # noqa: F401 pylint: disable=unused-import
-from rl_zoo import ALGOS, create_test_env, get_saved_hyperparams
-from rl_zoo.callbacks import tqdm
-from rl_zoo.exp_manager import ExperimentManager
-from rl_zoo.load_from_hub import download_from_hub
-from rl_zoo.utils import StoreDict, get_model_path
+import rl_zoo3.import_envs  # noqa: F401 pylint: disable=unused-import
+from rl_zoo3 import ALGOS, create_test_env, get_saved_hyperparams
+from rl_zoo3.callbacks import tqdm
+from rl_zoo3.exp_manager import ExperimentManager
+from rl_zoo3.load_from_hub import download_from_hub
+from rl_zoo3.utils import StoreDict, get_model_path
 
 
 def enjoy():  # noqa: C901