Merge pull request #18 from UoA-CARES/chore/requirements

Chore/requirements

.pylintrc

@@ -13,7 +13,7 @@ disable=
 max-line-length=130
 
 [MASTER]
-extension-pkg-whitelist=cv2
+extension-pkg-whitelist=cv2, pydantic
 
 [TYPECHECK]
 generated-members=cv2.*, numpy.*, torch.*

README.md

@@ -22,11 +22,13 @@ This package is a basic example of running the CARES RL algorithms on OpenAI/DMC
 
 An example is found below for running on the OpenAI, DMCS, and pyboy environments with TD3/NaSATD3 through console
 ```
+python train.py run --gym openai --task CartPole-v1 DQN
+
 python train.py run --gym openai --task HalfCheetah-v4 TD3
 
 python3 train.py run --gym dmcs --domain ball_in_cup --task catch TD3
 
-python3 train.py run --gym pyboy --task pokemon NaSATD3 --image_observation=True
+python3 train.py run --gym pyboy --task pokemon NaSATD3 --image_observation=1
 ```
 
 An example is found below for running using pre-defined configuration files - note these directories will need to exist on your machine for this to function. Examples of each configuration can be found in the `configs` folder and under https://github.com/UoA-CARES/cares_reinforcement_learning/

requirements.txt

@@ -1,5 +1,7 @@
 dm_control==1.0.15
 gymnasium==0.29.0
+gymnasium[classic-control]==0.29.0
+gymnasium[mujoco]==0.29.0
 numpy==1.23.5
 opencv-contrib-python==4.6.0.66
 pydantic==1.10.13

scripts/envrionments/environment_factory.py

@@ -34,4 +34,4 @@ def create_environment(self, config: GymEnvironmentConfig) -> GymEnvironment:
             env = create_pyboy_environment(config)
         else:
             raise ValueError(f"Unkown environment: {config.gym}")
-        return ImageWrapper(env) if config.image_observation else env
+        return ImageWrapper(env) if bool(config.image_observation) else env

scripts/train_loops/policy_loop.py

@@ -76,7 +76,7 @@ def policy_based_train(
 
     # Algorthm specific attributes - e.g. NaSA-TD3 dd
     intrinsic_on = (
-        alg_config.intrinsic_on if hasattr(alg_config, "intrinsic_on") else False
+        bool(alg_config.intrinsic_on) if hasattr(alg_config, "intrinsic_on") else False
     )
 
     min_noise = alg_config.min_noise if hasattr(alg_config, "min_noise") else 0

scripts/util/configurations.py

@@ -30,12 +30,12 @@ class GymEnvironmentConfig(EnvironmentConfig):
     gym: str = Field(description="Gym Environment <openai, dmcs, pyboy>")
     task: str
     domain: Optional[str] = ""
-    image_observation: Optional[bool] = False
+    image_observation: Optional[int] = 0
 
     rom_path: Optional[str] = f"{Path.home()}/cares_rl_configs"
     act_freq: Optional[int] = 24
     emulation_speed: Optional[int] = 0
-    headless: Optional[bool] = False
+    headless: Optional[int] = 0
 
 
 # class OpenAIEnvironmentConfig(GymEnvironmentConfig):