Merge pull request #117 from NREL/ray_v2_3

Update to use ray v2.3 and gymnasium v0.26.3

examples/benchmarks/RL_workshop.ipynb

@@ -696,7 +696,7 @@
    "outputs": [],
    "source": [
     "# Here, obs is a list of MoleculeState observations, with the first entry corresponding to the parent node\n",
-    "obs = env.reset()"
+    "obs, info = env.reset()"
    ]
   },
   {
@@ -728,7 +728,7 @@
    ],
    "source": [
     "done = False\n",
-    "obs = env.reset()\n",
+    "obs, info = env.reset()\n",
     "np.random.seed(0)\n",
     "\n",
     "while not done:\n",
@@ -770,7 +770,7 @@
    "outputs": [],
    "source": [
     "model = policy_model()\n",
-    "obs = env.reset()"
+    "obs, info = env.reset()"
    ]
   },
   {
@@ -1045,7 +1045,7 @@
     "\n",
     "env_preprocessor = get_preprocessor(env.observation_space)(env.observation_space)\n",
     "policy = trainer.get_policy()\n",
-    "obs = env.reset()"
+    "obs, info = env.reset()"
    ]
   },
   {

rlmolecule/molecule_state.py

@@ -3,7 +3,7 @@
 from dataclasses import dataclass
 from typing import Any, Dict, List, Optional, Sequence, Type, Union
 
-import gym
+import gymnasium as gym
 import nfp
 import numpy as np
 import ray

tests/test_molecule_policy.py

@@ -28,7 +28,7 @@ def molecule_env(qed_root: MoleculeState):
 
 def test_policy_model(molecule_env, single_layer_model):
 
-    observation, reward, terminal, info = molecule_env.step(0)
+    observation, reward, terminal, truncated, info = molecule_env.step(0)
 
     preprocessor = get_preprocessor(molecule_env.observation_space)
     obs = preprocessor(molecule_env.observation_space).transform(observation)

tests/test_molecule_state.py

@@ -46,12 +46,12 @@ def test_prune_terminal(builder):
     assert repr(env.state.children[-1]) == "C (t)"
 
     # select the terminal state
-    obs, reward, terminal, info = env.step(len(env.state.children) - 1)
+    obs, reward, terminal, truncated, info = env.step(len(env.state.children) - 1)
     assert terminal
     assert np.isclose(reward, 0.3597849378839701)
 
-    obs = env.reset()
-    obs, reward, terminal, info = env.step(len(env.state.children) - 1)
+    obs, info = env.reset()
+    obs, reward, terminal, truncated, info = env.step(len(env.state.children) - 1)
     assert not terminal
     assert np.isclose(reward, 0)
 
@@ -75,12 +75,12 @@ def test_prune_terminal_ray(ray_init):
     assert repr(env.state.children[-1]) == "C (t)"
 
     # select the terminal state
-    obs, reward, terminal, info = env.step(len(env.state.children) - 1)
+    obs, reward, terminal, truncated, info = env.step(len(env.state.children) - 1)
     assert terminal
     assert np.isclose(reward, 0.3597849378839701)
 
-    obs = env.reset()
-    obs, reward, terminal, info = env.step(len(env.state.children) - 1)
+    obs, info = env.reset()
+    obs, reward, terminal, truncated, info = env.step(len(env.state.children) - 1)
     assert not terminal
     assert np.isclose(reward, 0)