Fix: Forgot 'fi' to end if clause in bash snippet

.github/workflows/ensure-release-tagged.yaml

@@ -1,26 +0,0 @@
-name: Ensure Tagged Commits on Release
-
-on:
-  pull_request:
-    branches:
-      - release
-
-jobs:
-  check_tag:
-    runs-on: ubuntu-latest
-    steps:
-    - name: Check out code
-      uses: actions/checkout@v4
-      with:
-        fetch-depth: 0
-
-    - name: Check if base commit of PR is tagged
-      run: |
-        BASE_COMMIT=$(jq -r .pull_request.base.sha < "$GITHUB_EVENT_PATH")
-        TAG=$(git tag --contains $BASE_COMMIT)
-        if [ -z "$TAG" ]; then
-          echo "Base commit of PR is not tagged. PRs onto release must be tagged with the version number."
-          exit 1
-        fi
-        echo "Base commit of PR is tagged. Check passed."
-

.github/workflows/ensure-version-consistency.yaml

@@ -0,0 +1,52 @@
+name: Ensure Version Consistency on PR to Release
+
+on:
+  pull_request:
+    branches:
+      - release
+
+jobs:
+  check_version_and_tag:
+    runs-on: ubuntu-latest
+    strategy:
+      fail-fast: true  # Terminate the job immediately if any step fails
+    steps:
+      - name: Check out code
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 0 # Necessary to fetch all tags for comparison
+
+      - name: Set up Python
+        uses: actions/setup-python@v4
+        with:
+          python-version: '3.x'
+
+      - name: Install dependencies
+        run: |
+          python -m pip install toml
+
+      - name: Extract version from pyproject.toml
+        run: |
+          echo "Extracting version from pyproject.toml"
+          VERSION=$(python -c 'import toml; print(toml.load("pyproject.toml")["project"]["version"])')
+          echo "Version in pyproject.toml is $VERSION"
+          echo "VERSION=$VERSION" >> $GITHUB_ENV
+
+      - name: Get tag for the PR's head commit
+        run: |
+          PR_HEAD_SHA=$(jq -r .pull_request.head.sha < "$GITHUB_EVENT_PATH")
+          TAG=$(git tag --contains $PR_HEAD_SHA)
+          echo "Tag on PR's head commit is $TAG"
+          echo "TAG=$TAG" >> $GITHUB_ENV
+
+      - name: Compare version and tag
+        run: |
+          if [ -z "$TAG" ]; then
+            echo "Head commit of PR is not tagged. Ensure the head commit of PRs onto release is tagged with the version number."
+            exit 1
+          elif [ "$VERSION" != "$TAG" ]; then
+            echo "Version in pyproject.toml ($VERSION) does not match the git tag ($TAG)."
+            exit 1
+          else
+            echo "Version and git tag match. Check passed."
+          fi

.github/workflows/host_docs_static.yml

@@ -1,40 +0,0 @@
-name: Deploy static docs to Pages
-
-on:
-  push:
-    branches: ["release"]
-
-  # Allows you to run this workflow manually from the Actions tab
-  workflow_dispatch:
-
-# Sets permissions of the GITHUB_TOKEN to allow deployment to GitHub Pages
-permissions:
-  contents: read
-  pages: write
-  id-token: write
-
-# Allow only one concurrent deployment, skipping runs queued between the run in-progress and latest queued.
-# However, do NOT cancel in-progress runs as we want to allow these production deployments to complete.
-concurrency:
-  group: "pages"
-  cancel-in-progress: false
-
-jobs:
-  # Single deploy job since we're just deploying
-  deploy:
-    environment:
-      name: github-pages
-      url: ${{ steps.deployment.outputs.page_url }}
-    runs-on: ubuntu-latest
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v4
-      - name: Setup Pages
-        uses: actions/configure-pages@v4
-      - name: Upload artifact
-        uses: actions/upload-pages-artifact@v3
-        with:
-          path: 'docs/build/html'
-      - name: Deploy to GitHub Pages
-        id: deployment
-        uses: actions/deploy-pages@v4

.github/workflows/publish-to-pypi.yml

@@ -8,6 +8,8 @@ on:
 jobs:
   publish:
     name: Publish to PyPI
+    strategy:
+      fail-fast: true  # Terminate the job immediately if any step fails
     runs-on: ubuntu-latest
     steps:
     - name: Check out code
@@ -15,19 +17,24 @@ jobs:
       with:
         fetch-depth: 0  # This fetches all history for all branches and tags
 
-    - name: Check if commit is tagged
+    - name: Set up Python
+      uses: actions/setup-python@v4
+      with:
+        python-version: "3.x"
+
+    - name: Validate version against tag
       run: |
+        VERSION=$(python -c 'import toml; print(toml.load("pyproject.toml")["project"]["version"])')
         TAG=$(git tag --contains HEAD)
         if [ -z "$TAG" ]; then
           echo "Commit is not tagged. Failing the workflow."
           exit 1
         fi
-        echo "Commit is tagged. Proceeding with the workflow."
+        if [ "$VERSION" != "$TAG" ]; then
-
+          echo "Version in pyproject.toml ($VERSION) does not match the git tag ($TAG). Failing the workflow."
-    - name: Set up Python
+          exit 1
-      uses: actions/setup-python@v4
+        fi
-      with:
+        echo "Version and commit tag match. Proceeding with the workflow."
-        python-version: "3.x"
 
     - name: Install pypa/build/setuptools/twine
       run: >-
@@ -36,9 +43,6 @@ jobs:
         build setuptools twine
         --user
 
-    - name: Prevent fallback onto setup.py
-      run: rm setup.py
-
     - name: Build a binary wheel and a source tarball
       run: python3 -m build
 

docs/source/conf.py

@@ -1,13 +1,17 @@
 # This conf.py is in large parts inspired by the oen used by stable-baselines 3
 
+import toml
 import datetime
 
 project = 'Fancy Gym'
 author = 'Fabian Otto, Onur Celik, Dominik Roth, Hongyi Zhou'
 copyright = f'2020-{datetime.date.today().year}, {author}'
 
-release = '0.2'  # The full version, including alpha/beta/rc tags
+pyproject_content = toml.load("../../pyproject.toml")
-version = '0.2'  # The short X.Y version
+proj_version = pyproject_content["project"]["version"]
+
+release = proj_version  # The full version, including alpha/beta/rc tags
+version = proj_version  # The short X.Y version
 
 extensions = [
     'myst_parser',
@@ -50,4 +54,4 @@ html_context = {
 }
 
 def setup(app):
-    app.add_css_file("style.css")
+    app.add_css_file("style.css")

fancy_gym/envs/mujoco/ant_jump/ant_jump.py

@@ -115,7 +115,6 @@ class AntJumpEnv(AntEnvCustomXML):
                          contact_force_range=contact_force_range,
                          reset_noise_scale=reset_noise_scale,
                          exclude_current_positions_from_observation=exclude_current_positions_from_observation, **kwargs)
-        self.render_active = False
 
     def step(self, action):
         self.current_step += 1
@@ -154,15 +153,8 @@ class AntJumpEnv(AntEnvCustomXML):
         }
         truncated = False
 
-        if self.render_active and self.render_mode=='human':
-            self.render()
-
         return obs, reward, terminated, truncated, info
 
-    def render(self):
-        self.render_active = True
-        return super().render()
-
     def _get_obs(self):
         return np.append(super()._get_obs(), self.goal)
 

fancy_gym/envs/mujoco/beerpong/beerpong.py

@@ -44,7 +44,6 @@ class BeerPongEnv(MujocoEnv, utils.EzPickle):
     }
 
     def __init__(self, **kwargs):
-        utils.EzPickle.__init__(self)
         self._steps = 0
         # Small Context -> Easier. Todo: Should we do different versions?
         # self.xml_path = os.path.join(os.path.dirname(os.path.abspath(__file__)), "assets", "beerpong_wo_cup.xml")
@@ -90,7 +89,7 @@ class BeerPongEnv(MujocoEnv, utils.EzPickle):
             observation_space=self.observation_space,
             **kwargs
         )
-        self.render_active = False
+        utils.EzPickle.__init__(self)
 
     @property
     def start_pos(self):
@@ -170,15 +169,8 @@ class BeerPongEnv(MujocoEnv, utils.EzPickle):
 
         truncated = False
 
-        if self.render_active and self.render_mode=='human':
-            self.render()
-
         return ob, reward, terminated, truncated, infos
 
-    def render(self):
-        self.render_active = True
-        return super().render()
-
     def _get_obs(self):
         theta = self.data.qpos.flat[:7].copy()
         theta_dot = self.data.qvel.flat[:7].copy()

fancy_gym/envs/mujoco/box_pushing/box_pushing_env.py

@@ -4,7 +4,6 @@ import numpy as np
 from gymnasium import utils, spaces
 from gymnasium.envs.mujoco import MujocoEnv
 from fancy_gym.envs.mujoco.box_pushing.box_pushing_utils import rot_to_quat, get_quaternion_error, rotation_distance
-from fancy_gym.envs.mujoco.box_pushing.box_pushing_utils import rot_to_quat, get_quaternion_error, rotation_distance
 from fancy_gym.envs.mujoco.box_pushing.box_pushing_utils import q_max, q_min, q_dot_max, q_torque_max
 from fancy_gym.envs.mujoco.box_pushing.box_pushing_utils import desired_rod_quat
 
@@ -61,7 +60,6 @@ class BoxPushingEnvBase(MujocoEnv, utils.EzPickle):
                            frame_skip=self.frame_skip,
                            observation_space=self.observation_space, **kwargs)
         self.action_space = spaces.Box(low=-1, high=1, shape=(7,))
-        self.render_active = False
 
     def step(self, action):
         action = 10 * np.clip(action, self.action_space.low, self.action_space.high)
@@ -110,15 +108,8 @@ class BoxPushingEnvBase(MujocoEnv, utils.EzPickle):
         terminated = episode_end and infos['is_success']
         truncated = episode_end and not infos['is_success']
 
-        if self.render_active and self.render_mode=='human':
-            self.render()
-
         return obs, reward, terminated, truncated, infos
 
-    def render(self):
-        self.render_active = True
-        return super().render()
-
     def reset_model(self):
         # rest box to initial position
         self.set_state(self.init_qpos_box_pushing, self.init_qvel_box_pushing)

fancy_gym/envs/mujoco/half_cheetah_jump/half_cheetah_jump.py

@@ -60,11 +60,7 @@ class HalfCheetahEnvCustomXML(HalfCheetahEnv):
             default_camera_config=DEFAULT_CAMERA_CONFIG,
             **kwargs,
         )
-        self.render_active = False
 
-    def render(self):
-        self.render_active = True
-        return super().render()
 
 class HalfCheetahJumpEnv(HalfCheetahEnvCustomXML):
     """
@@ -124,9 +120,6 @@ class HalfCheetahJumpEnv(HalfCheetahEnvCustomXML):
             'max_height': self.max_height
         }
 
-        if self.render_active and self.render_mode=='human':
-            self.render()
-
         return observation, reward, terminated, truncated, info
 
     def _get_obs(self):

fancy_gym/envs/mujoco/hopper_jump/hopper_jump.py

@@ -88,12 +88,6 @@ class HopperEnvCustomXML(HopperEnv):
             **kwargs,
         )
 
-        self.render_active = False
-
-    def render(self):
-        self.render_active = True
-        return super().render()
-
 
 class HopperJumpEnv(HopperEnvCustomXML):
     """
@@ -207,10 +201,6 @@ class HopperJumpEnv(HopperEnvCustomXML):
             healthy=self.is_healthy,
             contact_dist=self.contact_dist or 0
         )
-
-        if self.render_active and self.render_mode=='human':
-            self.render()
-
         return observation, reward, terminated, truncated, info
 
     def _get_obs(self):

fancy_gym/envs/mujoco/hopper_jump/hopper_jump_on_box.py

@@ -140,9 +140,6 @@ class HopperJumpOnBoxEnv(HopperEnvCustomXML):
 
         truncated = self.current_step >= self.max_episode_steps and not terminated
 
-        if self.render_active and self.render_mode=='human':
-            self.render()
-
         return observation, reward, terminated, truncated, info
 
     def _get_obs(self):

fancy_gym/envs/mujoco/hopper_throw/hopper_throw.py

@@ -61,8 +61,6 @@ class HopperThrowEnv(HopperEnvCustomXML):
                          exclude_current_positions_from_observation=exclude_current_positions_from_observation,
                          **kwargs)
 
-        self.render_active = False
-
     def step(self, action):
         self.current_step += 1
         self.do_simulation(action, self.frame_skip)
@@ -96,15 +94,8 @@ class HopperThrowEnv(HopperEnvCustomXML):
         }
         truncated = False
 
-        if self.render_active and self.render_mode=='human':
-            self.render()
-
         return observation, reward, terminated, truncated, info
 
-    def render(self):
-        self.render_active = True
-        return super().render()
-
     def _get_obs(self):
         return np.append(super()._get_obs(), self.goal)
 

fancy_gym/envs/mujoco/hopper_throw/hopper_throw_in_basket.py

@@ -68,7 +68,6 @@ class HopperThrowInBasketEnv(HopperEnvCustomXML):
                          reset_noise_scale=reset_noise_scale,
                          exclude_current_positions_from_observation=exclude_current_positions_from_observation,
                          **kwargs)
-        self.render_active = False
 
     def step(self, action):
 
@@ -119,15 +118,8 @@ class HopperThrowInBasketEnv(HopperEnvCustomXML):
         }
         truncated = False
 
-        if self.render_active and self.render_mode=='human':
-            self.render()
-
         return observation, reward, terminated, truncated, info
 
-    def render(self):
-        self.render_active = True
-        return super().render()
-
     def _get_obs(self):
         return np.append(super()._get_obs(), self.basket_x)
 

fancy_gym/envs/mujoco/reacher/reacher.py

@@ -47,8 +47,6 @@ class ReacherEnv(MujocoEnv, utils.EzPickle):
                            **kwargs
                            )
 
-        self.render_active = False
-
     def step(self, action):
         self._steps += 1
 
@@ -79,15 +77,8 @@ class ReacherEnv(MujocoEnv, utils.EzPickle):
             goal=self.goal if hasattr(self, "goal") else None
         )
 
-        if self.render_active and self.render_mode=='human':
-            self.render()
-
         return ob, reward, terminated, truncated, info
 
-    def render(self):
-        self.render_active = True
-        return super().render()
-
     def distance_reward(self):
         vec = self.get_body_com("fingertip") - self.get_body_com("target")
         return -self._reward_weight * np.linalg.norm(vec)

fancy_gym/envs/mujoco/table_tennis/table_tennis_env.py

@@ -71,8 +71,6 @@ class TableTennisEnv(MujocoEnv, utils.EzPickle):
                            observation_space=self.observation_space,
                            **kwargs)
 
-        self.render_active = False
-
         if ctxt_dim == 2:
             self.context_bounds = CONTEXT_BOUNDS_2DIMS
         elif ctxt_dim == 4:
@@ -160,15 +158,8 @@ class TableTennisEnv(MujocoEnv, utils.EzPickle):
 
         terminated, truncated = self._terminated, False
 
-        if self.render_active and self.render_mode=='human':
-            self.render()
-
         return self._get_obs(), reward, terminated, truncated, info
 
-    def render(self):
-        self.render_active = True
-        return super().render()
-
     def _contact_checker(self, id_1, id_2):
         for coni in range(0, self.data.ncon):
             con = self.data.contact[coni]

fancy_gym/envs/mujoco/walker_2d_jump/walker_2d_jump.py

@@ -79,8 +79,6 @@ class Walker2dEnvCustomXML(Walker2dEnv):
             **kwargs,
         )
 
-        self.render_active = False
-
 
 class Walker2dJumpEnv(Walker2dEnvCustomXML):
     """
@@ -147,15 +145,8 @@ class Walker2dJumpEnv(Walker2dEnvCustomXML):
         }
         truncated = False
 
-        if self.render_active and self.render_mode=='human':
-            self.render()
-
         return observation, reward, terminated, truncated, info
 
-    def render(self):
-        self.render_active = True
-        return super().render()
-
     def _get_obs(self):
         return np.append(super()._get_obs(), self.goal)
 

fancy_gym/examples/example_replanning_envs.py

@@ -3,14 +3,14 @@ import fancy_gym
 
 
 def example_run_replanning_env(env_name="fancy_ProDMP/BoxPushingDenseReplan-v0", seed=1, iterations=1, render=False):
-    env = gym.make(env_name, render_mode='human' if render else None)
+    env = gym.make(env_name)
     env.reset(seed=seed)
     for i in range(iterations):
         while True:
             ac = env.action_space.sample()
             obs, reward, terminated, truncated, info = env.step(ac)
             if render:
-                env.render()
+                env.render(mode="human")
             if terminated or truncated:
                 env.reset()
                 break
@@ -38,13 +38,13 @@ def example_custom_replanning_envs(seed=0, iteration=100, render=True):
                         'replanning_schedule': lambda pos, vel, obs, action, t: t % 25 == 0,
                         'condition_on_desired': True}
 
-    base_env = gym.make(base_env_id, render_mode='human' if render else None)
+    base_env = gym.make(base_env_id)
     env = fancy_gym.make_bb(env=base_env, wrappers=wrappers, black_box_kwargs=black_box_kwargs,
                             traj_gen_kwargs=trajectory_generator_kwargs, controller_kwargs=controller_kwargs,
                             phase_kwargs=phase_generator_kwargs, basis_kwargs=basis_generator_kwargs,
                             seed=seed)
     if render:
-        env.render()
+        env.render(mode="human")
 
     obs = env.reset()
 

fancy_gym/examples/examples_dmc.py

@@ -17,7 +17,7 @@ def example_dmc(env_id="dm_control/fish-swim", seed=1, iterations=1000, render=T
     Returns:
 
     """
-    env = gym.make(env_id, render_mode='human' if render else None)
+    env = gym.make(env_id)
     rewards = 0
     obs = env.reset(seed=seed)
     print("observation shape:", env.observation_space.shape)
@@ -26,7 +26,7 @@ def example_dmc(env_id="dm_control/fish-swim", seed=1, iterations=1000, render=T
     for i in range(iterations):
         ac = env.action_space.sample()
         if render:
-            env.render()
+            env.render(mode="human")
         obs, reward, terminated, truncated, info = env.step(ac)
         rewards += reward
 
@@ -84,7 +84,7 @@ def example_custom_dmc_and_mp(seed=1, iterations=1, render=True):
     # basis_generator_kwargs = {'basis_generator_type': 'rbf',
     #                           'num_basis': 5
     #                           }
-    base_env = gym.make(base_env_id, render_mode='human' if render else None)
+    base_env = gym.make(base_env_id)
     env = fancy_gym.make_bb(env=base_env, wrappers=wrappers, black_box_kwargs={},
                             traj_gen_kwargs=trajectory_generator_kwargs, controller_kwargs=controller_kwargs,
                             phase_kwargs=phase_generator_kwargs, basis_kwargs=basis_generator_kwargs,
@@ -96,7 +96,7 @@ def example_custom_dmc_and_mp(seed=1, iterations=1, render=True):
     # It is also possible to change them mode multiple times when
     # e.g. only every nth trajectory should be displayed.
     if render:
-        env.render()
+        env.render(mode="human")
 
     rewards = 0
     obs = env.reset()
@@ -115,7 +115,7 @@ def example_custom_dmc_and_mp(seed=1, iterations=1, render=True):
     env.close()
     del env
 
-def main(render = False):
+def main(render = True):
     # # Standard DMC Suite tasks
     example_dmc("dm_control/fish-swim", seed=10, iterations=1000, render=render)
     #

fancy_gym/examples/examples_general.py

@@ -21,7 +21,7 @@ def example_general(env_id="Pendulum-v1", seed=1, iterations=1000, render=True):
 
     """
 
-    env = gym.make(env_id, render_mode='human' if render else None)
+    env = gym.make(env_id)
     rewards = 0
     obs = env.reset(seed=seed)
     print("Observation shape: ", env.observation_space.shape)
@@ -85,7 +85,7 @@ def example_async(env_id="fancy/HoleReacher-v0", n_cpu=4, seed=int('533D', 16),
     # do not return values above threshold
     return *map(lambda v: np.stack(v)[:n_samples], buffer.values()),
 
-def main(render = False):
+def main(render = True):
     # Basic gym task
     example_general("Pendulum-v1", seed=10, iterations=200, render=render)
 

fancy_gym/examples/examples_metaworld.py

@@ -2,7 +2,7 @@ import gymnasium as gym
 import fancy_gym
 
 
-def example_meta(env_id="metaworld/button-press-v2", seed=1, iterations=1000, render=True):
+def example_meta(env_id="fish-swim", seed=1, iterations=1000, render=True):
     """
     Example for running a MetaWorld based env in the step based setting.
     The env_id has to be specified as `task_name-v2`. V1 versions are not supported and we always
@@ -18,7 +18,7 @@ def example_meta(env_id="metaworld/button-press-v2", seed=1, iterations=1000, re
     Returns:
 
     """
-    env = gym.make(env_id, render_mode='human' if render else None)
+    env = gym.make(env_id)
     rewards = 0
     obs = env.reset(seed=seed)
     print("observation shape:", env.observation_space.shape)
@@ -27,7 +27,9 @@ def example_meta(env_id="metaworld/button-press-v2", seed=1, iterations=1000, re
     for i in range(iterations):
         ac = env.action_space.sample()
         if render:
-            env.render()
+            # THIS NEEDS TO BE SET TO FALSE FOR NOW, BECAUSE THE INTERFACE FOR RENDERING IS DIFFERENT TO BASIC GYM
+            # TODO: Remove this, when Metaworld fixes its interface.
+            env.render(False)
         obs, reward, terminated, truncated, info = env.step(ac)
         rewards += reward
         if terminated or truncated:
@@ -79,7 +81,7 @@ def example_custom_meta_and_mp(seed=1, iterations=1, render=True):
     basis_generator_kwargs = {'basis_generator_type': 'rbf',
                               'num_basis': 5
                               }
-    base_env = gym.make(base_env_id, render_mode='human' if render else None)
+    base_env = gym.make(base_env_id)
     env = fancy_gym.make_bb(env=base_env, wrappers=wrappers, black_box_kwargs={},
                             traj_gen_kwargs=trajectory_generator_kwargs, controller_kwargs=controller_kwargs,
                             phase_kwargs=phase_generator_kwargs, basis_kwargs=basis_generator_kwargs,
@@ -91,7 +93,7 @@ def example_custom_meta_and_mp(seed=1, iterations=1, render=True):
     # It is also possible to change them mode multiple times when
     # e.g. only every nth trajectory should be displayed.
     if render:
-        env.render()
+        env.render(mode="human")
 
     rewards = 0
     obs = env.reset(seed=seed)

fancy_gym/examples/examples_open_ai.py

@@ -13,13 +13,15 @@ def example_mp(env_name, seed=1, render=True):
     Returns:
 
     """
-    env = gym.make(env_name, render_mode='human' if render else None)
+    env = gym.make(env_name)
 
     returns = 0
     obs = env.reset(seed=seed)
     # number of samples/full trajectories (multiple environment steps)
     for i in range(10):
         if render and i % 2 == 0:
+            env.render(mode="human")
+        else:
             env.render()
         ac = env.action_space.sample()
         obs, reward, terminated, truncated, info = env.step(ac)

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "fancy_gym"
-version = "0.1.4"
+version = "0.3.0"
 description = "Fancy Gym: Unifying interface for various RL benchmarks with support for Black Box approaches."
 readme = "README.md"
 authors = [
@@ -26,6 +26,7 @@ classifiers = [
 ]
 
 dependencies = [
+    "toml",
     "mp_pytorch<=0.1.3",
     "mujoco==2.3.3",
     "gymnasium[mujoco]>=0.26.0"

setup.py

@@ -1,5 +1,6 @@
 # We still provide a setup.py for backwards compatability.
 # But the pyproject.toml should be prefered.
+import toml
 import itertools
 from pathlib import Path
 from typing import List
@@ -8,6 +9,9 @@ from setuptools import setup, find_packages
 
 print('[!] You are currently installing/building fancy_gym via setup.py. This is only provided for backwards-compatability. Please use the pyproject.toml instead.')
 
+pyproject_content = toml.load("pyproject.toml")
+project_version = pyproject_content["project"]["version"]
+
 # Environment-specific dependencies for dmc and metaworld
 extras = {
     'dmc': ['shimmy[dm-control]', 'Shimmy==1.0.0'],
@@ -38,7 +42,7 @@ def find_package_data(extensions_to_include: List[str]) -> List[str]:
 setup(
     author='Fabian Otto, Onur Celik, Dominik Roth, Hongyi Zhou',
     name='fancy_gym',
-    version='0.1.0',
+    version=project_version,
     classifiers=[
         'Development Status :: 4 - Beta',
         'Intended Audience :: Science/Research',
@@ -55,6 +59,7 @@ setup(
     ],
     extras_require=extras,
     install_requires=[
+        'toml',
         'mp_pytorch<=0.1.3',
         'mujoco==2.3.3',
         'gymnasium[mujoco]>=0.26.0'