Dynamics + Differentiable

.github/workflows/docker.yml

@@ -0,0 +1,43 @@
+name: Build Docker Image
+
+on: [push]
+
+jobs:
+  docker-build:
+    runs-on: ubuntu-latest
+    env:
+      IMAGE_NAME: ghcr.io/${{ github.repository_owner }}/posggym:latest
+
+
+    permissions:
+      contents: read
+      packages: write
+
+    steps:
+      - name: Checkout Repository
+        uses: actions/checkout@v3
+        with:
+          submodules: recursive
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
+
+      - name: Log in to GitHub Container Registry
+        uses: docker/login-action@v3
+        with:
+          registry: ghcr.io
+          username: ${{ github.actor }}
+          password: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Normalize tag name
+        run: echo "IMAGE_TAG=$(echo $IMAGE_NAME | tr '[:upper:]' '[:lower:]')" >> $GITHUB_ENV
+
+      - name: Build and Push Docker Image
+        uses: docker/build-push-action@v5
+        with:
+          context: .
+          file: Dockerfile
+          push: true
+          tags: ${{ env.IMAGE_TAG }}
+          cache-from: type=registry,ref=${{ env.IMAGE_TAG }}
+          cache-to: type=registry,ref=${{ env.IMAGE_TAG }},mode=max

.github/workflows/test.yml

@@ -5,14 +5,24 @@ on: [push]
 jobs:
   test:
     runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        python-version: ["3.10", "3.11", "3.12"]
 
     steps:
       - name: Checkout code
         uses: actions/checkout@v2
 
+      - name: Set up Python ${{ matrix.python-version }}
+        uses: actions/setup-python@v4
+        with:
+          python-version: ${{ matrix.python-version }}
 
-      - name: pip install
-        run: pip install --upgrade pip && pip install --user -e .[all] && pip install --user -e .[testing]
+      - name: Install dependencies
+        run: |
+          pip install --upgrade pip
+          pip install --user -e .[all]
+          pip install --user -e .[testing]
 
       - name: Run tests
-        run : pytest
+        run: pytest

.gitignore

@@ -140,3 +140,4 @@ dmypy.json
 
 # Ruff linter
 .ruff_cache/
+*.pickle

.pre-commit-config.yaml

@@ -30,9 +30,12 @@ repos:
       - id: black
   - repo: https://github.com/charliermarsh/ruff-pre-commit
     # Ruff version.
-    rev: 'v0.0.254'
+    rev: 'v0.4.10'
     hooks:
       - id: ruff
+        args:
+          - --fix
+          - --unsafe-fixes
   - repo: local
     hooks:
       - id: pyright

Dockerfile

@@ -0,0 +1,12 @@
+FROM pytorch/pytorch:2.2.2-cuda11.8-cudnn8-runtime
+
+WORKDIR /app
+
+COPY pyproject.toml ./
+COPY setup.py ./
+COPY ./posggym/__init__.py /app/posggym/__init__.py
+
+RUN pip install -e .[all]
+
+
+COPY . .

docs/conf.py

@@ -7,6 +7,7 @@
 # https://www.sphinx-doc.org/en/master/usage/configuration.html#project-information
 import posggym
 
+
 project = "POSGGym"
 copyright = "2023, Jonathon Schwartz"
 author = "Jonathon Schwartz"

docs/scripts/gen_agent_gifs.py

@@ -9,7 +9,7 @@
 import re
 from pathlib import Path
 from pprint import pprint
-from typing import Any, Dict, List
+from typing import Any
 
 import posggym
 import posggym.agents as pga
@@ -27,7 +27,7 @@
 
 def gen_agent_gif(
     env_id: str,
-    policy_ids: List[str],
+    policy_ids: list[str],
     ignore_existing: bool = False,
     length: int = 300,
     custom_env: bool = False,
@@ -43,7 +43,7 @@ def gen_agent_gif(
     for policy_id in policy_ids:
         try:
             pi_spec = pga.spec(policy_id)
-        except posggym.error.NameNotFound as e:
+        except posggym.error.NameNotFoundError as e:
             if "/" not in policy_id:
                 # try prepending env id
                 policy_id = f"{env_id}/{policy_id}"
@@ -65,8 +65,6 @@ def gen_agent_gif(
     env = posggym.make(
         env_id, disable_env_checker=True, render_mode="rgb_array", **env_args
     )
-    # env = posggym.wrappers.RescaleObservations(env, min_obs=-1.0, max_obs=1.0)
-    # env = posggym.wrappers.RescaleActions(env, min_action=-1.0, max_action=1.0)
 
     policies = {}
     for idx, spec in enumerate(policy_specs):
@@ -107,7 +105,7 @@ def gen_agent_gif(
         for _ in range(repeat):
             frames.append(Image.fromarray(frame))
 
-        actions: Dict[str, Any] = {}
+        actions: dict[str, Any] = {}
         for i in env.agents:
             if policies[i].observes_state:
                 actions[i] = policies[i].step(env.state)

docs/scripts/gen_agent_mds.py

@@ -5,7 +5,6 @@
 
 """
 import re
-from typing import Dict, List
 from pathlib import Path
 
 import posggym
@@ -22,7 +21,7 @@
 
 all_agents = list(pga.registry.values())
 # env_type -> env_name -> [PolicySpec]
-filtered_agents_by_env_type: Dict[str, Dict[str, List[PolicySpec]]] = {}
+filtered_agents_by_env_type: dict[str, dict[str, list[PolicySpec]]] = {}
 
 # Obtain filtered list
 for pi_spec in tqdm(all_agents):
@@ -100,7 +99,7 @@
         else:
             info = (
                 "These policies are for the "
-                + f"<a href='../../../environments/{env_type}/{snake_env_name}'>"
+                f"<a href='../../../environments/{env_type}/{snake_env_name}'>"
                 f"{title_env_name} environment</a>. Read environment page for detailed "
                 "information about the environment."
             )
@@ -124,7 +123,7 @@
         env_args_ids.sort()
 
         if None in filtered_agents_by_env_args_id:
-            env_args_ids = [None] + env_args_ids
+            env_args_ids = [None, *env_args_ids]
 
         for env_args_id in env_args_ids:
             policy_specs = filtered_agents_by_env_args_id[env_args_id]

docs/scripts/gen_env_mds.py

@@ -9,20 +9,21 @@
 
 import re
 from functools import reduce
-from typing import Dict, List
 from pathlib import Path
-from tqdm import tqdm
-from utils import kill_strs, trim
 
 import posggym
 from posggym.envs.registration import EnvSpec
+from tqdm import tqdm
+
+from utils import kill_strs, trim
+
 
 pattern = re.compile(r"(?<!^)(?=[A-Z])")
 
 posggym.logger.set_level(posggym.logger.DISABLED)
 
 all_envs = list(posggym.envs.registry.values())
-filtered_envs_by_type: Dict[str, Dict[str, EnvSpec]] = {}
+filtered_envs_by_type: dict[str, dict[str, EnvSpec]] = {}
 
 # Obtain filtered list
 for env_spec in tqdm(all_envs):
@@ -47,7 +48,7 @@
         split = str(type(env.unwrapped)).split(".")
         env_name = split[3]
 
-        if env_type not in filtered_envs_by_type.keys():
+        if env_type not in filtered_envs_by_type:
             filtered_envs_by_type[env_type] = {}
         # only store new entries and higher versions
         if env_name not in filtered_envs_by_type[env_type] or (
@@ -60,7 +61,7 @@
         print(e)
 
 # Sort
-filtered_envs: List = list(
+filtered_envs: list = list(
     reduce(
         lambda s, x: s + x,  # type: ignore
         (
@@ -120,15 +121,15 @@
         if "rgb_array" in env.metadata["render_modes"]:
             gif = (
                 "```{figure}"
-                + f" ../../_static/videos/{env_type}/{snake_env_name}.gif"
-                + f"\n:width: 200px\n:name: {snake_env_name}\n```"
+                f" ../../_static/videos/{env_type}/{snake_env_name}.gif"
+                f"\n:width: 200px\n:name: {snake_env_name}\n```"
             )
         else:
             gif = ""
         info = (
             "This environment is part of the "
-            + f"<a href='..'>{env_type_title} environments</a>. "
-            + "Please read that page first for general information."
+            f"<a href='..'>{env_type_title} environments</a>. "
+            "Please read that page first for general information."
         )
 
         act_spaces_str = str(env.action_spaces)
@@ -144,32 +145,17 @@
         env_table += f"| Symmetric | {env.is_symmetric} |\n"
 
         # if env.observation_space.shape:
-        #     env_table += f"| Observation Shape | {env.observation_space.shape} |\n"
 
         #     if hasattr(env.observation_space, "high"):
-        #         high = env.observation_space.high
 
         #         if hasattr(high, "shape"):
         #             if len(high.shape) == 3:
-        #                 high = high[0][0][0]
         #         if env_type == "mujoco":
-        #             high = high[0]
-        #         high = np.round(high, 2)
-        #         high = str(high).replace("\n", " ")
-        #         env_table += f"| Observation High | {high} |\n"
 
         #     if hasattr(env.observation_space, "low"):
-        #         low = env.observation_space.low
         #         if hasattr(low, "shape"):
         #             if len(low.shape) == 3:
-        #                 low = low[0][0][0]
         #         if env_type == "mujoco":
-        #             low = low[0]
-        #         low = np.round(low, 2)
-        #         low = str(low).replace("\n", " ")
-        #         env_table += f"| Observation Low | {low} |\n"
-        # else:
-        #     env_table += f"| Observation Space | {env.observation_space} |\n"
 
         env_table += f'| Import | `posggym.make("{env_spec.id}")` |\n'
 

docs/scripts/gen_envs_display.py

@@ -7,6 +7,7 @@
 import sys
 from pathlib import Path
 
+
 DOCS_DIR = Path(__file__).resolve().parent.parent
 
 all_envs = [
@@ -100,7 +101,7 @@ def generate_page(env, limit=-1, base_path=""):
         type_arg = sys.argv[1]
 
     for env in all_envs:
-        if type_arg == env["id"] or type_arg == "":
+        if type_arg in {env["id"], ""}:
             type_dict_arr.append(env)
 
     for type_dict in type_dict_arr:
@@ -127,7 +128,7 @@ def generate_page(env, limit=-1, base_path=""):
                 env_name = " ".join(type_id.split("_")).title()
                 fp.write(
                     f"# Complete List - {env_name}\n\n"
-                    + "```{raw} html\n:file: complete_list.html\n```"
+                    "```{raw} html\n:file: complete_list.html\n```"
                 )
         else:
             page = generate_page(type_dict)

docs/scripts/gen_gifs.py

@@ -71,7 +71,7 @@ def gen_gif(
             repeat = (
                 int(60 / env.metadata["render_fps"]) if env_type == "classic" else 1
             )
-            for i in range(repeat):
+            for _i in range(repeat):
                 frames.append(Image.fromarray(frame))
             action = {i: env.action_spaces[i].sample() for i in env.agents}
             _, _, _, _, done, _ = env.step(action)