continuous integration

.github/workflows/update_stompy_s3.yml

@@ -0,0 +1,61 @@
+name: Update Stompy S3 Model
+
+on:
+  release:
+    types: [created]
+  schedule:
+    - cron: "0 0 * * *"
+  workflow_dispatch:
+
+permissions:
+  contents: read
+  id-token: write
+
+concurrency:
+  group: "stompy-s3-model"
+  cancel-in-progress: true
+
+jobs:
+  publish:
+    timeout-minutes: 10
+    name: Update Stompy S3 Model
+
+    # We don't need to run on all platforms since this package is
+    # platform-agnostic. The output wheel is something like
+    # "monotonic_attention-<version>-py3-none-any.whl".
+    runs-on: ubuntu-latest
+
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v3
+
+      - name: Set up Python
+        uses: actions/setup-python@v4
+        with:
+          python-version: "3.10"
+
+      - name: Install dependencies
+        run: |
+          pip install kscale-onshape-library '.'
+
+      - name: Build package
+        env:
+          ONSHAPE_ACCESS_KEY: ${{ secrets.ONSHAPE_ACCESS_KEY }}
+          ONSHAPE_SECRET_KEY: ${{ secrets.ONSHAPE_SECRET_KEY }}
+          ONSHAPE_API: ${{ secrets.ONSHAPE_API }}
+        run: python -m sim.scripts.update_stompy_s3
+
+      - name: Configure AWS credentials
+        uses: aws-actions/configure-aws-credentials@v1
+        with:
+          aws-access-key-id: ${{ secrets.AWS_ACCESS_KEY_ID }}
+          aws-secret-access-key: ${{ secrets.AWS_SECRET_ACCESS_KEY }}
+          aws-region: us-east-1
+
+      - name: Upload to S3
+        env:
+          AWS_S3_BUCKET: ${{ secrets.AWS_S3_BUCKET }}
+        run: |
+          for file in stompy/*.tar.gz; do
+            aws s3 cp "$file" s3://${AWS_S3_BUCKET}/$(basename "$file")
+          done

.gitignore

@@ -25,4 +25,4 @@ out*/
 # Training artifacts
 wandb/
 runs/
-stompy/
+stompy/

sim/humanoid_gym/play.py

@@ -1,6 +1,7 @@
 """Script to replay a trained policy in the environment."""
 
 import argparse
+import logging
 import os
 from datetime import datetime
 
@@ -9,11 +10,17 @@
 from isaacgym import gymapi
 from tqdm import tqdm
 
+from sim.logging import configure_logging
+
+logger = logging.getLogger(__name__)
+
 from sim.env import run_dir
 from sim.humanoid_gym.envs import *  # noqa: F403
 
 
 def play(args: argparse.Namespace) -> None:
+    configure_logging()
+
     env_cfg, train_cfg = task_registry.get_cfgs(name=args.task)
     # override some parameters for testing
     env_cfg.env.num_envs = min(env_cfg.env.num_envs, 1)
@@ -31,7 +38,7 @@ def play(args: argparse.Namespace) -> None:
     env_cfg.noise.noise_level = 0.5
 
     train_cfg.seed = 123145
-    print("train_cfg.runner_class_name:", train_cfg.runner_class_name)
+    logger.info("train_cfg.runner_class_name: %s", train_cfg.runner_class_name)
 
     # prepare environment
     env, _ = task_registry.make_env(name=args.task, args=args, env_cfg=env_cfg)
@@ -44,7 +51,7 @@ def play(args: argparse.Namespace) -> None:
     ppo_runner, train_cfg = task_registry.make_alg_runner(env=env, name=args.task, args=args, train_cfg=train_cfg)
     policy = ppo_runner.get_inference_policy(device=env.device)
 
-    logger = Logger(env.dt)
+    env_logger = Logger(env.dt)
     robot_index = 0  # which robot is used for logging
     joint_index = 1  # which joint is used for logging
     stop_state_log = 1200  # number of steps before plotting states
@@ -58,8 +65,8 @@ def play(args: argparse.Namespace) -> None:
         camera_rotation = gymapi.Quat.from_axis_angle(gymapi.Vec3(-0.3, 0.2, 1), np.deg2rad(135))
         actor_handle = env.gym.get_actor_handle(env.envs[0], 0)
         body_handle = env.gym.get_actor_rigid_body_handle(env.envs[0], actor_handle, 0)
-        print(body_handle)
-        print(actor_handle)
+        logger.info("body_handle: %s", body_handle)
+        logger.info("actor_handle: %s", actor_handle)
         env.gym.attach_camera_to_body(
             h1, env.envs[0], body_handle, gymapi.Transform(camera_offset, camera_rotation), gymapi.FOLLOW_POSITION
         )
@@ -99,7 +106,7 @@ def play(args: argparse.Namespace) -> None:
 
             video.write(img[..., :3])  # Write only the RGB channels
 
-        logger.log_states(
+        env_logger.log_states(
             {
                 "dof_pos_target": actions[robot_index, joint_index].item() * env.cfg.control.action_scale,
                 "dof_pos": env.dof_pos[robot_index, joint_index].item(),
@@ -119,10 +126,10 @@ def play(args: argparse.Namespace) -> None:
         if infos["episode"]:
             num_episodes = env.reset_buf.sum().item()
             if num_episodes > 0:
-                logger.log_rewards(infos["episode"], num_episodes)
+                env_logger.log_rewards(infos["episode"], num_episodes)
 
-    logger.print_rewards()
-    logger.plot_states()
+    env_logger.print_rewards()
+    env_logger.plot_states()
 
     if RENDER:
         video.release()

sim/scripts/print_joints.py

@@ -27,7 +27,7 @@ def main() -> None:
 
     # Makes a "tree" of the joints using common prefixes.
     joint_names.sort()
-    joint_tree = {}
+    joint_tree: Dict = {}
     for joint_name in joint_names:
         parts = joint_name.split("_")
         current_tree = joint_tree

sim/scripts/update_stompy_s3.py

@@ -0,0 +1,83 @@
+"""Updates the Stompy model."""
+
+import tarfile
+from pathlib import Path
+
+from kol.logging import configure_logging
+from kol.onshape.converter import Converter
+
+STOMPY_MODEL = (
+    "https://cad.onshape.com/documents/71f793a23ab7562fb9dec82d/"
+    "w/6160a4f44eb6113d3fa116cd/e/1a95e260677a2d2d5a3b1eb3"
+)
+
+SUFFIX_TO_JOINT_EFFORT = {
+    "dof_x4_h": 1.5,
+    "dof_x4": 1.5,
+    "dof_x6": 3,
+    "dof_x8": 6,
+    "dof_x10": 12,
+    "knee_revolute": 13.9,
+    "ankle_revolute": 6,
+}
+
+
+def main() -> None:
+    configure_logging()
+
+    output_dir = Path("stompy")
+
+    # Gets the latest STL URDF.
+    converter = Converter(
+        document_url=STOMPY_MODEL,
+        output_dir=output_dir / "latest_stl_urdf",
+        suffix_to_joint_effort=list(SUFFIX_TO_JOINT_EFFORT.items()),
+        disable_mimics=True,
+        mesh_ext="stl",
+    )
+    converter.save_urdf()
+    latest_stl_urdf_path = converter.output_dir
+
+    # Manually builds the tarball.
+    with tarfile.open(output_dir / "latest_stl_urdf.tar.gz", "w:gz") as tar:
+        for suffix in (".urdf", ".stl"):
+            for file in latest_stl_urdf_path.rglob(f"**/*{suffix}"):
+                tar.add(file, arcname=file.relative_to(latest_stl_urdf_path))
+
+    # Gets the latest OBJ URDF.
+    converter = Converter(
+        document_url=STOMPY_MODEL,
+        output_dir=output_dir / "latest_obj_urdf",
+        suffix_to_joint_effort=list(SUFFIX_TO_JOINT_EFFORT.items()),
+        disable_mimics=True,
+        mesh_ext="obj",
+    )
+    converter.save_urdf()
+    latest_obj_urdf_path = converter.output_dir
+
+    # Manually builds the tarball.
+    with tarfile.open(output_dir / "latest_obj_urdf.tar.gz", "w:gz") as tar:
+        for suffix in (".urdf", ".obj"):
+            for file in latest_obj_urdf_path.rglob(f"**/*{suffix}"):
+                tar.add(file, arcname=file.relative_to(latest_obj_urdf_path))
+
+    # Gets the latest MJCF.
+    converter = Converter(
+        document_url=STOMPY_MODEL,
+        output_dir=output_dir / "latest_mjcf",
+        suffix_to_joint_effort=list(SUFFIX_TO_JOINT_EFFORT.items()),
+        disable_mimics=True,
+    )
+    converter.save_mjcf()
+    latest_mjcf_path = converter.output_dir
+
+    # Manually builds the tarball.
+    with tarfile.open(output_dir / "latest_mjcf.tar.gz", "w:gz") as tar:
+        for suffix in (".xml", ".stl"):
+            for file in latest_mjcf_path.rglob(f"**/*{suffix}"):
+                tar.add(file, arcname=file.relative_to(latest_mjcf_path))
+
+
+if __name__ == "__main__":
+    # python -m sim.scripts.update_stompy
+    main()