Initial commit

Author: luator

pyproject.toml

@@ -0,0 +1,3 @@
+[build-system]
+requires = ["setuptools>=46.4.0", "wheel"]
+build-backend = "setuptools.build_meta"

rrc2022/__init__.py

@@ -0,0 +1 @@
+__version__ = "0.1.0"

rrc2022/example.py

@@ -0,0 +1,26 @@
+"""Example policy for Real Robot Challenge 2022"""
+import numpy as np
+
+
+class MyPolicy:
+    def __init__(self, steps=1000):
+        position_up = np.array([0.5, 1.2, -2.4] * 3)
+        position_down = np.array([-0.08, 0.84, -1.2] * 3)
+        trajectory_range = position_up - position_down
+        # steps = 1000
+        step_size = trajectory_range / steps
+
+        self.trajectory = [position_down + i * step_size for i in range(steps)]
+        self.i = 0
+
+    def get_action(self, observation):
+        if self.i >= len(self.trajectory):
+            # reverse trajectory
+            self.trajectory = self.trajectory[::-1]
+            self.i = 0
+
+        target_position = self.trajectory[self.i]
+        self.i += 1
+
+        # TODO: should be changed to return torques instead of positions
+        return target_position

setup.cfg

@@ -0,0 +1,8 @@
+[metadata]
+name = rrc2022
+version = attr: rrc2022.__version__
+
+[options]
+packages = rrc2022
+install_requires =
+    numpy

trifinger.toml

@@ -0,0 +1,2 @@
+[policy]
+class = "rrc2022.example.MyPolicy"