Template repository to create new agents with custom spines for Upkie wheeled bipeds.

If you don't need a custom spine, you can implement your agent in Python directly, as done for instance in the MPC balancer.

1. Create a new repository from this template
2. Replace LICENSE with the license of your choice (the default one is Apache-2.0)
3. Search for the string "TODO" and update files accordingly:
   • The first instance is in Makefile
   • The second instance is in WORKSPACE
4. Start listing your dependencies in environment.yaml
5. Implement your agent starting from agent/main.py
6. Customize your C++ spines in the spines directory

• Install Python packages to a conda environment: conda env create -f environment.yaml -n your_agent_name
• Activate conda environment: conda activate your_agent_name
• Build and run the custom simulation spine: make run_bullet_spine
• Build the custom pi3hat spine: make build
• Upload the working directory to the robot: make upload
• Run the pi3hat spine on the robot: make run_pi3hat_spine
• Run your agent on the robot: python agent/main.py