THALis: streaming and hand-simulation separation

hand sim first with predictable inputs, no EEG involved. if a joint does something weird i need to know which side broke; small command replay tool might be useful later

archived reorganization: b249471 moves acquisition into src/unicornlsl/stream.py, introduces src/handsim/sim.py, and adds Shadow Hand E MuJoCo assets under data/mujoco_models/. The assets include collision/visual meshes, URDF descriptions and XML environment files.

MuJoCo hand model and launch code

Shadow Robot Gazebo installation guide

interface question: the acquisition layer produces timed EEG samples; the hand simulator consumes control values/poses. Keep those interfaces independent so the simulator can be tested with deterministic inputs before coupling it to a decoder.

• path history: data/mujoco_models/ later becomes data/mujoco/, and simulation code is relocated again. Use a commit-specific path when returning to this version; a path in the final archive may not exist in the initial tree.

• review questions: what joints/actuators are commanded, what units/ranges are expected, and how are limits handled? The presence of a dexterous hand model is not evidence that all its degrees of freedom were decoded from EEG.

• archived code references: b249471.