Tooling

BLE tooling for reading the Triki token (nRF52810 + LSM6DSL). The scripts live in tools/ and run with the project's Python venv (tools/.venv/, gitignored).

Installed

Tool	Version	Purpose
bleak (venv)	3.0.2	Cross-platform BLE client (CoreBluetooth on macOS)

Read the device

# 1. find the token over the air (name/address/RSSI)
./.venv/bin/python ble_scan.py 8

# 2. connect and dump the GATT map
./.venv/bin/python ble_dump.py "<name-or-address-from-step-1>"

ble_dump.py enumerates the device's GATT services/characteristics, reads the readable ones, and subscribes to notifications. The GATT map and protocol are on the BLE protocol page.

Decode the IMU stream

The headline tool — start the IMU stream and decode the 14-byte motion frames into scaled gyro (deg/s) + accel (g):

./.venv/bin/python ble_imu_stream.py "Triki" 15

Scripts

Run each with the venv interpreter. <target> is the token's BLE name or address (from ble_scan.py):

• ble_nus.py <target> — interactive console (write RX, watch TX).
• ble_opcode_sweep.py <target> — map the command interface by single-byte opcode.
• ble_rx_probe.py <target> — probe RX with multi-byte payloads, classify each by reaction.
• ble_enable_scan.py <target> — sweep for an "enable / active mode" command.
• ble_probe_ctrl.py <target> / ble_ctrl_toggle.py <target> — exercise / toggle the green-LED ctrl register.
• ble_listen.py <target> — optionally set ctrl, then listen.
• ble_poll.py <target> <opcodes> / ble_imu_test.py <target> — correlate response columns with motion.

Tip: the token sleeps within seconds when idle — keep it moving so it stays awake and reachable while a script connects.