digity 0.1.5

Python SDK for the Digity sensorized exohand

digity SDK

Python SDK for the Digity sensorized exohand — plug in the glove, stream live sensor data directly into your Python code in three lines.

The Digity exohand is a sensorized glove that captures real-time hand motion data: joint angles for every finger, 6-axis IMU (accelerometer + gyroscope), and 6-channel capacitive touch per node. This SDK handles the USB connection, binary protocol decoding, and threading — so you get clean, typed Python objects at ~50 Hz without any low-level plumbing.

from digity import GloveStream, AnglesSensor

with GloveStream() as stream:
    for frame in stream:
        for sensor in frame.sensors:
            if isinstance(sensor, AnglesSensor):
                print(sensor.finger, sensor.samples[-1].angles_deg)

Use cases: motion capture, rehabilitation, robotics control, gesture recognition, HCI research, data collection.

---

Requirements

• Python 3.9 or later
• The Digity exohand connected via USB

---

Installation

# SDK only
pip install digity

# SDK + real-time dashboard
pip install digity[viz]

---

Real-time dashboard

digity[viz] includes a full web-based dashboard that visualises live hand data, records sessions, and relays data from a remote agent — all locally, no internet required.

digity-viz              # opens the dashboard as a desktop window
digity-viz --port COM3  # explicit serial port
digity-viz --browser    # open in system browser instead

Or launch from Python:

import digity.viz
digity.viz.start()

The dashboard runs on http://localhost:5001/chiro/ and opens automatically. No login needed — it is intended for single-user local use.

---

Quick start

Plug in the glove, then run:

from digity import GloveStream, AnglesSensor, ImuSensor, TouchSensor

with GloveStream() as stream:
    for frame in stream:
        print(f"side={frame.side}  node={frame.node_id}  seq={frame.seq}")

        for sensor in frame.sensors:

            if isinstance(sensor, AnglesSensor):
                angles = sensor.samples[-1].angles_deg
                print(f"  finger {sensor.finger} angles: {angles}")

            elif isinstance(sensor, ImuSensor):
                s = sensor.samples[-1]
                print(f"  finger {sensor.finger} acc={s.acc}  gyro={s.gyro}")

            elif isinstance(sensor, TouchSensor):
                print(f"  finger {sensor.finger} touch={sensor.channels}")

The with block handles connection and cleanup automatically. Press Ctrl+C to stop.

---

Specifying the USB port

The SDK auto-detects the glove port on all platforms. If auto-detection fails, pass the port explicitly:

GloveStream(port="/dev/ttyUSB0")   # Linux
GloveStream(port="/dev/tty.usbserial-0001")  # macOS
GloveStream(port="COM3")           # Windows

---

Data types

GloveFrame

Every iteration of GloveStream yields one GloveFrame:

Field	Type	Description
ts	float	Host timestamp (seconds since epoch) when the frame arrived
side	str	"right" or "left"
group	str	"hand" or "arm"
node_id	int	PCB node number on the glove
seq	int	Packet counter 0–65535, wraps around
sensors	list[Sensor]	List of sensor readings in this frame

AnglesSensor

Joint angles for one finger node.

Field	Type	Description
finger	int	Finger index (0=thumb, 1=index, …, 4=pinky)
com	int	Communication line index
samples	list[AnglesSample]	One or more timestamped angle readings

AnglesSample

Field	Type	Description
ts_us	int	Sensor timestamp in microseconds
angles_deg	list[float]	Joint angles in degrees (5 values for hand group)

if isinstance(sensor, AnglesSensor):
    latest = sensor.samples[-1]
    print(latest.angles_deg)   # e.g. [12.3, 45.1, 30.0, 5.5, 2.0]

ImuSensor

6-axis IMU (accelerometer + gyroscope) for one finger node.

Field	Type	Description
finger	int	Finger index
com	int	Communication line index
samples	list[ImuSample]	One or more timestamped IMU readings

ImuSample

Field	Type	Description
ts_us	int	Sensor timestamp in microseconds
acc	tuple[int, int, int]	Accelerometer x/y/z, raw i16 counts
gyro	tuple[int, int, int]	Gyroscope x/y/z, raw i16 counts

if isinstance(sensor, ImuSensor):
    s = sensor.samples[-1]
    print(s.acc)    # e.g. (312, -128, 16384)
    print(s.gyro)   # e.g. (5, -12, 3)

TouchSensor

6-channel capacitive touch for one finger node.

Field	Type	Description
finger	int	Finger index
com	int	Communication line index
ts_us	int	Sensor timestamp in microseconds
channels	list[float]	6 values normalised 0.0–1.0
channels_raw	list[int]	6 raw ADC counts 0–4095

if isinstance(sensor, TouchSensor):
    print(sensor.channels)      # [0.0, 0.82, 0.0, 0.0, 0.41, 0.0]
    print(sensor.channels_raw)  # [0, 3358, 0, 0, 1680, 0]

Type annotation helper

from digity import Sensor  # Union[AnglesSensor, ImuSensor, TouchSensor]

def process(sensor: Sensor) -> None:
    if isinstance(sensor, AnglesSensor):
        ...

---

Common patterns

Record data to a file

import csv
from digity import GloveStream, AnglesSensor

with GloveStream() as stream, open("recording.csv", "w", newline="") as f:
    writer = csv.writer(f)
    writer.writerow(["ts", "finger", "a0", "a1", "a2", "a3", "a4"])

    for frame in stream:
        for sensor in frame.sensors:
            if isinstance(sensor, AnglesSensor):
                a = sensor.samples[-1].angles_deg
                writer.writerow([frame.ts, sensor.finger] + a)

Detect a closed fist

from digity import GloveStream, AnglesSensor

with GloveStream() as stream:
    finger_angles = {}

    for frame in stream:
        for sensor in frame.sensors:
            if isinstance(sensor, AnglesSensor):
                finger_angles[sensor.finger] = sensor.samples[-1].angles_deg

        if len(finger_angles) == 5:
            avg_flex = [sum(finger_angles[f][1] for f in range(1, 5)) / 4]
            if avg_flex[0] > 60:
                print("Fist detected!")

Stop the stream from another thread

import threading
from digity import GloveStream

stream = GloveStream()
stream.connect()

def stop_after(seconds):
    import time; time.sleep(seconds)
    stream.disconnect()

threading.Thread(target=stop_after, args=(10,), daemon=True).start()

for frame in stream:
    print(frame.seq)

---

Remote mode (ZMQ)

If the glove is connected to a remote machine running glove-core, you can receive frames over the network:

GloveStream(host="192.168.1.10")          # default port 5555
GloveStream(host="192.168.1.10", zmq_port=5556)

No USB connection is needed on the client machine in this mode.

---

Error handling

from digity import GloveStream, GloveNotFoundError

try:
    with GloveStream() as stream:
        for frame in stream:
            ...
except GloveNotFoundError:
    print("Glove not found — check the USB cable")
except KeyboardInterrupt:
    pass

GloveNotFoundError is raised at connection time if the glove port cannot be found.

---

Connection speed

The glove streams at approximately 50 Hz. Each call to for frame in stream blocks until the next frame arrives (up to 1 second timeout before checking for errors).

The SDK uses a background thread and an internal queue so that your processing code never blocks the serial buffer. If your code is slower than 50 Hz, frames are dropped to keep the stream live rather than accumulating memory.

---

License

© Digity. All rights reserved.