neuracore 1.5.2

Neuracore Client Library

Neuracore Python Client

Neuracore is a powerful robotics and machine learning client library for seamless robot data collection, model deployment, and interaction.

Features

• Easy robot initialization and connection
• Streaming data logging
• Model endpoint management
• Local and remote model support
• Flexible dataset creation

Installation

pip install neuracore

Quick Start

Ensure you have an account at neuracore.app

Authentication

import neuracore as nc

# This will save your API key locally
nc.login()

Robot Connection

# Connect to a robot
nc.connect_robot(
    robot_name="MyRobot", 
    urdf_path="/path/to/robot.urdf"
)

You can also upload MuJoCo MJCF rather than URDF. For that, ensure you install extra dependencies: pip install neuracore[mjcf].

nc.connect_robot(
    robot_name="MyRobot", 
    mjcf_path="/path/to/robot.xml"
)

Data Logging

# Log joint positions
nc.log_joint_positions({
    'joint1': 0.5, 
    'joint2': -0.3
})

# Log RGB camera image
nc.log_rgb("top_camera", image_array)

Documentation

• Creating Custom Algorithms
• Limitations

Development

To set up for development:

git clone https://github.com/neuraco/neuracore
cd neuracore
pip install -e .[dev]

Environment Variables

A few environment variables effect how this library operates, they are generally prefixed with NEURACORE_ and are case insensitive

Variable	Function	Valid Values	Default Value
NEURACORE_REMOTE_RECORDING_TRIGGER_ENABLED	Allows you to disable other machines starting a recording when logging. this does not affect the live data	true/false	true
NEURACORE_LIVE_DATA_ENABLED	Allows you to disable the streaming of data for live visualizations from this node. This does not affect recording.	true/false	true
NEURACORE_API_URL	The base url used to contact the neuracore platform.	A url e.g. https://api.neuracore.app/api	https://api.neuracore.app/api

Testing

export NEURACORE_API_URL=http://localhost:8000/api
pytest tests/

Contributing

Contributions are welcome!