mingruixia/brainnet-viewer repository overview

⁠BrainNet Viewer

BrainNet Viewer visualizes brain surfaces, networks and volumetric maps through a browser. The application runs in a background Docker container; MATLAB and a desktop Python installation are not required on the browsing computer.

⁠Public beta

Use mingruixia/brainnet-viewer:2.0.0-beta.1 for the public beta. Its Python package version is 2.0.0b1, and the startup page identifies Public Beta 1. The multi-platform tag selects Linux/amd64 or Linux/arm64 for your Docker engine. Architecture-specific tags are 2.0.0-beta.1-amd64 and 2.0.0-beta.1-arm64. Do not use latest for this beta.

This beta is intended for local, single-user use with software rendering. It does not provide authentication, isolated user sessions or a public hosting service. Do not expose the port to a LAN, the Internet or a tunnel. Host GPU availability does not imply container GPU acceleration.

⁠Start

For step-by-step Docker setup, online installation and troubleshooting, use the English guide⁠ or Chinese guide⁠. The Web ZIP on GitHub and NITRC includes a ready-to-use Compose file.

For a direct command-line installation, use a working Linux-container Docker engine. The recipe caps memory at 8 GiB and CPU use at 2 cores. Docker Desktop must remain running. Ensure port 1031 is free and that no existing container uses the name brainnet-viewer-web; do not replace an existing installation without saving its session first.

These single-line commands work in PowerShell and common macOS/Linux shells:

docker pull mingruixia/brainnet-viewer:2.0.0-beta.1
docker run --name brainnet-viewer-web --detach --restart unless-stopped --read-only --memory 8g --cpus 2 --pids-limit 256 --cap-drop ALL --security-opt no-new-privileges --tmpfs /tmp:rw,size=512m,mode=1777 -e LIBGL_ALWAYS_SOFTWARE=1 -p 127.0.0.1:1031:1031 mingruixia/brainnet-viewer:2.0.0-beta.1

Open http://127.0.0.1:1031/index.html⁠. Closing the terminal does not stop the container. This version disables the browser idle-exit timer in container mode. If port 1031 is already occupied, use a free local port, for example 127.0.0.1:1032:1031, and open that host port instead. Keep the loopback binding.

⁠Use and save

1. Add local input files or select bundled examples.
2. Configure Visualization Settings and click Apply to render.
3. Interact with the scene and export figures.

The package includes the original surface library, atlas-network examples, real tutorial data, dual-atlas Inspect information and shared documentation. Example connections were generated from resting-state fMRI data from a group of adult healthy participants.

Save Scene and verify its download before stopping, restarting, updating or removing the container. Uploads, generated files, settings and cameras are temporary session data. A new process does not restore them automatically. Use Open Scene and Apply to restore a saved portable Scene ZIP.

Export Figure downloads the image and its .provenance.json separately, not as a ZIP. Your browser may ask permission for multiple downloads; verify both files. Files already saved on your computer survive container shutdown.

⁠Stop and restart

After saving your Scene:

docker stop brainnet-viewer-web
docker start brainnet-viewer-web
docker logs --tail 60 brainnet-viewer-web

start reuses the existing container; it does not pull a new image or restore the previous scene. Refresh the browser after startup. For an update, save your Scene and keep the old container until you have checked the new version. Do not run two services on the same host port.

⁠Scope and support

Focused public-beta installation, startup and shutdown checks were completed on Windows-hosted Linux/amd64 and Apple Silicon Mac-hosted Linux/arm64. These checks do not establish hardware-GPU acceleration, remote multi-user support or compatibility with every host, display configuration and dataset.

If startup fails, check Docker status, the port mapping and container logs. Large uploads or exports can exhaust the temporary filesystem; save available results and reduce input/output size rather than silently relaxing limits.

• Source and desktop downloads⁠
• NITRC project⁠
• Report a problem⁠

Include the edition, version, OS, processor architecture and reproduction steps. Share only data that you are permitted to disclose.

Developer: Mingrui Xia — [email protected]⁠; [email protected]⁠.

Citation: Xia M, Wang J, He Y (2013). BrainNet Viewer: A Network Visualization Tool for Human Brain Connectomics. PLoS ONE 8: e68910. doi:10.1371/journal.pone.0068910⁠.

Software: GPL-3.0. Bundled third-party resources retain their own notices, terms and citations. See the accompanying license and resource notices.