A high-performance graph drawing library for Rust, Python, and WebAssembly. egraph provides comprehensive graph algorithms, layout techniques, and visualization tools for network analysis and visualization.

• Multiple Language Bindings: Native Rust, Python (via PyO3), and JavaScript/WebAssembly
• Rich Layout Algorithms: SGD variants, MDS, Stress Majorization, Kamada-Kawai, and more
• Multiple Drawing Spaces: Euclidean (2D/nD), Spherical, Hyperbolic, and Torus
• Community Detection: Louvain, Label Propagation, Spectral Clustering, InfoMap
• Graph Algorithms: Shortest path, triangulation, layering, connected components
• Quality Metrics: Comprehensive layout evaluation tools

Version: 6.0.0-alpha

The Python bindings provide a complete interface to egraph's functionality with a Pythonic API.

# Clone the repository
git clone https://github.com/likr/egraph-rs.git
cd egraph-rs/crates/python

# Install with pip (requires Rust toolchain)
pip install maturin
maturin develop --release

import egraph as eg

# Create a graph
graph = eg.Graph()
nodes = [graph.add_node() for _ in range(10)]
for i in range(9):
    graph.add_edge(nodes[i], nodes[i + 1])

# Apply a layout
drawing = eg.DrawingEuclidean2d.initial_placement(graph)
rng = eg.Rng.seed_from(42)

sgd = eg.FullSgd().build(graph, lambda i: 1.0, rng)
scheduler = sgd.scheduler_exponential(100)

def step(eta):
    sgd.shuffle(rng)
    sgd.apply(drawing, eta)

scheduler.run(step)

# Get positions
positions = {i: (drawing.x(i), drawing.y(i)) for i in range(graph.node_count())}

Full Python documentation: https://egraph-py-docs.netlify.app/

See crates/python/README.md for detailed Python binding documentation.

# Build all crates
cargo build --workspace

# Run tests
make test

# Run tests for specific crate
make test-crate CRATE=petgraph-layout-sgd

# Format and lint
make fmt
make lint

# Build Python bindings
make python-build

# Run Python tests
make python-test

# Build documentation
make python-docs

# Run doctests
make python-doctest

# Build WebAssembly module
npm run wasm-build

# Run WebAssembly tests
wasm-pack test --node crates/wasm

# Run examples
npm start

egraph-rs/
├── crates/
│   ├── algorithm/          # Graph algorithms
│   ├── clustering/         # Community detection
│   ├── drawing/           # Drawing implementations
│   ├── layout/            # Layout algorithms
│   ├── python/            # Python bindings (PyO3)
│   ├── wasm/              # WebAssembly bindings
│   └── quality-metrics/   # Layout quality metrics
├── js/                    # JavaScript examples and datasets
└── memory-bank/           # Project documentation

• SGD Family: Full, Sparse, Distance-Adjusted, Omega variants
• Classical: MDS (Classical/Pivot), Stress Majorization, Kamada-Kawai
• Specialized: Overlap Removal, Separation Constraints

• Louvain (modularity-based)
• Label Propagation
• Spectral Clustering
• InfoMap (information-theoretic)

• Shortest Path (Dijkstra)
• Delaunay Triangulation
• Hierarchical Layering
• Connected Components

• Python: https://egraph-py-docs.netlify.app/
• Rust: Run cargo doc --open for API documentation
• Examples: See js/examples/ and crates/python/examples/

Contributions are welcome! Please feel free to submit issues or pull requests.

MIT

If you use egraph in your research, please cite:

@software{egraph,
  title = {egraph: High-Performance Graph Drawing Library},
  author = {Yosuke Onoue},
  year = {2025},
  url = {https://github.com/likr/egraph-rs}
}