vgraph-forge 0.1.2

A Rust-powered toolkit for building visibility graphs from time series.

vgraph-forge

This project provides a Rust-powered library for creating and manipulating visibility graphs, with first-class Python bindings generated via PyO3 and maturin. It currently implements both the Natural Visibility Graph (NVG) and the Horizontal Visibility Graph (HVG), exposing the same data structures to Rust and Python callers.

Features

• Efficient storage for visibility graphs, including helper methods to inspect vertices and edges.
• Natural and horizontal visibility graph builders for scalar time-series data.
• Optional adjacency filtering and subgraph extraction utilities.
• Easy integration with Python via a compiled extension module.

Installation

Once published on PyPI, the package can be installed with pip:

pip install vgraph-forge

During development (or before publishing), build a local wheel with maturin build --release and install the resulting artifact from target/wheels/ in your downstream project via pip install <wheel>. Only contributors building from source need Rust and maturin; consumers installing the wheel do not.

Usage

Below is a compact Python example that constructs both NVG and HVG representations from a simple time series:

from vgraph_forge import (
	build_visibility_graph,
	build_horizontal_visibility_graph,
)

series = [1.0, 3.2, 2.5, 4.0, 1.7]

# Natural visibility graph (NVG)
nvg = build_visibility_graph(series)
print("NVG vertices:", nvg.vertices)
print("NVG edges:", nvg.edges)

# Horizontal visibility graph (HVG) with adjacent-edge filtering
hvg = build_horizontal_visibility_graph(series, skip_adjacent=True)
print("HVG edges (skip adjacent):", hvg.edges)

print("NVG edge count:", nvg.edge_count())
print("HVG node count:", hvg.node_count())

See examples/python/plot_visibility_graph.py for a Matplotlib visualization helper and examples/python/visibility_from_csv.py for a CSV-to-edge-list CLI.

Contributing

Contributions are welcome! Please open an issue or submit a pull request for any improvements or bug fixes.

License

This project is licensed under the MIT License. See the LICENSE file for more details.