Graphizy Documentation

Detection to Graph

A powerful, fast, and flexible Python library for building and analyzing graphs from 2D spatial data

Graphizy specializes in creating comprehensive graph types from point data, with advanced memory-enhanced analysis for temporal patterns, sophisticated weight computation systems, and real-time streaming capabilities. Built on OpenCV and igraph, it provides a modern unified API for graph construction and comprehensive analytics.

Quick Start

from graphizy import Graphing, GraphizyConfig, generate_and_format_positions

# Generate sample data
data = generate_and_format_positions(size_x=800, size_y=600, num_particles=100)

# Configure and create grapher
config = GraphizyConfig(dimension=(800, 600))
grapher = Graphing(config=config)

# Create different graph types using unified interface
delaunay_graph = grapher.make_graph("delaunay", data)
proximity_graph = grapher.make_graph("proximity", data, proximity_thresh=50.0)
knn_graph = grapher.make_graph("knn", data, k=4)

# Advanced analysis with new API
graph_info = grapher.get_graph_info(delaunay_graph)
print(f"Density: {graph_info.density:.3f}")
print(f"Connected: {graph_info.is_connected}")

# Use advanced analyzers
social_roles = graph_info.social_analyzer.identify_social_roles(delaunay_graph)
percolation_result = graph_info.percolation_analyzer.analyze_percolation_threshold(
    data, [20, 30, 40, 50, 60]
)

Key Features

  • One API for All Graphs

Create Delaunay, k-NN, MST, Gabriel, Proximity, and even custom graphs with a single make_graph() call. Plugin-friendly, smart defaults, and fully type-safe.

  • Temporal Memory System

Track how connections evolve over time. Use built-in memory features for persistence-aware analysis, temporal filtering, and age-based visualization.

  • Rich Graph Types, Easily Extended

From spatial graphs to domain-specific topologies: support includes Delaunay triangulations, proximity graphs, k-nearest neighbors, MSTs, and custom plugins.

  • Instant Network Analysis

Access over 200 igraph algorithms with real-time stats: clustering, centrality, components, and more. All robust to disconnections. NetworkX compatible.

  • Custom Weights, Real-Time Ready

Define weights using distance, inverse, Gaussian, or custom formulas. Memory-aware weight updates and vectorized for performance.

  • Advanced Tools for Spatial & Temporal Insights

Includes percolation thresholds, service accessibility, social dynamics, and time-aware community tracking — all tailored for dynamic networks.

  • Visualization & Streaming

Visualize network memory with age-based coloring and transparency. Stream updates in real time, or export static snapshots. Comes with CLI tools and interactive demos.

User Guide

Core Systems

Advanced Features

Research Applications

API Reference

Examples & Tutorials

Development

Indices and Tables