🤖 AI Summary
Existing open-source libraries lack efficient and user-friendly support for analyzing discrete-timestamp dynamic graphs, hindering rapid computation of key temporal metrics—including temporal distance, centrality, topological overlap, burstiness, and temporal diameter. To address this gap, we propose TGLib: the first lightweight, cross-language (C++/Python) library for time-respecting graph analysis. TGLib systematically integrates high-performance, temporally aware data structures and algorithms—specifically, adjacency-based temporal indexing, event-driven traversal, memory-efficient temporal edge storage, and multi-threaded acceleration. Experimental evaluation demonstrates that its core algorithms achieve 10–100× speedups over NetworkX and other baselines on both real-world and synthetic temporal graphs. TGLib thus fills a critical void in general-purpose, high-performance, cross-platform temporal graph analytics tools. It is already widely adopted in research and teaching.
📝 Abstract
We initiate an open-source library for the efficient analysis of temporal graphs. We consider one of the standard models of dynamic networks in which each edge has a discrete timestamp and transition time. Recently there has been a massive interest in analyzing such temporal graphs. Common computational data mining and analysis tasks include the computation of temporal distances, centrality measures, and network statistics like topological overlap, burstiness, or temporal diameter. To fulfill the increasing demand for efficient and easy-to-use imple-mentations of temporal graph algorithms, we introduce the open-source library Tglib,which integrates efficient data structures and algorithms for temporal graph analysis. Tglibis highly efficient and versatile, providing simple and convenient C++ and Python interfaces, targeting computer scientists, practitioners, students, and the (temporal) network research community.
