NetworkX is indeed the tool of the craft, at least for this book, and I would like to justify my choice.
The main winning points of NetworkX are fourfold:
NetworkX is painless to install. Since it is written in pure Python, it requires no compilation.
NetworkX has excellent online documentation, far superior to that of iGraph and graph-tool. Besides, it has an active community of supporters on StackOverflow.
NetworkX’s structure (functions, algorithms, and attributes) are in good agreement with CNA tasks.
NetworkX’s performance is acceptable up to about 100,000 nodes.
While it is true that NetworkX lacks some essential features (such as community detection and advanced visualization layouts), you can easily add these ...