NOTES.md

Major changes in networkx

These are the main changes from networkx 1.x to networkx 2.x that affect the code in the textbook.

Accessing node properties

The node property of a Graph object no longer contains a dict mapping node names to node properties, but a NodeView object that behaves like a dict. This means you can still get the properties of a node using

g.node['name']

but have to use

dict(g.nodes)

to get a dict of all nodes with their properties.

Also note that the node and nodes are now equivalent with both returning the same NodeView object.

Accessing edges properties

The edge property no longer exists. You can now see all edges using

dict(g.edges)

This returns different output from the old g.edge, namely 2-tuples of edge source and target pairs, rather than a mapping from the source node to all destination nodes that are neighbors. Use

g.edges['source', 'target']

to access the properties of a particular edge.

Degree of a node

The degree of node i in graph g is now found using

g.degree[i]

rather than g.degree(i).

Selecting a random node

This is now more complicated since the nodes property is now a view.

To select a random node from the graph:

random.choice(list(g.nodes))

To select a random neighbor for node i:

random.choice(list(g.neighbors(i)))