networkx-temporal 1.0b4

Python package to build and manipulate temporal NetworkX graphs.

networkx-temporal

Python package to build and manipulate temporal NetworkX graphs.

Requirements

• Python>=3.7
• networkx>=2.1
• pandas>=1.1.0

Install

Package is available to install on PyPI:

pip install networkx-temporal

Usage

A Jupyter notebook with the following code is also available at notebook.ipynb.

Build temporal graph

The Temporal{Di,Multi,MultiDi}Graph class uses NetworkX graphs internally to allow easy manipulation of its data structures:

import networkx_temporal as nxt

TG = nxt.TemporalDiGraph(t=4)

TG[0].add_edge("a", "b")
TG[1].add_edge("c", "b")
TG[2].add_edge("c", "b")
TG[2].add_edge("d", "c")
TG[2].add_edge("d", "e")
TG[3].add_edge("f", "e")
TG[3].add_edge("f", "a")
TG[3].add_edge("f", "b")

print(f"t = {len(TG)} time steps\n"
      f"V = {TG.order()} nodes ({TG.temporal_order()} unique, {TG.total_order()} total)\n"
      f"E = {TG.size()} edges ({TG.temporal_size()} unique, {TG.total_size()} total)")

# t = 4 time steps
# V = [2, 2, 4, 4] nodes (6 unique, 12 total)
# E = [1, 1, 3, 3] edges (7 unique, 8 total)

Draw snapshots

import matplotlib.pyplot as plt

draw_opts = {"arrows": True,
             "node_color": "#aaa",
             "node_size": 250,
             "with_labels": True}

fig, ax = plt.subplots(nrows=1, ncols=4, figsize=(8, 2), constrained_layout=True)

for t, G in enumerate(TG):
    nx.draw(G, pos=nx.kamada_kawai_layout(G), ax=ax[t], **draw_opts)
    ax[t].set_title(f"$t$ = {t}")

plt.show()

Slice into time bins

Once initialized, a specified number of bins can be returned in a new object of the same type using slice:

TGS = TG.slice(bins=2)
TGS.nodes()

# [NodeView(('a', 'b', 'c')),
#  NodeView(('a', 'b', 'c', 'd', 'e', 'f'))]

By default, created bins are composed of non-overlapping edges and might have uneven size. To balance them, pass qcut=True:

TGS = TG.slice(bins=2, qcut=True)
TGS.nodes()

# [NodeView(('a', 'b', 'c', 'd', 'e')),
#  NodeView(('a', 'b', 'e', 'f'))]

Note that in some cases, the qcut method may not be able to split the graph into the number of bins requested and will return the maximum number of bins possible.

Additionally, either duplicates=True (allows duplicate edges among bins) or rank_first=True (ranks edges in order of appearance) may be used to avoid exceptions.

---

Convert from static graph

Static graphs can carry temporal information either in the node- or edge-level attributes.

In the example below, we create a static multigraph in which both nodes and edges are attributed with the time step t in which they are observed:

import networkx as nx

G = nx.MultiDiGraph()

G.add_nodes_from([
    ("a", {"t": 0}),
    ("b", {"t": 0}),
    ("c", {"t": 1}),
    ("d", {"t": 2}),
    ("e", {"t": 3}),
    ("f", {"t": 3}),
])

G.add_edges_from([
    ("a", "b", {"t": 0}),
    ("c", "b", {"t": 1}),
    ("d", "c", {"t": 2}),
    ("d", "e", {"t": 2}),
    ("c", "b", {"t": 2}),
    ("f", "e", {"t": 3}),
    ("f", "a", {"t": 3}),
    ("f", "b", {"t": 3}),
])

print(G)

# MultiDiGraph with 6 nodes and 8 edges

Node-level time attribute

Converting a static graph with node-level temporal data to a temporal graph object (node_level considers the source node's time by default when slicing edges):

TG = nxt.from_static(G).slice(attr="t", attr_level="node", node_level="source", bins=None, qcut=None)
TG.edges(data=True)

# [OutMultiEdgeDataView([('a', 'b', {'t': 0})]),
#  OutMultiEdgeDataView([('c', 'b', {'t': 1}), ('c', 'b', {'t': 2})]),
#  OutMultiEdgeDataView([('d', 'c', {'t': 2}), ('d', 'e', {'t': 2})]),
#  OutMultiEdgeDataView([('f', 'e', {'t': 3}), ('f', 'a', {'t': 3}), ('f', 'b', {'t': 3})])]

Note that considering node-level attributes resulted in misplacing the edge (c, b, 2) in the conversion from static to temporal, as it is duplicated at times 1 and 2.

Edge-level time attribute

Converting a static graph with edge-level temporal data to a temporal graph object (edge's time applies to both source and target nodes):

TG = nxt.from_static(G).slice(attr="t", attr_level="edge", bins=None, qcut=None)
TG.edges(data=True)

# [OutMultiEdgeDataView([('a', 'b', {'t': 0})]),
#  OutMultiEdgeDataView([('c', 'b', {'t': 1})]),
#  OutMultiEdgeDataView([('c', 'b', {'t': 2}), ('d', 'c', {'t': 2}), ('d', 'e', {'t': 2})]),
#  OutMultiEdgeDataView([('f', 'e', {'t': 3}), ('f', 'a', {'t': 3}), ('f', 'b', {'t': 3})])]

Both methods result in the same number of edges, but a higher number of nodes, as they appear in more than one bin in order to preserve all edges in the static graph.

---

Transform temporal graph

Once a temporal graph is instantiated, some methods are implemented that allow returning snaphots, events or unified temporal graphs.

Get snapshots

Returns a list of graphs internally stored under data in the temporal graph object, also accessible by iterating through the object:

STG = TG.to_snapshots()
STG

# [<networkx.classes.multidigraph.MultiDiGraph at 0x7fa0ac6faf50>,
#  <networkx.classes.multidigraph.MultiDiGraph at 0x7fa0ac72aa90>,
#  <networkx.classes.multidigraph.MultiDiGraph at 0x7fa0ac725c10>,
#  <networkx.classes.multidigraph.MultiDiGraph at 0x7fa0ac726290>]

Get static graph

Builds a static or flattened graph containing all the edges found at each time step.

G = TG.to_static()
fig = plt.figure(figsize=(2, 2))
nx.draw(G, pos=nx.kamada_kawai_layout(G), **draw_opts)
plt.show()

Note that the above graph is a MultiGraph, but the visualization is a simple graph drawing a single edge among each node pair.

Get sequence of events

An event-based temporal graph (ETG) is a sequence of 3- or 4-tuple edge-based events.

• 3-tuples: (u, v, t), where elements are the source node, target node, and time step of the observed event (also known as a stream graph);

• 4-tuples: (u, v, t, e), where e is either a positive (1) or negative (-1) unity for edge addition and deletion, respectively.

ETG = TG.to_events()  # stream=True (default)
ETG

# [('a', 'b', 0),
#  ('c', 'b', 1),
#  ('c', 'b', 2),
#  ('d', 'c', 2),
#  ('d', 'e', 2),
#  ('f', 'e', 3),
#  ('f', 'a', 3),
#  ('f', 'b', 3)]

ETG = TG.to_events(stream=False)
ETG

# [('a', 'b', 0, 1),
#  ('c', 'b', 1, 1),
#  ('a', 'b', 1, -1),
#  ('d', 'c', 2, 1),
#  ('d', 'e', 2, 1),
#  ('f', 'e', 3, 1),
#  ('f', 'a', 3, 1),
#  ('f', 'b', 3, 1),
#  ('c', 'b', 3, -1),
#  ('d', 'c', 3, -1),
#  ('d', 'e', 3, -1)]

Get unified temporal graph

The unified temporal graph (UTG) is a single graph that contains the original data plus proxy nodes and edge couplings connecting sequential temporal nodes.

UTG = TG.to_unified(add_couplings=True,
                    add_proxy_nodes=False,
                    proxy_nodes_with_attr=True,
                    prune_proxy_nodes=True)

print(UTG)

# DiGraph with 12 nodes and 14 edges

nodes = sorted(TG.temporal_nodes())
pos = {
    node: (nodes.index(node.rsplit("_")[0]), -int(node.rsplit("_")[1]))
    for node in UTG.nodes()
}
fig = plt.figure(figsize=(4, 4))
nx.draw(UTG, pos=pos, connectionstyle="arc3,rad=0.25", **draw_opts)
plt.show()

Convert back to TemporalGraph object

Functions to convert a newly created STG, ETG, or UTG back to a temporal graph object are also implemented.

nxt.from_snapshots(STG)
nxt.from_events(ETG, directed=True, multigraph=False)
nxt.from_unified(UTG)

---

Get temporal information

All methods implemented by networkx, e.g., {in_,out_}degree, are also available to be executed sequentially on the stored time slices.

A few additional methods that consider all time slices are also implemented for convenience, e.g., temporal_{in_,out_}degree.

Node degrees

TG.degree()

# [DiMultiDegreeView({'b': 1, 'a': 1}),
#  DiMultiDegreeView({'c': 1, 'b': 1}),
#  DiMultiDegreeView({'b': 1, 'c': 2, 'd': 2, 'e': 1}),
#  DiMultiDegreeView({'a': 1, 'b': 1, 'e': 1, 'f': 3})]

To obtain the degrees of nodes at a specific time step, use the degree method with the temporal graph index:

TG[0].degree()

# DiMultiDegreeView({'b': 1, 'a': 1})

And to obtain the degree of all nodes or a specific node considering all time steps:

TG.temporal_degree()

# {'c': 3, 'a': 2, 'f': 3, 'd': 2, 'b': 4, 'e': 2}

TG.temporal_degree("a")

# 2

Node neighborhoods

TG.neighbors("c")

# [[], ['b'], ['b'], []]

To obtain the temporal neighborhood of a node considering all time steps, use the method temporal_neighbors:

TG.temporal_neighbors("c")

# ['b']

Order and size

TG.order(), TG.size()

# ([2, 2, 4, 4], [1, 1, 3, 3])

Note that the temporal order and size are defined as the number of unique nodes and edges, respectively, across all time steps:

TG.temporal_order(), TG.temporal_size()

# (6, 7)

To consider nodes or edges with distinct attributes as non-unique, pass data=True:

TG.temporal_order(data=True), TG.temporal_size(data=True)

# (6, 8)

And to obtain the total number of nodes and edges across all time steps, use the total_order and total_size methods instead:

TG.total_order(), TG.total_size()  # sum(TG.order()), sum(TG.size())

# (12, 8)

---

References

• NetworkX
• Pandas