Integration with matplotlib

The drawing capabilities of graph-tool (see draw module) can be integrated with matplotlib, as we demonstrate in the following.

Note

Integration with matplotlib works with every backend, but vector drawing only works with a cairo-based backend (e.g. cairo or GTK3Cairo). The backend can be changed by calling matplotlib.pyplot.switch_backend():

import matplotlib.pyplot as plt
plt.switch_backend("cairo")

When using a backend not based on cairo, rasterization will be used instead. In this case, the resolution can be controlled via the dpi parameter of matplotlib.figure.Figure.

Drawing with matplotlib is done by calling graph_draw() and passing a container (e.g. matplotlib.axes.Axes) as the mplfig parameter. When this option is passed, the function will return a GraphArtist() object that has been added to the figure.

Tip

Axis autoscaling will work as expected with GraphArtist(), but the aspect ratio will be set the by the figure axis shape, rather than the node positions themselves.

In order to restore the aspect ratio independently of the axis shape, GraphArtist() offers a fit_view() method that does this automatically.

Furthermore, when calling graph_draw() without integrating with matplotlib, the node positions correspond to cairo coordinates, which have an origin in the upper left corner, and with the y axis increasing from top to bottom.

In order for the visualization to be the same when matplotlib is being used, the y axis needs to be flipped by inverting the limits with matplotlib.axes.Axes.set_ylim().

Alternatively, the option yflip can be passed to fit_view() for this to be done automatically.

The example below shows how to plot several graphs in different subplots of the same figure.

import graph_tool.all as gt
import matplotlib.pyplot as plt

plt.switch_backend("cairo")   # to enable vector drawing

fig, ax = plt.subplots(2, 2, figsize=(12, 11.5))

g = gt.collection.data["polbooks"]

a = gt.graph_draw(g, g.vp.pos, vertex_size=1.5, mplfig=ax[0,0])
a.fit_view(yflip=True)

ax[0,0].set_xlabel("$x$ coordinate")
ax[0,0].set_ylabel("$y$ coordinate")

state = gt.minimize_nested_blockmodel_dl(g)

a = state.draw(mplfig=ax[0,1])[0]
a.fit_view(yflip=True)

ax[0,1].set_xlabel("$x$ coordinate")
ax[0,1].set_ylabel("$y$ coordinate")

g = gt.collection.data["lesmis"]
a = gt.graph_draw(g, g.vp.pos, vertex_size=1.5, mplfig=ax[1,0])
a.fit_view(yflip=True)

ax[1,0].set_xlabel("$x$ coordinate")
ax[1,0].set_ylabel("$y$ coordinate")

state = gt.minimize_nested_blockmodel_dl(g)

a = state.draw(mplfig=ax[1,1])[0]
a.fit_view(yflip=True)

ax[1,1].set_xlabel("$x$ coordinate")
ax[1,1].set_ylabel("$y$ coordinate")

plt.subplots_adjust(left=0.08, right=0.99, top=0.99, bottom=0.06)
fig.savefig("gt-mpl.svg")

Four subplots showing networks drawn using graph-tool.

Integration with basemap

As a slightly more elaborate example, below we show how we can draw the European airline graph on a map using mpl_toolkits.basemap.

from itertools import chain
from mpl_toolkits.basemap import Basemap

fig, ax = plt.subplots(1, 1, figsize=(8, 8))

g = gt.collection.ns["eu_airlines"]
pos = gt.group_vector_property([g.vp.nodeLong, g.vp.nodeLat])

m = Basemap(projection='ortho', resolution=None,
            lat_0=g.vp.nodeLat.fa.mean(), lon_0=g.vp.nodeLong.fa.mean())

m.shadedrelief(scale=.2)

lats = m.drawparallels(np.linspace(-90, 90, 13))
lons = m.drawmeridians(np.linspace(-180, 180, 13))
lat_lines = chain(*(tup[1][0] for tup in lats.items()))
lon_lines = chain(*(tup[1][0] for tup in lons.items()))
all_lines = chain(lat_lines, lon_lines)
for line in all_lines:
    line.set(linestyle='-', alpha=0.3, color='w')

a = gt.graph_draw(g, pos=pos.t(lambda x: m(*x)),          # project positions
                  edge_color=(.1,.1,.1,.1), mplfig=ax)

a.fit_view()
a.set_zorder(10)
tight_layout()
fig.savefig("gt-map.svg")

Network of European fligths drawn on the globe with an orthographic projection.