Overview

Graphviz is a well-known tool dedicated to export graph in various formats (e.g. svg, jpg, etc.). If you are working in a Jupyter ipython notebook, it is possible to display Graphviz results directly in your Jupyter notebook using the gvmagic extension.

Note that graphviz provides several engine to compute how to place the vertices of your graph in the output image. For small graphs, you should use dot. For large graphs, prefer fdp which is significantly faster.

Graphviz

Installation

• Debian / Ubuntu

sudo apt update
sudo apt install graphviz

Usage

See man dot.

Jupyter notebook

Installation

• Windows: Install anaconda

  conda install -c conda-forge jupyter_contrib_nbextensions

• Debian / Ubuntu

 sudo apt update
 sudo apt install jupyter-notebook

• Under Linux, anaconda-based installation may also be considered.

Usage

• Anaconda installation

jupyter notebook

• Debian / Ubuntu

jupyter-notebook

Drawing graphs

• Minimal example (in a Jupyter notebook):

from pybgl.graph import *
from pybgl.ipynb import ipynb_display_graph

g = DirectedGraph(2)
add_edge(0, 1, g)
print(g.to_dot())       # Print the Graphviz description of the graph
ipynb_display_graph(g)  # Render the graph in Jupyter using Graphviz

• If you use dark theme, it is recommended to change the default foreground color:

from pybgl.graphviz import GraphvizStyle
GraphvizStyle.set_fg_color("grey")

• Note that ipynb_display_graph and .to_dot support optional parameter to customize the graph rendering (see e.g. automaton.py):
  • dv: default vertex style;
  • de: default edge style;
  • dg: graph style;
  • dpv: vertex-dependant style;
  • dpe: edge-dependant style;
  • extra_style: extra style (e.g. splines).

gvmagic

gvmagic is a jupyter notebook extension enabling some magics like %dotstr that may be handy in some situations.

Installation

• Anaconda

sudo jupyter nbextension install https://raw.github.com/cjdrake/ipython-magic/master/gvmagic.py
jupyter nbextension list
mkdir -p ~/.ipython/extensions/
ln -s /usr/local/share/jupyter/nbextensions/gvmagic.py ~/.ipython/extensions/gvmagic.py

• Debian / Ubuntu

sudo jupyter-nbextension install https://raw.github.com/cjdrake/ipython-magic/master/gvmagic.py
jupyter-nbextension list
mkdir -p ~/.ipython/extensions
ln -s /usr/local/share/jupyter/nbextensions/gvmagic.py ~/.ipython/extensions/gvmagic.py

Usage

In your notebook, execute a code cell containing:

%reload_ext gvmagic
%dotstr 'digraph G {0 -> 1}'