Examples

Graph and Digraph objects have a _repr_svg_() method so that they can be rendered and displayed directly in a Jupyter notebook.

Simple example

import graphviz

dot = graphviz.Digraph("hello-pythonistas", comment="Hello world example")

dot.edge("Hello", "Pythonistas!")

dot

You can also display the source code with:

print(dot.source)

// Hello world example
digraph "hello-pythonistas" {
        Hello -> "Pythonistas!"
}

It is also possible to output the comment or other elements of the source text, for example with:

print(dot.comment)

Hello world example

You can also use data from a pandas DataFrame, for example:

import pandas as pd


j = {
    "action": [
        "single use",
        "teamwork",
        "convert",
        "Java, R, Julia etc.",
        "extend",
    ],
    "view": ["Jupyter", "JupyterHub", "nbconvert", "kernels", "extensions"],
}

df = pd.DataFrame(j)

df

	action	view
0	single use	Jupyter
1	teamwork	JupyterHub
2	convert	nbconvert
3	Java, R, Julia etc.	kernels
4	extend	extensions

jm = graphviz.Graph("jupyter_moons", comment="Jupyter moons")

jm.node("What do you want to do?")

for index, row in df.iterrows():
    jm.edge(
        "What do you want to do?", str(row["view"]), label=(str(row["action"]))
    )

jm

Styling

You can use graph_attr, node_attr and edge_attr arguments of the Graph and Digraph constuctors to change the `default attributes of Graphviz](https://www.graphviz.org/doc/info/attrs.html) for your graphs, nodes and edges, for example:

dot = graphviz.Digraph(
    "hello-pythonistas",
    comment="Hello world example",
    node_attr={"shape": "plaintext"},
)

dot.edge("Hello", "Pythonistas!")

dot

The graph_attr-, node_attr- and edge_attr arguments can also be applied to instances:

dot.graph_attr["rankdir"] = "LR"

dot

To add att_stmt statements directly, call the attr() attr() method of the Graph or Digraph instance with the desired target as the first argument and the attributes as the keyword argument.

Hinweis:

Attribute statements affect all subsequent graphs, nodes or edges within the same (sub)graph.

Engines

In addition to Dot, various other layout engines can also be used.

pvl = graphviz.Graph("python_visualisation_landscape", engine="neato")

pvl.edge("Matplotlib", "pandas")
pvl.edge("pandas", "GeoPandas")
pvl.edge("Matplotlib", "Geoplot")
pvl.edge("Matplotlib", "descartes")
pvl.edge("Matplotlib", "seaborn")
pvl.edge("Matplotlib", "ggpy")
pvl.edge("Matplotlib", "plotnine")
pvl.edge("Matplotlib", "scikit_plot")

pvl

You can also change the engine attribute of an existing instance:

pvl.engine = "circo"

pvl

You can get a complete overview of the engines with:

for engine in sorted(graphviz.ENGINES):
    print(engine)

circo
dot
fdp
neato
osage
patchwork
sfdp
twopi

Read and write files

pvl.save()

'python_visualisation_landscape.gv'

graphviz.Source.from_file("python_visualisation_landscape.gv")

Converting dot files

A dot file can be converted to another format, such as PDF, PNG, SVG etc., with render:

from graphviz import render


render("dot", "svg", "python_visualisation_landscape.gv")

'python_visualisation_landscape.gv.svg'

Siehe auch:

• graphviz manual

• examples/graphviz-notebook

• examples/graphviz-engines

• examples/graphviz-escapes.ipynb

• Graphviz Online