Cytoscape Layouts

The layout parameter of cyto.Cytoscape takes as argument a
dictionary specifying how the nodes should be positioned on the screen.
Every graph requires this dictionary with a value specified for the
name key. It represents a built-in display method, which is one of the
following:
- preset
- random
- grid
- circle
- concentric
- breadthfirst
- cose

All those layouts, along with their options, are described in the
official Cytoscape documentation.
There, you can find the exact keys accepted by your dictionary, enabling
advanced fine-tuning (demonstrated below).

If preset is given, the positions will be rendered based on the positions
specified in the elements. Otherwise, the positions will be computed by
Cytoscape.js behind the scene, based on the given items of the layout
dictionary. Let’s start with an example of declaring a graph with a preset
layout:

from dash import Dash, html
import dash_cytoscape as cyto

app = Dash()

nodes = [
    {
        'data': {'id': short, 'label': label},
        'position': {'x': 20 * lat, 'y': -20 * long}
    }
    for short, label, long, lat in (
        ('la', 'Los Angeles', 34.03, -118.25),
        ('nyc', 'New York', 40.71, -74),
        ('to', 'Toronto', 43.65, -79.38),
        ('mtl', 'Montreal', 45.50, -73.57),
        ('van', 'Vancouver', 49.28, -123.12),
        ('chi', 'Chicago', 41.88, -87.63),
        ('bos', 'Boston', 42.36, -71.06),
        ('hou', 'Houston', 29.76, -95.37)
    )
]

edges = [
    {'data': {'source': source, 'target': target}}
    for source, target in (
        ('van', 'la'),
        ('la', 'chi'),
        ('hou', 'chi'),
        ('to', 'mtl'),
        ('mtl', 'bos'),
        ('nyc', 'bos'),
        ('to', 'hou'),
        ('to', 'nyc'),
        ('la', 'nyc'),
        ('nyc', 'bos')
    )
]

elements = nodes + edges

app.layout = html.Div([
    cyto.Cytoscape(
        id='cytoscape-layout-1',
        elements=elements,
        style={'width': '100%', 'height': '350px'},
        layout={
            'name': 'preset'
        }
    )
])

if __name__ == '__main__':
    app.run(debug=True)

Here, we provided toy elements using geographically positioned nodes. If
you’d like to reproduce this example by yourself, check out the code
below.

View Elements Declaration

nodes = [
    {
        'data': {'id': short, 'label': label},
        'position': {'x': 20*lat, 'y': -20*long}
    }
    for short, label, long, lat in (
        ('la', 'Los Angeles', 34.03, -118.25),
        ('nyc', 'New York', 40.71, -74),
        ('to', 'Toronto', 43.65, -79.38),
        ('mtl', 'Montreal', 45.50, -73.57),
        ('van', 'Vancouver', 49.28, -123.12),
        ('chi', 'Chicago', 41.88, -87.63),
        ('bos', 'Boston', 42.36, -71.06),
        ('hou', 'Houston', 29.76, -95.37)
    )
]

edges = [
    {'data':
         {'source': source, 'target': target}
     }
    for source, target in (
        ('van', 'la'),
        ('la', 'chi'),
        ('hou', 'chi'),
        ('to', 'mtl'),
        ('mtl', 'bos'),
        ('nyc', 'bos'),
        ('to', 'hou'),
        ('to', 'nyc'),
        ('la', 'nyc'),
        ('nyc', 'bos')
    )
]

elements = nodes + edges

Display Methods

In most cases, the position of the nodes will not be given. In these
cases, one of the built-in methods can be used. Let’s see what happens
when the value of name is set to 'circle' or 'grid'

from dash import Dash, html
import dash_cytoscape as cyto

app = Dash()

nodes = [
    {
        'data': {'id': short, 'label': label},
        'position': {'x': 20 * lat, 'y': -20 * long}
    }
    for short, label, long, lat in (
        ('la', 'Los Angeles', 34.03, -118.25),
        ('nyc', 'New York', 40.71, -74),
        ('to', 'Toronto', 43.65, -79.38),
        ('mtl', 'Montreal', 45.50, -73.57),
        ('van', 'Vancouver', 49.28, -123.12),
        ('chi', 'Chicago', 41.88, -87.63),
        ('bos', 'Boston', 42.36, -71.06),
        ('hou', 'Houston', 29.76, -95.37)
    )
]

edges = [
    {'data': {'source': source, 'target': target}}
    for source, target in (
        ('van', 'la'),
        ('la', 'chi'),
        ('hou', 'chi'),
        ('to', 'mtl'),
        ('mtl', 'bos'),
        ('nyc', 'bos'),
        ('to', 'hou'),
        ('to', 'nyc'),
        ('la', 'nyc'),
        ('nyc', 'bos')
    )
]

elements = nodes + edges

app.layout = html.Div([
    cyto.Cytoscape(
        id='cytoscape-layout-2',
        elements=elements,
        style={'width': '100%', 'height': '350px'},
        layout={
            'name': 'circle'
        }
    )
])

if __name__ == '__main__':
    app.run(debug=True)
from dash import Dash, html
import dash_cytoscape as cyto

app = Dash()

nodes = [
    {
        'data': {'id': short, 'label': label},
        'position': {'x': 20 * lat, 'y': -20 * long}
    }
    for short, label, long, lat in (
        ('la', 'Los Angeles', 34.03, -118.25),
        ('nyc', 'New York', 40.71, -74),
        ('to', 'Toronto', 43.65, -79.38),
        ('mtl', 'Montreal', 45.50, -73.57),
        ('van', 'Vancouver', 49.28, -123.12),
        ('chi', 'Chicago', 41.88, -87.63),
        ('bos', 'Boston', 42.36, -71.06),
        ('hou', 'Houston', 29.76, -95.37)
    )
]

edges = [
    {'data': {'source': source, 'target': target}}
    for source, target in (
        ('van', 'la'),
        ('la', 'chi'),
        ('hou', 'chi'),
        ('to', 'mtl'),
        ('mtl', 'bos'),
        ('nyc', 'bos'),
        ('to', 'hou'),
        ('to', 'nyc'),
        ('la', 'nyc'),
        ('nyc', 'bos')
    )
]

elements = nodes + edges

app.layout = html.Div([
    cyto.Cytoscape(
        id='cytoscape-layout-3',
        elements=elements,
        style={'width': '100%', 'height': '350px'},
        layout={
            'name': 'grid'
        }
    )
])

if __name__ == '__main__':
    app.run(debug=True)

Fine-tuning the Layouts

For any given name item, a collection of keys are accepted by the layout
dictionary. For example, the
grid layout
will accept row and cols, the
circle layout radius
and startAngle, and so forth. Here is the grid layout
with the same graph as above, but with different layout options:

from dash import Dash, html
import dash_cytoscape as cyto

app = Dash()

nodes = [
    {
        'data': {'id': short, 'label': label},
        'position': {'x': 20 * lat, 'y': -20 * long}
    }
    for short, label, long, lat in (
        ('la', 'Los Angeles', 34.03, -118.25),
        ('nyc', 'New York', 40.71, -74),
        ('to', 'Toronto', 43.65, -79.38),
        ('mtl', 'Montreal', 45.50, -73.57),
        ('van', 'Vancouver', 49.28, -123.12),
        ('chi', 'Chicago', 41.88, -87.63),
        ('bos', 'Boston', 42.36, -71.06),
        ('hou', 'Houston', 29.76, -95.37)
    )
]

edges = [
    {'data': {'source': source, 'target': target}}
    for source, target in (
        ('van', 'la'),
        ('la', 'chi'),
        ('hou', 'chi'),
        ('to', 'mtl'),
        ('mtl', 'bos'),
        ('nyc', 'bos'),
        ('to', 'hou'),
        ('to', 'nyc'),
        ('la', 'nyc'),
        ('nyc', 'bos')
    )
]

elements = nodes + edges

app.layout = html.Div([
    cyto.Cytoscape(
        id='cytoscape-layout-4',
        elements=elements,
        style={'width': '100%', 'height': '350px'},
        layout={
            'name': 'grid',
            'rows': 3
        }
    )
])

if __name__ == '__main__':
    app.run(debug=True)

In the case of the circle layout, we can force the nodes to start and end
at a certain angle in radians(import math for this example):

import math

from dash import Dash, html
import dash_cytoscape as cyto

app = Dash()

nodes = [
    {
        'data': {'id': short, 'label': label},
        'position': {'x': 20 * lat, 'y': -20 * long}
    }
    for short, label, long, lat in (
        ('la', 'Los Angeles', 34.03, -118.25),
        ('nyc', 'New York', 40.71, -74),
        ('to', 'Toronto', 43.65, -79.38),
        ('mtl', 'Montreal', 45.50, -73.57),
        ('van', 'Vancouver', 49.28, -123.12),
        ('chi', 'Chicago', 41.88, -87.63),
        ('bos', 'Boston', 42.36, -71.06),
        ('hou', 'Houston', 29.76, -95.37)
    )
]

edges = [
    {'data': {'source': source, 'target': target}}
    for source, target in (
        ('van', 'la'),
        ('la', 'chi'),
        ('hou', 'chi'),
        ('to', 'mtl'),
        ('mtl', 'bos'),
        ('nyc', 'bos'),
        ('to', 'hou'),
        ('to', 'nyc'),
        ('la', 'nyc'),
        ('nyc', 'bos')
    )
]

elements = nodes + edges

app.layout = html.Div([
    cyto.Cytoscape(
        id='cytoscape-layout-5',
        elements=elements,
        style={'width': '100%', 'height': '350px'},
        layout={
            'name': 'circle',
            'radius': 250,
            'startAngle': math.pi * 1 / 6,
            'sweep': math.pi * 2 / 3
        }
    )
])

if __name__ == '__main__':
    app.run(debug=True)

For the breadthfirst layout, a tree is created from the existing nodes
by performing a breadth-first search of the graph. By default, the root(s)
of the tree is inferred, but can also be specified as an option. Here is
how the graph would look like if we choose New York City as the root:

from dash import Dash, html
import dash_cytoscape as cyto

app = Dash()

nodes = [
    {
        'data': {'id': short, 'label': label},
        'position': {'x': 20 * lat, 'y': -20 * long}
    }
    for short, label, long, lat in (
        ('la', 'Los Angeles', 34.03, -118.25),
        ('nyc', 'New York', 40.71, -74),
        ('to', 'Toronto', 43.65, -79.38),
        ('mtl', 'Montreal', 45.50, -73.57),
        ('van', 'Vancouver', 49.28, -123.12),
        ('chi', 'Chicago', 41.88, -87.63),
        ('bos', 'Boston', 42.36, -71.06),
        ('hou', 'Houston', 29.76, -95.37)
    )
]

edges = [
    {'data': {'source': source, 'target': target}}
    for source, target in (
        ('van', 'la'),
        ('la', 'chi'),
        ('hou', 'chi'),
        ('to', 'mtl'),
        ('mtl', 'bos'),
        ('nyc', 'bos'),
        ('to', 'hou'),
        ('to', 'nyc'),
        ('la', 'nyc'),
        ('nyc', 'bos')
    )
]

elements = nodes + edges

app.layout = html.Div([
    cyto.Cytoscape(
        id='cytoscape-layout-6',
        elements=elements,
        style={'width': '100%', 'height': '350px'},
        layout={
            'name': 'breadthfirst',
            'roots': '[id = "nyc"]'
        }
    )
])

if __name__ == '__main__':
    app.run(debug=True)

Here is what would happen if we chose Montreal and Vancouver instead:

from dash import Dash, html
import dash_cytoscape as cyto

app = Dash()

nodes = [
    {
        'data': {'id': short, 'label': label},
        'position': {'x': 20 * lat, 'y': -20 * long}
    }
    for short, label, long, lat in (
        ('la', 'Los Angeles', 34.03, -118.25),
        ('nyc', 'New York', 40.71, -74),
        ('to', 'Toronto', 43.65, -79.38),
        ('mtl', 'Montreal', 45.50, -73.57),
        ('van', 'Vancouver', 49.28, -123.12),
        ('chi', 'Chicago', 41.88, -87.63),
        ('bos', 'Boston', 42.36, -71.06),
        ('hou', 'Houston', 29.76, -95.37)
    )
]

edges = [
    {'data': {'source': source, 'target': target}}
    for source, target in (
        ('van', 'la'),
        ('la', 'chi'),
        ('hou', 'chi'),
        ('to', 'mtl'),
        ('mtl', 'bos'),
        ('nyc', 'bos'),
        ('to', 'hou'),
        ('to', 'nyc'),
        ('la', 'nyc'),
        ('nyc', 'bos')
    )
]

elements = nodes + edges

app.layout = html.Div([
    cyto.Cytoscape(
        id='cytoscape-layout-7',
        elements=elements,
        style={'width': '100%', 'height': '350px'},
        layout={
            'name': 'breadthfirst',
            'roots': '#van, #mtl'
        }
    )
])

if __name__ == '__main__':
    app.run(debug=True)

Notice here that we are not giving the ID of the nodes to the roots
key, but instead using a specific syntax to select the desired elements.
This concept of selector is extensively documented in Cytoscape.js,
and will be further explored in part 3.
We follow the same syntax as the Javascript library.

For preset layouts, you can also specify the positions for which you would like to render each
of your nodes:

from dash import Dash, html
import dash_cytoscape as cyto

app = Dash()

nodes = [
    {
        'data': {'id': short, 'label': label},
        'position': {'x': 20 * lat, 'y': -20 * long}
    }
    for short, label, long, lat in (
        ('la', 'Los Angeles', 34.03, -118.25),
        ('nyc', 'New York', 40.71, -74),
        ('to', 'Toronto', 43.65, -79.38),
        ('mtl', 'Montreal', 45.50, -73.57),
        ('van', 'Vancouver', 49.28, -123.12),
        ('chi', 'Chicago', 41.88, -87.63),
        ('bos', 'Boston', 42.36, -71.06),
        ('hou', 'Houston', 29.76, -95.37)
    )
]

edges = [
    {'data': {'source': source, 'target': target}}
    for source, target in (
        ('van', 'la'),
        ('la', 'chi'),
        ('hou', 'chi'),
        ('to', 'mtl'),
        ('mtl', 'bos'),
        ('nyc', 'bos'),
        ('to', 'hou'),
        ('to', 'nyc'),
        ('la', 'nyc'),
        ('nyc', 'bos')
    )
]

elements = nodes + edges

app.layout = html.Div([
    cyto.Cytoscape(
        id='cytoscape-layout-8',
        elements=elements,
        style={'width': '100%', 'height': '350px'},
        layout={
            'name': 'preset',
            'positions': {
                node['data']['id']: node['position']
                for node in nodes
            }
        }
    )
])

if __name__ == '__main__':
    app.run(debug=True)

In the callbacks chapter, you will learn how to interactively update your layout; in order
to use preset, you will need to specify the position of each node.

Physics-based Layouts

Additionally, the cose layout can be used to position the nodes using
a force-directed layout by simulating attraction and repulsion among the
elements, based on the paper by
Dogrusoz et al, 2009.

from dash import Dash, html
import dash_cytoscape as cyto

app = Dash()

nodes = [
    {
        'data': {'id': short, 'label': label},
        'position': {'x': 20 * lat, 'y': -20 * long}
    }
    for short, label, long, lat in (
        ('la', 'Los Angeles', 34.03, -118.25),
        ('nyc', 'New York', 40.71, -74),
        ('to', 'Toronto', 43.65, -79.38),
        ('mtl', 'Montreal', 45.50, -73.57),
        ('van', 'Vancouver', 49.28, -123.12),
        ('chi', 'Chicago', 41.88, -87.63),
        ('bos', 'Boston', 42.36, -71.06),
        ('hou', 'Houston', 29.76, -95.37)
    )
]

edges = [
    {'data': {'source': source, 'target': target}}
    for source, target in (
        ('van', 'la'),
        ('la', 'chi'),
        ('hou', 'chi'),
        ('to', 'mtl'),
        ('mtl', 'bos'),
        ('nyc', 'bos'),
        ('to', 'hou'),
        ('to', 'nyc'),
        ('la', 'nyc'),
        ('nyc', 'bos')
    )
]

elements = nodes + edges

app.layout = html.Div([
    cyto.Cytoscape(
        id='cytoscape-layout-9',
        elements=elements,
        style={'width': '100%', 'height': '350px'},
        layout={
            'name': 'cose'
        }
    )
])

if __name__ == '__main__':
    app.run(debug=True)

Loading External Layout

External layouts are now available! Update your dash-cytoscape to
version 0.1.1 or later.

The following external layouts are distributed with the official dash-cytoscape library: * cose-bilkent * cola * euler * spread * dagre * klay

In order to use them, you will need to use the load_extra_layouts() function from
dash_cytoscape:

from dash import Dash, dcc, html
import dash_cytoscape as cyto

# Load extra layouts
cyto.load_extra_layouts()

app = Dash()
server = app.server

We also provided a
demo app directly derived from usage-elements,
but with the option to use the external layouts.

Make sure you use external layouts only when necessary. The distribution package takes
almost 3x more space than the regular bundle, which means that it will take more time to
load your apps, especially on slower networks.
This image
shows how long it would take to load the dev package on a slower network.