pyvis

Pyvis is a Python library designed to construct and visualize network graphs within a single workflow. It leverages the VisJS library to provide interactive web-based visualizations.

Key capabilities include:

• Customization: Configure nodes and edges individually with specific colors, sizes, labels, and metadata.
• Interactivity: Graphs support user interactions such as dragging, hovering, and selecting nodes and edges.
• Layout Control: Access and tweak layout algorithms to optimize the rendering of large-scale graphs.

You can inject a UI into your HTML output to dynamically adjust network settings (like physics) in the browser. Use show_buttons(filter_=['physics']) to limit the UI to specific categories. You can then copy the generated options from the UI and use set_options(options_string) in your Python code to finalize the configuration.

# Show only physics settings in the UI
net.show_buttons(filter_=['physics'])

# Later, after experimenting in the browser:
net.set_options(generated_options_string)

If you prefer to install from a downloaded archive, unpack the project and run the installation command from the top-level directory.

$ python setup.py install

To create a simple network visualization, instantiate a Network object, add nodes using add_node(), add edges using add_edge(), and then export the visualization to an HTML file using show().

from pyvis.network import Network

g = Network()
g.add_node(0)
g.add_node(1)
g.add_edge(0, 1)
g.show("basic.html")

To start creating a network visualization, you must instantiate the Network class. You can optionally pass parameters like height, width, bgcolor, and font_color to the constructor.

from pyvis.network import Network

# Basic instantiation
net = Network()

# Custom instantiation
got_net = Network(height="750px", width="100%", bgcolor="#222222", font_color="white")

You can convert a networkx.Graph object directly into a Pyvis network using the from_nx(nx_graph) method. NetworkX node properties with names matching Pyvis attributes (like title) are translated automatically.

import networkx as nx
from pyvis.network import Network

nx_graph = nx.cycle_graph(10)
# Add properties to NetworkX nodes
nx_graph.nodes[1]['title'] = 'Number 1'

nt = Network('500px', '500px')
tn.from_nx(nx_graph)
tn.show('nx.html')

The recommended way to install pyvis is using pip.

$ pip install pyvis

To embed a Pyvis network in a Jupyter Notebook, instantiate the Network class with notebook=True.

Note for Chrome users: If the graph fails to render, you may need to pass cdn_resources='remote' or cdn_resources='inline' to the Network constructor.

# Standard Jupyter usage
from pyvis.network import Network
net = Network(notebook=True)

# If using Chrome and experiencing rendering issues:
net = Network(notebook=True, cdn_resources='remote')

Pyvis provides two interactive features for exploring large networks:

1. Highlighting: Set select_menu=True in the Network constructor. This adds a dropdown that allows you to select a node; the selected node and its neighbors will be highlighted while others are greyed out.
2. Filtering: Set filter_menu=True in the Network constructor. This allows you to build queries to filter nodes or edges based on specific attributes and values.

# Enable highlighting via dropdown
net_highlight = Network(select_menu=True)

# Enable attribute filtering
net_filter = Network(filter_menu=True)

Pyvis provides interactive UI elements to explore large networks:

• select_menu=True: Adds a dropdown menu to select a node. Selecting a node highlights it and its neighboring edges and nodes.
• filter_menu=True: Adds a menu to filter nodes and edges based on their properties (e.g., group, color, or custom attributes). Any custom property added to a node or edge will be available in this menu.

Example initialization:

got_net = Network(notebook=True, select_menu=True, filter_menu=True)

To create a network visualization, import the Network class from pyvis.network. When working in Jupyter notebooks, set notebook=True. The cdn_resources parameter controls how dependencies are loaded: use 'in_line' to embed resources directly or 'remote' to load them from a CDN.

Basic usage involves creating a Network instance, adding nodes and edges, and calling .show("filename.html") to generate the visualization.

from pyvis.network import Network

# notebook=True for Jupyter, cdn_resources='in_line' embeds resources
g = Network(notebook=True, cdn_resources='in_line')
g.add_nodes(range(5))
g.add_edges([(0, 2), (0, 3), (0, 4), (1, 1), (1, 3), (1, 2)])
g.show("example.html")

Use the get_node(id) method to retrieve the dictionary of properties for a specific node using its unique ID.

net.add_nodes(["a", "b", "c"])
node_data = net.get_node("c")
# Returns: {'id': 'c', 'label': 'c', 'shape': 'dot'}