Appearance

  2. Tutorials
  3. Using Ogma in Jupyter Notebooks

Using Ogma in Jupyter Notebooks ​

You can use Ogma together with Jupyter Notebooks to create interactive notebooks. This is a great way to explore the Ogma API and to create visualizations of your models.

In order to start off using Ogma in a Jupyter Notebook, you just need to import it as a Javascript library and a block, then load data, styles and behaviours, like layout, transformations or other things that you need.

Dependencies and setting up the cell ​

Add the following dependencies to your notebook:

python
from IPython.display import HTML, Javascript, display
from string import Template

Now you need to load Ogma and set up the <div> element that will contain the graph. You can do this by using the page function and a simple HTML template:

python
def page(script):
  return HTML(Template('''
  <style>
  #graph-container {
    top: 0; bottom: 0;
    left: 0; right: 0;
    width: 100%;
    height: 400px;
    margin: 0;
    overflow: hidden;
  }
  </style>
  <div id="graph-container"></div>
  <script src="path/to/ogma.js"></script>
  <script>$script</script>
  ''').substitute(script=script))

Note that you will need to change the path to the ogma.js file to the correct one.

Populating your visualization ​

Now you can load the data and styles that you need. For example, you can create a random graph and color the nodes randomly. You can do this by using the Javascript block and putting the content of one of our examples in it:

python
script = Javascript(
'''
// Create Ogma instance
const ogma = new Ogma({ container: 'graph-container' });

const COLORS = [
  '#161344',
  '#3f1c4c',
  '#632654',
  '#86315b',
  '#a93c63',
  '#cd476a',
  '#f35371',
  'white'
];

ogma.setOptions({ background: 'rgb(113,199,183)' });

// Rule to randomly color the nodes.
// This random function must be deterministic.
ogma.styles.addNodeRule({
  color: n => COLORS[n.getId() % COLORS.length]
});

const pulse_options = {
  number: 5,
  duration: 1700,
  interval: 600,
  startColor: 'inherit',
  endColor: 'rgb(113,199,183)',
  width: 5,
  startRatio: 1,
  endRatio: 5
};

// Generate a random graph
Ogma.generate.random({ nodes: 10, edges: 0 })
  .then(graph => ogma.setGraph(graph))
  .then(() => ogma.view.locateGraph());
''')

display(page(script.data))

In a similar way, you can pass the JSON data to the cell, for that you will need to modify the page function to accept a JSON object and then use it in the Javascript block:

python
def vis_with_data(script, data):
  return HTML(Template('''
  <style>
  #graph-container {
    top: 0; bottom: 0;
    left: 0; right: 0;
    width: 100%;
    height: 400px;
    margin: 0;
    overflow: hidden;
  }
  </style>
  <div id="graph-container"></div>
  <!-- use real file location -->
  <script src="path/to/ogma.js"></script>
  <!-- visualisation setup -->
  <script>$script</script>
  <!-- second script to load the data -->
  <script>ogma.setGraph($data).then(() => ogma.view.locateGraph())</script>
  ''').substitute(script=script, data=data))

Note the second script that loads the data.

python
import json
data = {
    'nodes': [{
        'id': 0,
        'attributes': {
            'text': 'New node'
        }
    }]
}
script = Javascript(
'''
const ogma = new Ogma({ container: 'graph-container' });

const COLORS = [
  '#161344',
  '#3f1c4c',
  '#632654',
  '#86315b',
  '#a93c63',
  '#cd476a',
  '#f35371',
  'white'
];

ogma.setOptions({ background: 'rgb(113,199,183)' });

ogma.styles.addNodeRule({
  color: n => COLORS[n.getId() % COLORS.length]
});
''')

display(vis_with_data(script.data, json.dumps(data)))

Now you can see the node that you have just passed to your visualisation cell.

Jupyter Notebook