# HIDDEN
# Clear previously defined variables
%reset -f
# HIDDEN
import numpy as np
np.set_printoptions(threshold=999)
np.random.seed(64)

nbinteract

nbinteract is a Python package that provides a command-line tool to generate interactive web pages from Jupyter notebooks. It allows Jupyter widgets to remain interactive even when the notebook is converted to static HTML by using Binder servers as the computational backend.

nbinteract also provides Python functions for simple, interactive plots. These interactions are driven by data, not callbacks, allowing authors to focus on the logic of their programs.

nbinteract is useful for:

  • Data scientists that want to create simple interactive blog posts without having to know / work with Javascript.
  • Instructors that want to include interactive examples in their textbooks.
  • Students that want to publish data analysis that contains interactive demos.

Converting Notebooks

From the command line:

# Run on the command line to convert the notebook into a publishable HTML page.
#
# nbinteract {NOTEBOOK.ipynb} -s {BINDER_SPEC}
#
# Replace {BINDER_SPEC} with a Binder spec in the format
# {username}/{repo}/{branch} (e.g. SamLau95/nbinteract-image/master).
# The branch is optional; if omitted, defaults to `master`
#
# Replace {NOTEBOOK.ipynb} with the name of the notebook file to convert.
#
# For example:
nbinteract homepage.ipynb -s SamLau95/nbinteract-image

After initializing a GitHub repo and running nbinteract init, you may omit the Binder spec and simply write:

nbinteract homepage.ipynb

For more information on Binder specs and conversion, see the tutorial which has a complete walkthrough on publishing a notebook to the web.

Plotting

Most plotting functions from other libraries (e.g. matplotlib) take data as input. nbinteract's plotting methods instead take in functions that return data.

In the example below, the normal function generates data that we then plot using nbi.hist().

import numpy as np
import nbinteract as nbi

def normal(mean, sd):
    '''Returns 1000 points drawn at random fron N(mean, sd)'''
    return np.random.normal(mean, sd, 1000)

normal(10, 1.0)
array([11.10032294,  8.01737258,  8.84975049, ...,  9.86721442,
       11.06511688, 10.88371858])
# Plot aesthetics
options = {
    'xlim': (-2, 12),
    'ylim': (0, 0.7),
    'bins': 20
}

# Pass in the `normal` function and let user change mean and sd.
# Whenever the user interacts with the sliders, the `normal` function
# is called and the returned data are plotted.
nbi.hist(normal, mean=(0, 10), sd=(0, 2.0), options=options)

# Clicking the Show widget button below loads all widgets on the page.
# Widgets will automatically load for all subsequent pages until you close
# the tab/window.

Simulations are easy to create using nbinteract. In this simulation, we roll a die and plot the running average of the rolls. We can see that with more rolls, the average gets closer to the expected value: 3.5.

rolls = np.random.choice([1, 2, 3, 4, 5, 6], size=300)
averages = np.cumsum(rolls) / np.arange(1, 301)

def x_vals(num_rolls):
    return range(num_rolls)

# The function to generate y-values gets called with the
# x-values as its first argument.
def y_vals(xs):
    return averages[:len(xs)]
nbi.line(x_vals, y_vals, num_rolls=(1, 300))

Installation

Using pip:

pip install nbinteract

# The next two lines can be skipped for notebook version 5.3 and above
jupyter nbextension enable --py --sys-prefix widgetsnbextension
jupyter nbextension enable --py --sys-prefix bqplot

You may now import the nbinteract package in Python code and use the nbinteract CLI command to convert notebooks to HTML pages.

Documentation

Access the tutorials, examples, and documentation for nbinteract using the links in the sidebar.

Feedback

If you have any questions or comments, send us a message on the Gitter channel. We appreciate your feedback!