This project is in sunset mode. It is no longer actively maintained.
To serve widgets and rich outputs, we recommend using voila instead. See this issue for a discussion.
Ever wanted to display Jupyter widgets outside of a notebook environment? This lets you do it.
For more information, view the docs page.
Warnings
You probably should not use this in production yet.
This only supports Python 3.6 at present (PRs welcome)!
You should install ipywidgets_server with pip
:
$ pip install ipywidgets_server
To install from source, clone this repository and run:
$ cd ipywidgets_server/
$ pip install -e .
Create a python file (not a notebook) that renders your favourite widget. Let's, for
instance, create a file called example.py
that contains:
# example.py
from ipywidgets import IntSlider, Text, VBox
s = IntSlider(max=200, value=100)
t = Text()
def update_text(change=None):
t.value = str(float(s.value) ** 2)
s.observe(update_text, names='value')
update_text()
vbox = VBox([s, t])
To serve the widget, go to the directory containing example.py
and run:
$ ipywidgets-server example:vbox
This will serve the widget on http://localhost:8866
:
The command line argument is <module name>:<object name>
, where <module name>
is the name of a Python module that can be imported, and <object name>
is the name of the variable that holds the widget to display.
For information on other command line arguments, run:
$ ipywidgets-server --help
Much of the frontend code for this is inspired by the web3 example in ipywidgets.
The original idea of writing a custom kernel to stop arbitrary code execution on the server came from Maarten Breddels.
Most of the implementation is inspired by various parts of ipywidgets and Jupyter.