uclnlp / egal

easy drawing in jupyter
Apache License 2.0
257 stars 21 forks source link
jupyter svg

egal

easy SVG drawing in jupyter and elsewhere...

Example

Instructions

Click on the "brush" icon in the jupyter toolbar to create an SVG cell below the current selection.

Labels and Latex

To change the label of a shape, double-click on it and type. Hit enter to finish. To use latex, start and finish the label with $$ as in $$\sum_i i$$.

Animation

The egal canvas shows/overlays a window of frames. This window is specified by the two numbers between the left and right arrow icons. The left number is the first frame included in the window, the right number the last frame. Each object (circles, boxes, freestyle shapes etc.) is associated with its own frame window which appears on the right when selecting the object. An object is rendered on the egal canvas when the egal frame window overlaps with the object frame window.

Using this setup, to generate an animate build-in sequence you can:

To click through the animation, set the egal window to (1,1). Then click on the right arrow which will move the frame window to (2,2), and so on. It will look something like this:

Features

egal's focus is on drawing simple graphs:

How Does it Work?

egal creates a raw jupyter cell and stores the edited SVG in the source code field of that cell.

Installation

Get the Python Package

Manually by cloning and changing the python path:

git clone https://github.com/uclmr/egal.git
cd egal
export PYTHONPATH=. 

or (experimental) install python package directly:

pip3 install git+https://github.com/uclmr/egal.git

Install and Enable Extension

jupyter nbextension install --py egal 
jupyter nbextension enable --py egal 

Optional: Server Extension

If you want to use egal outside of a notebook you need to install the server extensions (which allows clients to save the SVG on the jupyter server):

jupyter serverextension enable --py egal 

Then you can edit an SVG on the server via accessing http://localhost:8888/files/draw.html (assuming you run jupyter notebook locally).