diagrams-magic 0.0.1

Jupyter notebook magic cell wrapper for diagrams.

diagrams-magic

The diagrams-magic is an IPython magic that enables calling diagrams within IPython kernel to draw graphs with pure texts. The most common use case, however, would be to use this magic within Jupyter Notebook, which integrates the IPython kernel within it.

Requirements

• diagrams

  The mechanism of this magic is to call diagrams installed on local machine. So please ensure that diagrams is accessible on your system or in your virtual environment.

  If you have node and npm installed, you may install diagrams with

  npm install -g diagrams

• A (virtual) display driver

  diagrams flowchart and diagrams sequence starts an Electron app that requires a display to run. If you are using Windows or MacOS, this shouldn't be a problem.

  However if you are to deploy this magic on a headless Linux backend, diagrams flowchart and diagrams sequence will report 'Cannot open display' error.

  This can be circumvented with a virtual display. We suggest installing xvfb so that our magic will automatically detect packaged xvfb-run and use it. The solution is detailed at Electron's webpage.

Installation

• From source code

  git clone https://github.com/chunxy/diagrams-magic.git
  cd diagrams-magic
  pip install ./
  

Usage

As with usual IPython magic, remember to %load_ext diagrams-magic before using this magic.

This is a cell magic. In the first line of your cell, type %%diagrams (flowchart|dot|sequence|railroad) [name] to indicate which driver you are to use and the name of image file. If no name is provided, no image file will be generated in your current directory (but the image will still be available in the output cell). In the rest of this cell, simply type in the graph descriptions.

See the demo file for reference.

Known Issue(s)

• Invalid asm.js: Function definition doesn't match use

  There might be error like above when you are using the dot driver. This is a known issue caused by the upstream viz.js. Usually the output won't be affected and thus you can ignore it, as long as the red error message won't be too harsh for your eyes.