Components



In [1]:

    
from IPython.display import display, Image, HTML
from talktools import website, nbviewer

Jupyter/Google coLaboratory

In May of 2014, the Project Jupyter and Google Research announced a collaboration that will bring the Jupyter Notebook to Google Drive.

Main features:

Store notebooks on Google Drive
Share notebooks and collaborate in realtime
Post comments on notebooks
Run the Python kernel in two ways:
- In the browser using Portable Native Client (PNaCl)
- On your local system using the standard notebook server

Code and release:

Fully open source
Uses public Google APIs for Drive, Comments and Realtime Collaboration
Public but beta on the Chrome App Store

IPython Console

IPython started as a terminal based interactive console with tab completion, integrated help, plotting support, etc. This runs the Python based kernel in proces or out of process:



In [2]:

    
Image('images/ipython_console.png')









    Out[2]:

IPython Qt Console

The Qt Console Console provides another frontend with inline plotting and multiline code editing.



In [3]:

    
Image('images/ipython_qtconsole.png')









    Out[3]:

IPython.parallel

This same abstraction of running code remotely enables a high-level parallel computing framework.



In [4]:

    
Image("images/ParallelKernels.png")









    Out[4]:

On November 14, 2013, IBM announced that it was making its Jeopardy playing supercomputer, Watson, available to developers on the internet as a service. In the summer of 2013, researchers on the Watson team revealed that they were using the IPython Notebook and IPython.parallel to improve Watson's performance and capabilities.

Before:

8000 lines of Java code
2 minutes per analysis run.

After:

220 lines of Python code.
2 seconds per analysis run.