NbConvert provides a command line interface to convert from IPython notebooks to standard formats.

• ReST
• Markdown
• HTML
• Python script
• Reveal
• LaTeX
• Sphinx

As these functions mature, they will be merged into IPython.

To install the necessary dependencies on a Linux machine:

pip install jinja2 
pip install markdown 
curl http://docutils.svn.sourceforge.net/viewvc/docutils/trunk/docutils/?view=tar > docutils.tgz 
pip install -U docutils.tgz 
pip install pygments 
sudo easy_install -U sphinx 
sudo apt-get install texlive-full 
sudo apt-get install pandoc

If you're on a Mac, the last two commands are executed in a different manner (since apt-get doesn't exist.)

Install PanDoc via the installer http://code.google.com/p/pandoc/downloads/list
Install MacTex via the .pkg http://www.tug.org/mactex/

Now, to export a notebook you can call

python nbconvert.py sphinx_howto book.ipynb

Where book.ipynb is the name of the notebook you'd like to convert and sphinx_howto is the output template name. See /nbconvert/templates/ and /nbconvert/templates/sphinx/ for more. NOTE: the template extension should not be included in the template name.

This will create a file.ext (converted file) and /book_files/ directory with all of the output figures.

If you want to compile a PDF from LaTeX output, move the file.tex (conversion results) into the directory with all the figures /book_files/. The You can then run

cd book_files       
PdfLatex file.tex

If you want to convert a notebook to a reveal-powered slideshow, you have to call

python nbconvert.py reveal your_slideshow.ipynb

and then, you will get a your_slideshow.reveal.html file.

But the slideshow will not be rendered correctly if you do not download and unzip the reveal.js library in the same folder where your_slideshow.reveal.html file is located.

Note: You can get the reveal.js library from this link: https://github.com/hakimel/reveal.js/archive/2.4.0.zip

Finally, to see the slideshow, just serve it

python -m SimpleHTTPServer 8000

and point your browser to http://localhost:8000.

Please try to run the tests to avoid regression when committing a patch, and create new tests when adding features. :

$ pip install nose
$ nosetests