Skip to content

Instantly share code, notes, and snippets.

What would you like to do?
First draft of an ipython notebook html exporter for blogger
#!/usr/bin/env python
"""A really simple notebook to rst/html exporter.
./ file.ipynb
Produces 'file.rst' and 'file.html', along with auto-generated figure files
called nb_figure_NN.png.
NOTE: this *needs* docutils 0.9 for the code directive, which you'll
need to download and install yourself. docutils 0.8.x (the most recent one as of Jan/2012) is not sufficient.
import os
import subprocess
import sys
from IPython.nbformat import current as nbformat
from IPython.utils.text import wrap_paragraphs, indent
# Cell converters
def markdown_cell(cell):
"""convert a markdown cell to rst
Returns list."""
return [cell.source]
def rst_directive(directive, text):
return [directive, '', indent(text), '']
def code_cell(cell):
"""Convert a code cell to rst
Returns list."""
if not cell.input:
return []
lines = ['In[%s]:' % cell.prompt_number, '']
lines.extend(rst_directive('.. code:: python', cell.input))
for output in cell.outputs:
conv = converters[output.output_type]
return lines
# Converters for parts of a cell.
figures_counter = 1
def out_display(output):
"""convert display data from the output of a code cell to rst.
Returns list.
global figures_counter
lines = []
if 'png' in output:
fname = 'nb_figure_%s.png' % figures_counter
with open(fname, 'w') as f:
figures_counter += 1
lines.append('.. image:: %s' % fname)
return lines
def out_pyout(output):
"""convert pyout part of a code cell to rst
Returns list."""
lines = ['Out[%s]:' % output.prompt_number, '']
if 'latex' in output:
lines.extend(rst_directive('.. math::', output.latex))
if 'text' in output:
lines.extend(rst_directive('.. parsed-literal::', output.text))
return lines
converters = dict(code = code_cell,
markdown = markdown_cell,
pyout = out_pyout,
display_data = out_display,
def convert_notebook(nb):
lines = []
for cell in nb.worksheets[0].cells:
conv = converters[cell.cell_type]
return '\n'.join(lines)
def nb2rst(fname):
"Convert notebook to rst"
with open(fname) as f:
nb =, 'json')
rst = convert_notebook(nb)
newfname = os.path.splitext(fname)[0] + '.rst'
with open(newfname, 'w') as f:
return newfname
def rst2simplehtml(fname):
"""Convert a rst file to simplified html suitable for blogger.
This just runs rst2html with certain parameters to produce really simple
html and strips the document header, so the resulting file can be easily
pasted into a blogger edit window.
# This is the template for the rst2html call that produces the cleanest,
# simplest html I could find. This should help in making it easier to
# paste into the blogspot html window, though I'm still having problems
# with linebreaks there...
cmd_template = ("rst2html --link-stylesheet --no-xml-declaration "
"--no-generator --no-datestamp --no-source-link "
"--no-toc-backlinks --no-section-numbering "
"--strip-comments ")
cmd = "%s %s" % (cmd_template, fname)
proc = subprocess.Popen(cmd,
html, stderr = proc.communicate()
if stderr:
raise IOError(stderr)
# Make an iterator so breaking out holds state. Our implementation of
# searching for the html body below is basically a trivial little state
# machine, so we need this.
walker = iter(html.splitlines())
# Find start of main text, break out to then print until we find end /div.
# This may only work if there's a real title defined so we get a 'div class'
# tag, I haven't really tried.
for line in walker:
if line.startswith('<div class'):
newfname = os.path.splitext(fname)[0] + '.html'
with open(newfname, 'w') as f:
for line in walker:
if line.startswith('</div>'):
return newfname
def main(fname):
"""Convert a notebook to html in one step"""
if __name__ == '__main__':
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment