Run 2to3 on IPython notebooks

2to3_nb.py

#!/usr/bin/env python3
"""
To run: python3 nb2to3.py notebook-or-directory
"""
# Authors: Thomas Kluyver, Fernando Perez
# See: https://gist.github.com/takluyver/c8839593c615bb2f6e80

import argparse
import pathlib
from nbformat import read, write

import lib2to3
from lib2to3.refactor import RefactoringTool, get_fixers_from_package


def refactor_notebook_inplace(rt, path):
    
    def refactor_cell(src):
        #print('\n***SRC***\n', src)
        try:
            tree = rt.refactor_string(src+'\n', str(path) + '/cell-%d' % i)
        except (lib2to3.pgen2.parse.ParseError,
                lib2to3.pgen2.tokenize.TokenError):
            return src
        else:
            return str(tree)[:-1]

    
    print("Refactoring:", path)
    nb = read(str(path), as_version=4)
    
    # Run 2to3 on code
    for i, cell in enumerate(nb.cells, start=1):
        if cell.cell_type == 'code':
            if cell.execution_count in (' ', '*'):
                cell.execution_count = None

            if cell.source.startswith('%%'):
                # For cell magics, try to refactor the body, in case it's
                # valid python
                head, source = cell.source.split('\n', 1)
                cell.source = head + '\n' + refactor_cell(source)
            else:
                cell.source = refactor_cell(cell.source)
                   

    # Update notebook metadata
    nb.metadata.kernelspec = {
        'display_name': 'Python 3',
        'name': 'python3',
        'language': 'python',
    }
    if 'language_info' in nb.metadata:
        nb.metadata.language_info.codemirror_mode = {
            'name': 'ipython',
            'version': 3,
        }
        nb.metadata.language_info.pygments_lexer = 'ipython3'
        nb.metadata.language_info.pop('version', None)

    write(nb, str(path))

def main(argv=None):
    ap = argparse.ArgumentParser()
    ap.add_argument('path', type=pathlib.Path,
        help="Notebook or directory containing notebooks")
    
    options = ap.parse_args(argv)
    
    avail_fixes = set(get_fixers_from_package('lib2to3.fixes'))
    rt = RefactoringTool(avail_fixes)
    
    if options.path.is_dir():
        for nb_path in options.path.rglob('*.ipynb'):
            refactor_notebook_inplace(rt, nb_path)
    else:
        refactor_notebook_inplace(rt, options.path)

if __name__ == '__main__':
    main()

Traceback (most recent call last):
  File "2to3_nb.py", line 80, in <module>
    main()
  File "2to3_nb.py", line 75, in main
    refactor_notebook_inplace(rt, nb_path)
  File "2to3_nb.py", line 41, in refactor_notebook_inplace
    head, source = cell.source.split('\n', 1)
ValueError: not enough values to unpack (expected 2, got 1)

I think the offending bit of the ipynb file is this:

   "source": [
    "%%timeit p={\"$push\":{'events':event}}"
   ]

Catching the exception and ignoring looks to solve it (:

I used this to make an extension for Jupyter Notebooks that can convert cells (or whole notebooks) from python 2 to 3 with the click of a button. It has been added to jupyter_contrib_nbextensions. I mentioned your contribution in the readme.

Find it here.

Thanks. Very useful for me.

Great docs but where is the simplied, "pip install this", "conda install this" and you're good to go instructions?

Since @EWouters turned this script into an nbextension, you can refer to the jupyter-contrib-nbextensions docs.