Create .py and .html on save of Jupyter notebook

jupyter_notebook_config.py

import os
import re

from nbconvert.nbconvertapp import NbConvertApp
from nbconvert.postprocessors.base import PostProcessorBase


class CopyToSubfolderPostProcessor(PostProcessorBase):
    def __init__(self, subfolder=None):
        self.subfolder = subfolder
        super(CopyToSubfolderPostProcessor, self).__init__()

    def postprocess(self, input):
        """ Save converted file to a separate directory. """
        if self.subfolder is None:
            return

        dirname, filename = os.path.split(input)
        new_dir = os.path.join(dirname, self.subfolder)
        new_path = os.path.join(new_dir, filename)

        if not os.path.exists(new_dir):
            os.mkdir(new_dir)

        with open(input, 'r') as f:
            text = f.read()

        with open(new_path, 'w') as f:
            f.write(re.sub(r'\n#\sIn\[(([0-9]+)|(\s))\]:\n{2}', '', text))

        os.remove(input)


SAVE_PROGRESS_INDICATOR_FILE = '.ipynb_saveprogress'


def post_save(model, os_path, contents_manager):
    """post-save hook for converting notebooks to .py scripts and html
       in a separate folder with the same name
    """
    # only do this for notebooks
    if model['type'] != 'notebook':
        return

    # only do this if we've added the special indicator file to the working directory
    cwd = os.path.dirname(os_path)
    save_progress_indicator = os.path.join(cwd, SAVE_PROGRESS_INDICATOR_FILE)
    should_convert = os.path.exists(save_progress_indicator)

    if should_convert:
        d, fname = os.path.split(os_path)
        subfolder = os.path.splitext(fname)[0]

        converter = NbConvertApp()
        converter.postprocessor = CopyToSubfolderPostProcessor(subfolder=subfolder)
        converter.export_format = 'script'
        converter.initialize(argv=[])
        converter.notebooks = [os_path]
        converter.convert_notebooks()

c.FileContentsManager.post_save_hook = post_save

Hey Peter, what's the recommended way to use this file? Found the link from your slides, but unfortunately wasn't able to attend the accompanying talk.

from the looks of it, I think you would just add these contents to your regular ~/.jupyter/jupyter_notebook_config.py

Adding this to the end of the postprocess function (instead of lines 25 and 26) will remove the cell numbers from the .py file 🎉

with open(input, 'r') as f:
    text = f.read()

with open(new_path, 'w') as f:
    f.write(re.sub(r'\n#\sIn\[(([0-9]+)|(\s))\]:\n{2}', '', text))

os.remove(input)

Thanks @ejm714, I updated with that fix