html2ipynb.py

from bs4 import BeautifulSoup
import json


def strip_empty_lines(code):
    return '\n'.join([
        line for line in code.split('\n')
        if line.strip() != ''])


def convert_html(html):
    soup = BeautifulSoup(html, 'lxml')
    notebook = {'nbformat': 4, 'nbformat_minor': 1, 'cells': [], 'metadata': {}}
    for el in soup.findAll(['div', 'p', 'figure']):
        if el.name == 'div' and 'code-sample' in el.attrs['class']:
            notebook['cells'].append({
                'metadata': {},
                'outputs': [],
                'source': [strip_empty_lines(el.get_text())],
                'execution_count': None,
                'cell_type': 'code'
            })
        elif el.name == 'p':
            notebook['cells'].append({
                'metadata': {},
                'source': [str(el)],
                'cell_type': 'markdown'
            })
        elif el.name == 'figure':
            code = str(el)
            code = code.replace(
                'src="./',
                'src="https://third-bit.com/sdxpy/interpreter/')
            notebook['cells'].append({
                'metadata': {},
                'source': [code],
                'cell_type': 'markdown'
            })
    return notebook


html = open("ch3.html").read()
notebook = convert_html(html)
with open('notebook.ipynb', 'w') as output:
    output.write(json.dumps(notebook, indent=2))