euler/ipython/publish.py

#!/usr/bin/env python

import jinja2
import os
import subprocess
import shutil
from operator import itemgetter
from collections import namedtuple
from os.path import getmtime
from bs4 import BeautifulSoup
import json


def extract_metadata(ipynb_file):
    Metadata = namedtuple('Metadata', ['tags', 'completion_date'])
    with open(ipynb_file, 'r') as f:
        j = json.load(f)
    try:
        tags = j['metadata']['tags']
        completion_date = j['metadata']['completion_date']
    except KeyError:
        raise Exception('Failed extracting meta from {}.'.format(ipynb_file))
    return Metadata(tags, completion_date)


def file_name_to_solution(name):
    Solution = namedtuple('Solution', ["number", "ipynb", "html", "name", "metadata"])
    number = int(name.replace("EulerProblem", "").replace(".ipynb", ""))
    ipynb = name
    metadata =  extract_metadata(ipynb)
    html = name.replace(".ipynb", ".html")
    name = name.replace("EulerProblem", "Problem ").replace(".ipynb", "")
    return Solution(number, ipynb, html, name, metadata)


def get_solution_list(directory="./"):
    l = [file_name_to_solution(f) for f in os.listdir(directory)
        if f.endswith(".ipynb")
        if f.startswith("EulerProblem")
        if not f.endswith("000.ipynb")]
    l.sort(key=itemgetter(0))
    return l


def render_solutions(solutions):
    loader = jinja2.FileSystemLoader(searchpath="./")
    env = jinja2.Environment(loader=loader)
    template = env.get_template("template.html")
    d = {"solutions": solutions}
    with open("html/index.html", 'w') as f:
        f.write(template.render(**d))


def convert_solutions_to_html(solutions):
    for s in solutions:
        html = os.path.join("html", s.html)
        if not os.path.isfile(html) or getmtime(html) < getmtime(s.ipynb):
            args = ["jupyter-nbconvert", s.ipynb, "--output-dir=html"]
            subprocess.call(args)
            post_process(html)


def post_process(html):

    def tag(t, string="", children=[],  **attrs):
        t = BeautifulSoup("", 'html.parser').new_tag(t, **attrs)
        if string:
            t.string = string
        for c in children:
            t.append(c)
        return t

    with open(html, 'r') as f:
        soup = BeautifulSoup(f.read(), 'html.parser')

    soup_h1 = soup.find("h1")
    soup_p = tag("p", children=[tag("a", href="/euler", string="Back to overview.")])
    soup_h1.insert_after(soup_p)

    with open(html, 'w') as f:
        f.write(str(soup))


def ship_to_failx():
    args = ["rsync", "-r", "html/", "failx@felixm.de:/home/failx/html/euler"]
    subprocess.call(args)


if __name__ == "__main__":
    solutions = get_solution_list()
    convert_solutions_to_html(solutions)
    render_solutions(solutions)
    ship_to_failx()