Checks Brandon Sanderson’s website for book progress bar updates and tweets them.

sanderson-progress.py

import json
import os

from bs4 import BeautifulSoup
import requests
import twitter


TWITTER_CLIENT = twitter.Twitter(
        auth=twitter.OAuth('REDACTED',
                           'REDACTED',
                           'REDACTED', 'REDACTED'))


def get_progress():
    soup = BeautifulSoup(requests.get('http://brandonsanderson.com').text, 'html5lib')
    progressdiv =  soup.find('div', attrs={'class': 'vc_progress_bar'})
    books = [i.text for i in progressdiv.findAll('small', attrs={'class': 'vc_label'})]
    progress = {}
    for i, book in enumerate(books):
        s = book.split(' ')
        percent = s.pop()
        title = ' '.join(s)
        progress[title] = percent
    return progress


def print_progress(progress):
    for k, v in progress.items():
        print(f'{k}: {v}')


def save_progress(progress):
    with open('progress.json', 'w') as f:
        json.dump(progress, f)


def compare_progress(current):
    previous = json.load(open('progress.json'))
    if previous != current:
        for title, percent in previous.items():
            if title in current:
                if percent != current[title]:
                    report(f'"{title}" has progressed from {percent} to {current[title]}.')
            else:
                report(f'"{title}" was removed from the progress list at {percent}.')
        for title, percent in current.items():
            if title not in previous:
                report(f'"{title}" was added and is {percent} complete.')
        return False
    return True


def report(update):
    print(update)
    try:
        TWITTER_CLIENT.statuses.update(status=update)
    except twitter.api.TwitterHTTPError as err:
        print(err)


def main():
    os.chdir(os.path.dirname(os.path.realpath(__file__)))
    progress = get_progress()
    if not compare_progress(progress):
        with open('progress.json', 'w') as f:
            json.dump(progress, f)


if __name__ == '__main__':
    main()