Get revisions for a Wiki article, in this case, 'Once Upon a Time in Hollywood'

wiki_revs.py

"""
I got really interested in how many edits 'Once Upon a Time in Hollywood' got after
reading about the contentious issue of whether or not the Plot section should include
information deemed to be a spoiler, and the kinda-heated debate on the Talk page.

This gist also serves as my own documentation for using the WikiMedia query/revisions API.
"""
import json
import pickle
import pprint
from urllib import parse, request

from collections import Counter

revs_pkl = 'revs.pkl'

# Once Upon a Time in Hollywood
pageid = '56717294'

def download():
    base_path = 'https://en.wikipedia.org/w/api.php?%s'
    base_params = {
        'action':'query',
        'prop':'revisions',
        'pageids':pageid,
        'rvprop':'timestamp',
        'rvlimit':'max',
        'format':'json'
    }

    i = 1
    revs = []
    path = base_path % parse.urlencode(base_params)
    while True:
        print(i)
        with request.urlopen(path) as f:
            d = json.loads(f.read())

        revs.extend([x['timestamp'] for x in d['query']['pages'][pageid]['revisions']])

        if 'continue' in d:
            base_params['rvcontinue'] = d['continue']['rvcontinue']
        else:
            break

        path = base_path % parse.urlencode(base_params)
        i += 1

    with open(revs_pkl, 'w+b') as f:
        f.write(pickle.dumps(revs))

try:
    print('Trying to read %s' % revs_pkl)
    with open(revs_pkl, 'r+b') as f:
        revs = pickle.loads(f.read())
except OSError:
    print('Couldn\'t find %s, downloading...' % revs_pkl)
    download()
    print('Trying to read %s' % revs_pkl)
    with open(revs_pkl, 'r+b') as f:
        revs = pickle.loads(f.read())

print('%d revisions' % len(revs))

# For an RFC 3339 datetime, use full date and just the hour
counter = Counter([x[:13] for x in revs])

print('10 edits/hour, or more')
pprint.pprint([x for x in counter.most_common() if x[1] >= 10])