Find follower accounts that have profile pics with faces in them

.env

OAUTH_CONSUMER_KEY=
OAUTH_CONSUMER_SECRET=
OAUTH_TOKEN=
OAUTH_TOKEN_SECRET=
IMGIX_TOKEN=

requirements.txt

imgix==1.0.0
requests==2.9.1
requests-oauthlib==0.6.0
tablib==0.11.2
python-dotenv==0.3.0

twitter_face_detect.py

# encoding=utf-8

import sys
import shelve
import os
import json
from os.path import join, dirname

from clint.textui import progress
from dotenv import load_dotenv
load_dotenv(join(dirname(__file__), '.env'))
import requests
from requests_oauthlib import OAuth1Session
import imgix
import tablib


builder = imgix.UrlBuilder(
    "takumi-twitter-dp.imgix.net",
    sign_key=os.environ['IMGIX_TOKEN']
)


def chunks(l, n):
    """Yield successive n-sized chunks from l."""
    for i in xrange(0, len(l), n):
        yield l[i:i + n]

twitter = OAuth1Session(os.environ['OAUTH_CONSUMER_KEY'],
                        client_secret=os.environ['OAUTH_CONSUMER_SECRET'],
                        resource_owner_key=os.environ['OAUTH_TOKEN'],
                        resource_owner_secret=os.environ['OAUTH_TOKEN_SECRET'])

api_url = 'https://api.twitter.com/1.1'
url = api_url + '/users/lookup.json'


def main(path, size=None):
    # Split into chunks of 100
    book = tablib.Dataset(headers=['ID', 'Name', 'Handle', 'Bio', 'Image', 'Location', 'Verified', 'URL'])
    scratch = shelve.open('.scratch', writeback=True)
    scratch.setdefault('skippable', [])

    followers = map(str, json.load(sys.stdin)['data']['followers'])
    sys.stdin = open('/dev/tty')

    if os.path.exists(path):
        with open(path) as fp:
            book.load(fp.read())

    booked_ids = book['ID']

    if size is not None:
        followers = followers[:int(size)]

    for i, _followers in enumerate(progress.bar(list(chunks(followers, 100)))):
        _followers = [f for f in _followers if f not in (booked_ids + scratch['skippable'])]
        if not _followers:
            continue
        response = twitter.get(
            url + '?user_id=' + ','.join(_followers),
            headers={'content-type': 'application/json'},
        )
        for user in response.json():
            if user['id_str'] in (booked_ids + scratch['skippable']):
                continue
            imgix_url = builder.create_url(user['profile_image_url'], opts=dict(faces=1, fm='json'))
            imgix_response = requests.get(imgix_url)
            if imgix_response.status_code != 200:
                continue
            if imgix_response.json().get('Faces'):
                book.append((
                    user['id_str'],
                    user['name'],
                    user['screen_name'],
                    user['description'],
                    user['profile_image_url'],
                    user['location'],
                    user['verified'],
                    user['url'],
                ))
            else:
                scratch['skippable'].append(user['id_str'])

        with open(path, 'w') as fp:
            fp.write(book.csv)

        scratch.sync()

    scratch.close()


if __name__ == "__main__":
    from ipdb import launch_ipdb_on_exception
    with launch_ipdb_on_exception():
        main(*sys.argv[1:])