derpibooru から、 JSON と画像を取得する。ループが 3 秒ぐらいになってる。負荷を掛けないように各自要調整。

derpibooru.py

# coding: utf8
from __future__ import division, print_function, unicode_literals

import json
import mimetypes
import os

import requests


class DerpibooruDocument(object):

    def __init__(self, json_dict):
        self._data = json_dict
        self._image_request = None

    def image_url(self):
        return 'http:' + self._data['image']

    def _fetch_image(self):
        if self._image_request is not None:
            return self._image_request
        r = requests.get(self.image_url(), stream=True)
        if r.status_code != 200:
            r.raise_for_status()
        self._image_request = r
        return r

    def guess_fetched_image_ext(self):
        r = self._fetch_image()
        ct = r.headers['content-type']
        ext = mimetypes.guess_extension(ct)
        if ext is not None:
            return ext
        path, ext = os.path.splitext(r.url)
        return ext

    def save_image(self, fo):
        r = self._fetch_image()
        for chunk in r.iter_content(1024):
            fo.write(chunk)

    @classmethod
    def from_number(cls, num):
        url = 'http://derpibooru.org/{}.json'.format(num)
        return cls.from_url(url)

    @classmethod
    def from_url(cls, url):
        j = requests.get(url).json()
        return cls(j)

    def to_file(self, fo):
        json.dump(self._data, fo)


if __name__ == '__main__':
    import time
    try:
        os.makedirs('images')
        os.makedirs('jsons')
    except OSError:
        pass
    for num in xrange(0, 599500):
        time.sleep(1)
        try:
            d = DerpibooruDocument.from_number(num)
        except Exception:
            continue
        if 'image' not in d._data:
            continue
        with open('jsons/{}.json'.format(num), 'wb') as fo:
            d.to_file(fo)
        try:
            ext = d.guess_fetched_image_ext()
        except Exception:
            continue
        with open('images/{}{}'.format(num, ext), 'wb') as fio:
            d.save_image(fio)
        time.sleep(2)