# -*- coding: utf-8 -*- # Copyright 2015 Mike Fährmann # # This program is free software; you can redistribute it and/or modify # it under the terms of the GNU General Public License version 2 as # published by the Free Software Foundation. """Extract images from http://www.hbrowse.com/""" from .common import Extractor, Message from .. import text class HbrowseExtractor(Extractor): category = "hbrowse" directory_fmt = ["{category}", "{gallery-id} {title}"] filename_fmt = "{category}_{gallery-id}_{num:>03}.{extension}" pattern = [r"(?:https?://)?(?:www\.)?hbrowse\.com/(\d+)/(c\d+)"] url_base = "http://www.hbrowse.com/thumbnails/" def __init__(self, match): Extractor.__init__(self) self.gid, self.chapter = match.groups() def items(self): page = self.request(self.url_base + self.gid + "/" + self.chapter).text data = self.get_job_metadata(page) yield Message.Version, 1 yield Message.Directory, data for num, url in enumerate(self.get_image_urls(page), 1): data["num"] = num text.nameext_from_url(url, data) yield Message.Url, url, data def get_job_metadata(self, page): """Collect metadata for extractor-job""" data = { "category": self.category, 'gallery-id': self.gid, 'chapter': int(self.chapter[1:]), } return text.extract_all(page, ( ('title' , '