X-Git-Url: https://git.rapsys.eu/youtubedl/blobdiff_plain/ced7488f6d3a519b2c1b1cbd31048743fb8285bd..87a0165ca7e39af4dacb7ec637063b2cd35ae40b:/youtube_dl/extractor/iwara.py diff --git a/youtube_dl/extractor/iwara.py b/youtube_dl/extractor/iwara.py new file mode 100644 index 0000000..8d7e7f4 --- /dev/null +++ b/youtube_dl/extractor/iwara.py @@ -0,0 +1,77 @@ +# coding: utf-8 +from __future__ import unicode_literals + +from .common import InfoExtractor +from ..compat import compat_urllib_parse_urlparse +from ..utils import remove_end + + +class IwaraIE(InfoExtractor): + _VALID_URL = r'https?://(?:www\.|ecchi\.)?iwara\.tv/videos/(?P[a-zA-Z0-9]+)' + _TESTS = [{ + 'url': 'http://iwara.tv/videos/amVwUl1EHpAD9RD', + 'md5': '1d53866b2c514b23ed69e4352fdc9839', + 'info_dict': { + 'id': 'amVwUl1EHpAD9RD', + 'ext': 'mp4', + 'title': '【MMD R-18】ガールフレンド carry_me_off', + 'age_limit': 18, + }, + }, { + 'url': 'http://ecchi.iwara.tv/videos/Vb4yf2yZspkzkBO', + 'md5': '7e5f1f359cd51a027ba4a7b7710a50f0', + 'info_dict': { + 'id': '0B1LvuHnL-sRFNXB1WHNqbGw4SXc', + 'ext': 'mp4', + 'title': '[3D Hentai] Kyonyu Ã\x97 Genkai Ã\x97 Emaki Shinobi Girls.mp4', + 'age_limit': 18, + }, + 'add_ie': ['GoogleDrive'], + }, { + 'url': 'http://www.iwara.tv/videos/nawkaumd6ilezzgq', + 'md5': '1d85f1e5217d2791626cff5ec83bb189', + 'info_dict': { + 'id': '6liAP9s2Ojc', + 'ext': 'mp4', + 'age_limit': 0, + 'title': '[MMD] Do It Again Ver.2 [1080p 60FPS] (Motion,Camera,Wav+DL)', + 'description': 'md5:590c12c0df1443d833fbebe05da8c47a', + 'upload_date': '20160910', + 'uploader': 'aMMDsork', + 'uploader_id': 'UCVOFyOSCyFkXTYYHITtqB7A', + }, + 'add_ie': ['Youtube'], + }] + + def _real_extract(self, url): + video_id = self._match_id(url) + + webpage, urlh = self._download_webpage_handle(url, video_id) + + hostname = compat_urllib_parse_urlparse(urlh.geturl()).hostname + # ecchi is 'sexy' in Japanese + age_limit = 18 if hostname.split('.')[0] == 'ecchi' else 0 + + entries = self._parse_html5_media_entries(url, webpage, video_id) + + if not entries: + iframe_url = self._html_search_regex( + r']+src=([\'"])(?P[^\'"]+)\1', + webpage, 'iframe URL', group='url') + return { + '_type': 'url_transparent', + 'url': iframe_url, + 'age_limit': age_limit, + } + + title = remove_end(self._html_search_regex( + r'([^<]+)', webpage, 'title'), ' | Iwara') + + info_dict = entries[0] + info_dict.update({ + 'id': video_id, + 'title': title, + 'age_limit': age_limit, + }) + + return info_dict