[crunchyroll] extract playlist entries ids
[youtube-dl] / youtube_dl / extractor / crunchyroll.py
1 # coding: utf-8
2 from __future__ import unicode_literals
3
4 import re
5 import json
6 import base64
7 import zlib
8
9 from hashlib import sha1
10 from math import pow, sqrt, floor
11 from .common import InfoExtractor
12 from ..compat import (
13     compat_etree_fromstring,
14     compat_urllib_parse_urlencode,
15     compat_urllib_request,
16     compat_urlparse,
17 )
18 from ..utils import (
19     ExtractorError,
20     bytes_to_intlist,
21     intlist_to_bytes,
22     int_or_none,
23     lowercase_escape,
24     remove_end,
25     sanitized_Request,
26     unified_strdate,
27     urlencode_postdata,
28     xpath_text,
29     extract_attributes,
30 )
31 from ..aes import (
32     aes_cbc_decrypt,
33 )
34
35
36 class CrunchyrollBaseIE(InfoExtractor):
37     _LOGIN_URL = 'https://www.crunchyroll.com/login'
38     _LOGIN_FORM = 'login_form'
39     _NETRC_MACHINE = 'crunchyroll'
40
41     def _login(self):
42         (username, password) = self._get_login_info()
43         if username is None:
44             return
45
46         login_page = self._download_webpage(
47             self._LOGIN_URL, None, 'Downloading login page')
48
49         def is_logged(webpage):
50             return '<title>Redirecting' in webpage
51
52         # Already logged in
53         if is_logged(login_page):
54             return
55
56         login_form_str = self._search_regex(
57             r'(?P<form><form[^>]+?id=(["\'])%s\2[^>]*>)' % self._LOGIN_FORM,
58             login_page, 'login form', group='form')
59
60         post_url = extract_attributes(login_form_str).get('action')
61         if not post_url:
62             post_url = self._LOGIN_URL
63         elif not post_url.startswith('http'):
64             post_url = compat_urlparse.urljoin(self._LOGIN_URL, post_url)
65
66         login_form = self._form_hidden_inputs(self._LOGIN_FORM, login_page)
67
68         login_form.update({
69             'login_form[name]': username,
70             'login_form[password]': password,
71         })
72
73         response = self._download_webpage(
74             post_url, None, 'Logging in', 'Wrong login info',
75             data=urlencode_postdata(login_form),
76             headers={'Content-Type': 'application/x-www-form-urlencoded'})
77
78         # Successful login
79         if is_logged(response):
80             return
81
82         error = self._html_search_regex(
83             '(?s)<ul[^>]+class=["\']messages["\'][^>]*>(.+?)</ul>',
84             response, 'error message', default=None)
85         if error:
86             raise ExtractorError('Unable to login: %s' % error, expected=True)
87
88         raise ExtractorError('Unable to log in')
89
90     def _real_initialize(self):
91         self._login()
92
93     def _download_webpage(self, url_or_request, *args, **kwargs):
94         request = (url_or_request if isinstance(url_or_request, compat_urllib_request.Request)
95                    else sanitized_Request(url_or_request))
96         # Accept-Language must be set explicitly to accept any language to avoid issues
97         # similar to https://github.com/rg3/youtube-dl/issues/6797.
98         # Along with IP address Crunchyroll uses Accept-Language to guess whether georestriction
99         # should be imposed or not (from what I can see it just takes the first language
100         # ignoring the priority and requires it to correspond the IP). By the way this causes
101         # Crunchyroll to not work in georestriction cases in some browsers that don't place
102         # the locale lang first in header. However allowing any language seems to workaround the issue.
103         request.add_header('Accept-Language', '*')
104         return super(CrunchyrollBaseIE, self)._download_webpage(request, *args, **kwargs)
105
106     @staticmethod
107     def _add_skip_wall(url):
108         parsed_url = compat_urlparse.urlparse(url)
109         qs = compat_urlparse.parse_qs(parsed_url.query)
110         # Always force skip_wall to bypass maturity wall, namely 18+ confirmation message:
111         # > This content may be inappropriate for some people.
112         # > Are you sure you want to continue?
113         # since it's not disabled by default in crunchyroll account's settings.
114         # See https://github.com/rg3/youtube-dl/issues/7202.
115         qs['skip_wall'] = ['1']
116         return compat_urlparse.urlunparse(
117             parsed_url._replace(query=compat_urllib_parse_urlencode(qs, True)))
118
119
120 class CrunchyrollIE(CrunchyrollBaseIE):
121     _VALID_URL = r'https?://(?:(?P<prefix>www|m)\.)?(?P<url>crunchyroll\.(?:com|fr)/(?:media(?:-|/\?id=)|[^/]*/[^/?&]*?)(?P<video_id>[0-9]+))(?:[/?&]|$)'
122     _TESTS = [{
123         'url': 'http://www.crunchyroll.com/wanna-be-the-strongest-in-the-world/episode-1-an-idol-wrestler-is-born-645513',
124         'info_dict': {
125             'id': '645513',
126             'ext': 'mp4',
127             'title': 'Wanna be the Strongest in the World Episode 1 – An Idol-Wrestler is Born!',
128             'description': 'md5:2d17137920c64f2f49981a7797d275ef',
129             'thumbnail': 'http://img1.ak.crunchyroll.com/i/spire1-tmb/20c6b5e10f1a47b10516877d3c039cae1380951166_full.jpg',
130             'uploader': 'Yomiuri Telecasting Corporation (YTV)',
131             'upload_date': '20131013',
132             'url': 're:(?!.*&amp)',
133         },
134         'params': {
135             # rtmp
136             'skip_download': True,
137         },
138     }, {
139         'url': 'http://www.crunchyroll.com/media-589804/culture-japan-1',
140         'info_dict': {
141             'id': '589804',
142             'ext': 'flv',
143             'title': 'Culture Japan Episode 1 – Rebuilding Japan after the 3.11',
144             'description': 'md5:2fbc01f90b87e8e9137296f37b461c12',
145             'thumbnail': r're:^https?://.*\.jpg$',
146             'uploader': 'Danny Choo Network',
147             'upload_date': '20120213',
148         },
149         'params': {
150             # rtmp
151             'skip_download': True,
152         },
153         'skip': 'Video gone',
154     }, {
155         'url': 'http://www.crunchyroll.com/rezero-starting-life-in-another-world-/episode-5-the-morning-of-our-promise-is-still-distant-702409',
156         'info_dict': {
157             'id': '702409',
158             'ext': 'mp4',
159             'title': 'Re:ZERO -Starting Life in Another World- Episode 5 – The Morning of Our Promise Is Still Distant',
160             'description': 'md5:97664de1ab24bbf77a9c01918cb7dca9',
161             'thumbnail': r're:^https?://.*\.jpg$',
162             'uploader': 'TV TOKYO',
163             'upload_date': '20160508',
164         },
165         'params': {
166             # m3u8 download
167             'skip_download': True,
168         },
169     }, {
170         'url': 'http://www.crunchyroll.com/konosuba-gods-blessing-on-this-wonderful-world/episode-1-give-me-deliverance-from-this-judicial-injustice-727589',
171         'info_dict': {
172             'id': '727589',
173             'ext': 'mp4',
174             'title': "KONOSUBA -God's blessing on this wonderful world! 2 Episode 1 – Give Me Deliverance from this Judicial Injustice!",
175             'description': 'md5:cbcf05e528124b0f3a0a419fc805ea7d',
176             'thumbnail': r're:^https?://.*\.jpg$',
177             'uploader': 'Kadokawa Pictures Inc.',
178             'upload_date': '20170118',
179             'series': "KONOSUBA -God's blessing on this wonderful world!",
180             'season_number': 2,
181             'episode': 'Give Me Deliverance from this Judicial Injustice!',
182             'episode_number': 1,
183         },
184         'params': {
185             # m3u8 download
186             'skip_download': True,
187         },
188     }, {
189         'url': 'http://www.crunchyroll.fr/girl-friend-beta/episode-11-goodbye-la-mode-661697',
190         'only_matching': True,
191     }, {
192         # geo-restricted (US), 18+ maturity wall, non-premium available
193         'url': 'http://www.crunchyroll.com/cosplay-complex-ova/episode-1-the-birth-of-the-cosplay-club-565617',
194         'only_matching': True,
195     }, {
196         # A description with double quotes
197         'url': 'http://www.crunchyroll.com/11eyes/episode-1-piros-jszaka-red-night-535080',
198         'info_dict': {
199             'id': '535080',
200             'ext': 'mp4',
201             'title': '11eyes Episode 1 – Piros éjszaka - Red Night',
202             'description': 'Kakeru and Yuka are thrown into an alternate nightmarish world they call "Red Night".',
203             'uploader': 'Marvelous AQL Inc.',
204             'upload_date': '20091021',
205         },
206         'params': {
207             # Just test metadata extraction
208             'skip_download': True,
209         },
210     }]
211
212     _FORMAT_IDS = {
213         '360': ('60', '106'),
214         '480': ('61', '106'),
215         '720': ('62', '106'),
216         '1080': ('80', '108'),
217     }
218
219     def _decrypt_subtitles(self, data, iv, id):
220         data = bytes_to_intlist(base64.b64decode(data.encode('utf-8')))
221         iv = bytes_to_intlist(base64.b64decode(iv.encode('utf-8')))
222         id = int(id)
223
224         def obfuscate_key_aux(count, modulo, start):
225             output = list(start)
226             for _ in range(count):
227                 output.append(output[-1] + output[-2])
228             # cut off start values
229             output = output[2:]
230             output = list(map(lambda x: x % modulo + 33, output))
231             return output
232
233         def obfuscate_key(key):
234             num1 = int(floor(pow(2, 25) * sqrt(6.9)))
235             num2 = (num1 ^ key) << 5
236             num3 = key ^ num1
237             num4 = num3 ^ (num3 >> 3) ^ num2
238             prefix = intlist_to_bytes(obfuscate_key_aux(20, 97, (1, 2)))
239             shaHash = bytes_to_intlist(sha1(prefix + str(num4).encode('ascii')).digest())
240             # Extend 160 Bit hash to 256 Bit
241             return shaHash + [0] * 12
242
243         key = obfuscate_key(id)
244
245         decrypted_data = intlist_to_bytes(aes_cbc_decrypt(data, key, iv))
246         return zlib.decompress(decrypted_data)
247
248     def _convert_subtitles_to_srt(self, sub_root):
249         output = ''
250
251         for i, event in enumerate(sub_root.findall('./events/event'), 1):
252             start = event.attrib['start'].replace('.', ',')
253             end = event.attrib['end'].replace('.', ',')
254             text = event.attrib['text'].replace('\\N', '\n')
255             output += '%d\n%s --> %s\n%s\n\n' % (i, start, end, text)
256         return output
257
258     def _convert_subtitles_to_ass(self, sub_root):
259         output = ''
260
261         def ass_bool(strvalue):
262             assvalue = '0'
263             if strvalue == '1':
264                 assvalue = '-1'
265             return assvalue
266
267         output = '[Script Info]\n'
268         output += 'Title: %s\n' % sub_root.attrib['title']
269         output += 'ScriptType: v4.00+\n'
270         output += 'WrapStyle: %s\n' % sub_root.attrib['wrap_style']
271         output += 'PlayResX: %s\n' % sub_root.attrib['play_res_x']
272         output += 'PlayResY: %s\n' % sub_root.attrib['play_res_y']
273         output += """
274 [V4+ Styles]
275 Format: Name, Fontname, Fontsize, PrimaryColour, SecondaryColour, OutlineColour, BackColour, Bold, Italic, Underline, StrikeOut, ScaleX, ScaleY, Spacing, Angle, BorderStyle, Outline, Shadow, Alignment, MarginL, MarginR, MarginV, Encoding
276 """
277         for style in sub_root.findall('./styles/style'):
278             output += 'Style: ' + style.attrib['name']
279             output += ',' + style.attrib['font_name']
280             output += ',' + style.attrib['font_size']
281             output += ',' + style.attrib['primary_colour']
282             output += ',' + style.attrib['secondary_colour']
283             output += ',' + style.attrib['outline_colour']
284             output += ',' + style.attrib['back_colour']
285             output += ',' + ass_bool(style.attrib['bold'])
286             output += ',' + ass_bool(style.attrib['italic'])
287             output += ',' + ass_bool(style.attrib['underline'])
288             output += ',' + ass_bool(style.attrib['strikeout'])
289             output += ',' + style.attrib['scale_x']
290             output += ',' + style.attrib['scale_y']
291             output += ',' + style.attrib['spacing']
292             output += ',' + style.attrib['angle']
293             output += ',' + style.attrib['border_style']
294             output += ',' + style.attrib['outline']
295             output += ',' + style.attrib['shadow']
296             output += ',' + style.attrib['alignment']
297             output += ',' + style.attrib['margin_l']
298             output += ',' + style.attrib['margin_r']
299             output += ',' + style.attrib['margin_v']
300             output += ',' + style.attrib['encoding']
301             output += '\n'
302
303         output += """
304 [Events]
305 Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text
306 """
307         for event in sub_root.findall('./events/event'):
308             output += 'Dialogue: 0'
309             output += ',' + event.attrib['start']
310             output += ',' + event.attrib['end']
311             output += ',' + event.attrib['style']
312             output += ',' + event.attrib['name']
313             output += ',' + event.attrib['margin_l']
314             output += ',' + event.attrib['margin_r']
315             output += ',' + event.attrib['margin_v']
316             output += ',' + event.attrib['effect']
317             output += ',' + event.attrib['text']
318             output += '\n'
319
320         return output
321
322     def _extract_subtitles(self, subtitle):
323         sub_root = compat_etree_fromstring(subtitle)
324         return [{
325             'ext': 'srt',
326             'data': self._convert_subtitles_to_srt(sub_root),
327         }, {
328             'ext': 'ass',
329             'data': self._convert_subtitles_to_ass(sub_root),
330         }]
331
332     def _get_subtitles(self, video_id, webpage):
333         subtitles = {}
334         for sub_id, sub_name in re.findall(r'\bssid=([0-9]+)"[^>]+?\btitle="([^"]+)', webpage):
335             sub_page = self._download_webpage(
336                 'http://www.crunchyroll.com/xml/?req=RpcApiSubtitle_GetXml&subtitle_script_id=' + sub_id,
337                 video_id, note='Downloading subtitles for ' + sub_name)
338             id = self._search_regex(r'id=\'([0-9]+)', sub_page, 'subtitle_id', fatal=False)
339             iv = self._search_regex(r'<iv>([^<]+)', sub_page, 'subtitle_iv', fatal=False)
340             data = self._search_regex(r'<data>([^<]+)', sub_page, 'subtitle_data', fatal=False)
341             if not id or not iv or not data:
342                 continue
343             subtitle = self._decrypt_subtitles(data, iv, id).decode('utf-8')
344             lang_code = self._search_regex(r'lang_code=["\']([^"\']+)', subtitle, 'subtitle_lang_code', fatal=False)
345             if not lang_code:
346                 continue
347             subtitles[lang_code] = self._extract_subtitles(subtitle)
348         return subtitles
349
350     def _real_extract(self, url):
351         mobj = re.match(self._VALID_URL, url)
352         video_id = mobj.group('video_id')
353
354         if mobj.group('prefix') == 'm':
355             mobile_webpage = self._download_webpage(url, video_id, 'Downloading mobile webpage')
356             webpage_url = self._search_regex(r'<link rel="canonical" href="([^"]+)" />', mobile_webpage, 'webpage_url')
357         else:
358             webpage_url = 'http://www.' + mobj.group('url')
359
360         webpage = self._download_webpage(self._add_skip_wall(webpage_url), video_id, 'Downloading webpage')
361         note_m = self._html_search_regex(
362             r'<div class="showmedia-trailer-notice">(.+?)</div>',
363             webpage, 'trailer-notice', default='')
364         if note_m:
365             raise ExtractorError(note_m)
366
367         mobj = re.search(r'Page\.messaging_box_controller\.addItems\(\[(?P<msg>{.+?})\]\)', webpage)
368         if mobj:
369             msg = json.loads(mobj.group('msg'))
370             if msg.get('type') == 'error':
371                 raise ExtractorError('crunchyroll returned error: %s' % msg['message_body'], expected=True)
372
373         if 'To view this, please log in to verify you are 18 or older.' in webpage:
374             self.raise_login_required()
375
376         video_title = self._html_search_regex(
377             r'(?s)<h1[^>]*>((?:(?!<h1).)*?<span[^>]+itemprop=["\']title["\'][^>]*>(?:(?!<h1).)+?)</h1>',
378             webpage, 'video_title')
379         video_title = re.sub(r' {2,}', ' ', video_title)
380         video_description = self._parse_json(self._html_search_regex(
381             r'<script[^>]*>\s*.+?\[media_id=%s\].+?({.+?"description"\s*:.+?})\);' % video_id,
382             webpage, 'description', default='{}'), video_id).get('description')
383         if video_description:
384             video_description = lowercase_escape(video_description.replace(r'\r\n', '\n'))
385         video_upload_date = self._html_search_regex(
386             [r'<div>Availability for free users:(.+?)</div>', r'<div>[^<>]+<span>\s*(.+?\d{4})\s*</span></div>'],
387             webpage, 'video_upload_date', fatal=False, flags=re.DOTALL)
388         if video_upload_date:
389             video_upload_date = unified_strdate(video_upload_date)
390         video_uploader = self._html_search_regex(
391             r'<a[^>]+href="/publisher/[^"]+"[^>]*>([^<]+)</a>', webpage,
392             'video_uploader', fatal=False)
393
394         available_fmts = []
395         for a, fmt in re.findall(r'(<a[^>]+token=["\']showmedia\.([0-9]{3,4})p["\'][^>]+>)', webpage):
396             attrs = extract_attributes(a)
397             href = attrs.get('href')
398             if href and '/freetrial' in href:
399                 continue
400             available_fmts.append(fmt)
401         if not available_fmts:
402             for p in (r'token=["\']showmedia\.([0-9]{3,4})p"', r'showmedia\.([0-9]{3,4})p'):
403                 available_fmts = re.findall(p, webpage)
404                 if available_fmts:
405                     break
406         video_encode_ids = []
407         formats = []
408         for fmt in available_fmts:
409             stream_quality, stream_format = self._FORMAT_IDS[fmt]
410             video_format = fmt + 'p'
411             streamdata_req = sanitized_Request(
412                 'http://www.crunchyroll.com/xml/?req=RpcApiVideoPlayer_GetStandardConfig&media_id=%s&video_format=%s&video_quality=%s'
413                 % (video_id, stream_format, stream_quality),
414                 compat_urllib_parse_urlencode({'current_page': url}).encode('utf-8'))
415             streamdata_req.add_header('Content-Type', 'application/x-www-form-urlencoded')
416             streamdata = self._download_xml(
417                 streamdata_req, video_id,
418                 note='Downloading media info for %s' % video_format)
419             stream_info = streamdata.find('./{default}preload/stream_info')
420             video_encode_id = xpath_text(stream_info, './video_encode_id')
421             if video_encode_id in video_encode_ids:
422                 continue
423             video_encode_ids.append(video_encode_id)
424
425             video_file = xpath_text(stream_info, './file')
426             if not video_file:
427                 continue
428             if video_file.startswith('http'):
429                 formats.extend(self._extract_m3u8_formats(
430                     video_file, video_id, 'mp4', entry_protocol='m3u8_native',
431                     m3u8_id='hls', fatal=False))
432                 continue
433
434             video_url = xpath_text(stream_info, './host')
435             if not video_url:
436                 continue
437             metadata = stream_info.find('./metadata')
438             format_info = {
439                 'format': video_format,
440                 'format_id': video_format,
441                 'height': int_or_none(xpath_text(metadata, './height')),
442                 'width': int_or_none(xpath_text(metadata, './width')),
443             }
444
445             if '.fplive.net/' in video_url:
446                 video_url = re.sub(r'^rtmpe?://', 'http://', video_url.strip())
447                 parsed_video_url = compat_urlparse.urlparse(video_url)
448                 direct_video_url = compat_urlparse.urlunparse(parsed_video_url._replace(
449                     netloc='v.lvlt.crcdn.net',
450                     path='%s/%s' % (remove_end(parsed_video_url.path, '/'), video_file.split(':')[-1])))
451                 if self._is_valid_url(direct_video_url, video_id, video_format):
452                     format_info.update({
453                         'url': direct_video_url,
454                     })
455                     formats.append(format_info)
456                     continue
457
458             format_info.update({
459                 'url': video_url,
460                 'play_path': video_file,
461                 'ext': 'flv',
462             })
463             formats.append(format_info)
464         self._sort_formats(formats)
465
466         metadata = self._download_xml(
467             'http://www.crunchyroll.com/xml', video_id,
468             note='Downloading media info', query={
469                 'req': 'RpcApiVideoPlayer_GetMediaMetadata',
470                 'media_id': video_id,
471             })
472
473         subtitles = self.extract_subtitles(video_id, webpage)
474
475         # webpage provide more accurate data than series_title from XML
476         series = self._html_search_regex(
477             r'id=["\']showmedia_about_episode_num[^>]+>\s*<a[^>]+>([^<]+)',
478             webpage, 'series', default=xpath_text(metadata, 'series_title'))
479
480         episode = xpath_text(metadata, 'episode_title')
481         episode_number = int_or_none(xpath_text(metadata, 'episode_number'))
482
483         season_number = int_or_none(self._search_regex(
484             r'(?s)<h4[^>]+id=["\']showmedia_about_episode_num[^>]+>.+?</h4>\s*<h4>\s*Season (\d+)',
485             webpage, 'season number', default=None))
486
487         return {
488             'id': video_id,
489             'title': video_title,
490             'description': video_description,
491             'thumbnail': xpath_text(metadata, 'episode_image_url'),
492             'uploader': video_uploader,
493             'upload_date': video_upload_date,
494             'series': series,
495             'season_number': season_number,
496             'episode': episode,
497             'episode_number': episode_number,
498             'subtitles': subtitles,
499             'formats': formats,
500         }
501
502
503 class CrunchyrollShowPlaylistIE(CrunchyrollBaseIE):
504     IE_NAME = 'crunchyroll:playlist'
505     _VALID_URL = r'https?://(?:(?P<prefix>www|m)\.)?(?P<url>crunchyroll\.com/(?!(?:news|anime-news|library|forum|launchcalendar|lineup|store|comics|freetrial|login))(?P<id>[\w\-]+))/?(?:\?|$)'
506
507     _TESTS = [{
508         'url': 'http://www.crunchyroll.com/a-bridge-to-the-starry-skies-hoshizora-e-kakaru-hashi',
509         'info_dict': {
510             'id': 'a-bridge-to-the-starry-skies-hoshizora-e-kakaru-hashi',
511             'title': 'A Bridge to the Starry Skies - Hoshizora e Kakaru Hashi'
512         },
513         'playlist_count': 13,
514     }, {
515         # geo-restricted (US), 18+ maturity wall, non-premium available
516         'url': 'http://www.crunchyroll.com/cosplay-complex-ova',
517         'info_dict': {
518             'id': 'cosplay-complex-ova',
519             'title': 'Cosplay Complex OVA'
520         },
521         'playlist_count': 3,
522         'skip': 'Georestricted',
523     }, {
524         # geo-restricted (US), 18+ maturity wall, non-premium will be available since 2015.11.14
525         'url': 'http://www.crunchyroll.com/ladies-versus-butlers?skip_wall=1',
526         'only_matching': True,
527     }]
528
529     def _real_extract(self, url):
530         show_id = self._match_id(url)
531
532         webpage = self._download_webpage(self._add_skip_wall(url), show_id)
533         title = self._html_search_regex(
534             r'(?s)<h1[^>]*>\s*<span itemprop="name">(.*?)</span>',
535             webpage, 'title')
536         episode_paths = re.findall(
537             r'(?s)<li id="showview_videos_media_(\d+)"[^>]+>.*?<a href="([^"]+)"',
538             webpage)
539         entries = [
540             self.url_result('http://www.crunchyroll.com' + ep, 'Crunchyroll', ep_id)
541             for ep_id, ep in episode_paths
542         ]
543         entries.reverse()
544
545         return {
546             '_type': 'playlist',
547             'id': show_id,
548             'title': title,
549             'entries': entries,
550         }