X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;f=youtube_dl%2FInfoExtractors.py;h=7221ae41e4796adc0830b4ae789556ec3d1e7d26;hb=392105265ccb3e155fe92b3c0ac15b2b4ff906d3;hp=0fc39163ee2a74131ed84442c68f388a72fad0f1;hpb=9ab3406ddb84a84a80f1be788af745d5a407ae23;p=youtube-dl diff --git a/youtube_dl/InfoExtractors.py b/youtube_dl/InfoExtractors.py index 0fc39163e..7221ae41e 100644 --- a/youtube_dl/InfoExtractors.py +++ b/youtube_dl/InfoExtractors.py @@ -13,6 +13,8 @@ import urllib import urllib2 import email.utils import xml.etree.ElementTree +import random +import math from urlparse import parse_qs try: @@ -181,8 +183,8 @@ class YoutubeIE(InfoExtractor): start = "%02i:%02i:%02i,%03i" %(start/(60*60), start/60%60, start%60, start%1*1000) end = "%02i:%02i:%02i,%03i" %(end/(60*60), end/60%60, end%60, end%1*1000) caption = unescapeHTML(caption) - caption = unescapeHTML(caption) # double cycle, inentional - srt += str(n) + '\n' + caption = unescapeHTML(caption) # double cycle, intentional + srt += str(n+1) + '\n' srt += start + ' --> ' + end + '\n' srt += caption + '\n\n' return srt @@ -311,6 +313,11 @@ class YoutubeIE(InfoExtractor): self._downloader.trouble(u'ERROR: "token" parameter not in video info for unknown reason') return + # Check for "rental" videos + if 'ypc_video_rental_bar_text' in video_info and 'author' not in video_info: + self._downloader.trouble(u'ERROR: "rental" videos not supported') + return + # Start extracting information self.report_information_extraction(video_id) @@ -361,7 +368,8 @@ class YoutubeIE(InfoExtractor): srt_list = urllib2.urlopen(request).read() except (urllib2.URLError, httplib.HTTPException, socket.error), err: raise Trouble(u'WARNING: unable to download video subtitles: %s' % str(err)) - srt_lang_list = re.findall(r'lang_code="([\w\-]+)"', srt_list) + srt_lang_list = re.findall(r'name="([^"]*)"[^>]+lang_code="([\w\-]+)"', srt_list) + srt_lang_list = dict((l[1], l[0]) for l in srt_lang_list) if not srt_lang_list: raise Trouble(u'WARNING: video has no closed captions') if self._downloader.params.get('subtitleslang', False): @@ -369,14 +377,16 @@ class YoutubeIE(InfoExtractor): elif 'en' in srt_lang_list: srt_lang = 'en' else: - srt_lang = srt_lang_list[0] + srt_lang = srt_lang_list.keys()[0] if not srt_lang in srt_lang_list: raise Trouble(u'WARNING: no closed captions found in the specified language') - request = urllib2.Request('http://video.google.com/timedtext?hl=en&lang=%s&v=%s' % (srt_lang, video_id)) + request = urllib2.Request('http://www.youtube.com/api/timedtext?lang=%s&name=%s&v=%s' % (srt_lang, srt_lang_list[srt_lang], video_id)) try: srt_xml = urllib2.urlopen(request).read() except (urllib2.URLError, httplib.HTTPException, socket.error), err: raise Trouble(u'WARNING: unable to download video subtitles: %s' % str(err)) + if not srt_xml: + raise Trouble(u'WARNING: unable to download video subtitles') video_subtitles = self._closed_captions_xml_to_srt(srt_xml.decode('utf-8')) except Trouble as trouble: self._downloader.trouble(trouble[0]) @@ -1227,7 +1237,7 @@ class YoutubeSearchIE(InfoExtractor): InfoExtractor.__init__(self, downloader) def report_download_page(self, query, pagenum): - """Report attempt to download playlist page with given number.""" + """Report attempt to download search page with given number.""" query = query.decode(preferredencoding()) self._downloader.to_screen(u'[youtube] query "%s": Downloading page %s' % (query, pagenum)) @@ -1463,8 +1473,8 @@ class YoutubePlaylistIE(InfoExtractor): _VALID_URL = r'(?:https?://)?(?:\w+\.)?youtube\.com/(?:(?:course|view_play_list|my_playlists|artist|playlist)\?.*?(p|a|list)=|user/.*?/user/|p/|user/.*?#[pg]/c/)(?:PL)?([0-9A-Za-z-_]+)(?:/.*?/([0-9A-Za-z_-]+))?.*' _TEMPLATE_URL = 'http://www.youtube.com/%s?%s=%s&page=%s&gl=US&hl=en' - _VIDEO_INDICATOR_TEMPLATE = r'/watch\?v=(.+?)&list=PL%s&' - _MORE_PAGES_INDICATOR = r'(?m)>\s*Next\s*' + _VIDEO_INDICATOR_TEMPLATE = r'/watch\?v=(.+?)&list=(PL)?%s&' + _MORE_PAGES_INDICATOR = r'yt-uix-pager-next' IE_NAME = u'youtube:playlist' def __init__(self, downloader=None): @@ -1614,6 +1624,98 @@ class YoutubeUserIE(InfoExtractor): self._downloader.download(['http://www.youtube.com/watch?v=%s' % video_id]) +class BlipTVUserIE(InfoExtractor): + """Information Extractor for blip.tv users.""" + + _VALID_URL = r'(?:(?:(?:https?://)?(?:\w+\.)?blip\.tv/)|bliptvuser:)([^/]+)/*$' + _PAGE_SIZE = 12 + IE_NAME = u'blip.tv:user' + + def __init__(self, downloader=None): + InfoExtractor.__init__(self, downloader) + + def report_download_page(self, username, pagenum): + """Report attempt to download user page.""" + self._downloader.to_screen(u'[%s] user %s: Downloading video ids from page %d' % + (self.IE_NAME, username, pagenum)) + + def _real_extract(self, url): + # Extract username + mobj = re.match(self._VALID_URL, url) + if mobj is None: + self._downloader.trouble(u'ERROR: invalid url: %s' % url) + return + + username = mobj.group(1) + + page_base = 'http://m.blip.tv/pr/show_get_full_episode_list?users_id=%s&lite=0&esi=1' + + request = urllib2.Request(url) + + try: + page = urllib2.urlopen(request).read().decode('utf-8') + mobj = re.search(r'data-users-id="([^"]+)"', page) + page_base = page_base % mobj.group(1) + except (urllib2.URLError, httplib.HTTPException, socket.error), err: + self._downloader.trouble(u'ERROR: unable to download webpage: %s' % str(err)) + return + + + # Download video ids using BlipTV Ajax calls. Result size per + # query is limited (currently to 12 videos) so we need to query + # page by page until there are no video ids - it means we got + # all of them. + + video_ids = [] + pagenum = 1 + + while True: + self.report_download_page(username, pagenum) + + request = urllib2.Request( page_base + "&page=" + str(pagenum) ) + + try: + page = urllib2.urlopen(request).read().decode('utf-8') + except (urllib2.URLError, httplib.HTTPException, socket.error), err: + self._downloader.trouble(u'ERROR: unable to download webpage: %s' % str(err)) + return + + # Extract video identifiers + ids_in_page = [] + + for mobj in re.finditer(r'href="/([^"]+)"', page): + if mobj.group(1) not in ids_in_page: + ids_in_page.append(unescapeHTML(mobj.group(1))) + + video_ids.extend(ids_in_page) + + # A little optimization - if current page is not + # "full", ie. does not contain PAGE_SIZE video ids then + # we can assume that this page is the last one - there + # are no more ids on further pages - no need to query + # again. + + if len(ids_in_page) < self._PAGE_SIZE: + break + + pagenum += 1 + + all_ids_count = len(video_ids) + playliststart = self._downloader.params.get('playliststart', 1) - 1 + playlistend = self._downloader.params.get('playlistend', -1) + + if playlistend == -1: + video_ids = video_ids[playliststart:] + else: + video_ids = video_ids[playliststart:playlistend] + + self._downloader.to_screen(u"[%s] user %s: Collected %d video ids (downloading %d of them)" % + (self.IE_NAME, username, all_ids_count, len(video_ids))) + + for video_id in video_ids: + self._downloader.download([u'http://blip.tv/'+video_id]) + + class DepositFilesIE(InfoExtractor): """Information extractor for depositfiles.com""" @@ -1912,7 +2014,7 @@ class BlipTVIE(InfoExtractor): else: cchar = '?' json_url = url + cchar + 'skin=json&version=2&no_wrap=1' - request = urllib2.Request(json_url) + request = urllib2.Request(json_url.encode('utf-8')) self.report_extraction(mobj.group(1)) info = None try: @@ -1970,6 +2072,7 @@ class BlipTVIE(InfoExtractor): self._downloader.trouble(u'ERROR: unable to parse video information: %s' % repr(err)) return + std_headers['User-Agent'] = 'iTunes/10.6.1' return [info] @@ -2188,12 +2291,14 @@ class EscapistIE(InfoExtractor): self.report_extraction(showName) try: - webPageBytes = urllib2.urlopen(url).read() + webPage = urllib2.urlopen(url) + webPageBytes = webPage.read() + m = re.match(r'text/html; charset="?([^"]+)"?', webPage.headers['Content-Type']) + webPage = webPageBytes.decode(m.group(1) if m else 'utf-8') except (urllib2.URLError, httplib.HTTPException, socket.error), err: self._downloader.trouble(u'ERROR: unable to download webpage: ' + unicode(err)) return - webPage = webPageBytes.decode('utf-8') descMatch = re.search('[A-Za-z0-9]+)\.html' + IE_NAME = u'Youku' + + def __init__(self, downloader=None): + InfoExtractor.__init__(self, downloader) + + def report_download_webpage(self, file_id): + """Report webpage download.""" + self._downloader.to_screen(u'[Youku] %s: Downloading webpage' % file_id) + + def report_extraction(self, file_id): + """Report information extraction.""" + self._downloader.to_screen(u'[Youku] %s: Extracting information' % file_id) + + def _gen_sid(self): + nowTime = int(time.time() * 1000) + random1 = random.randint(1000,1998) + random2 = random.randint(1000,9999) + + return "%d%d%d" %(nowTime,random1,random2) + + def _get_file_ID_mix_string(self, seed): + mixed = [] + source = list("abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ/\:._-1234567890") + seed = float(seed) + for i in range(len(source)): + seed = (seed * 211 + 30031 ) % 65536 + index = math.floor(seed / 65536 * len(source) ) + mixed.append(source[int(index)]) + source.remove(source[int(index)]) + #return ''.join(mixed) + return mixed + + + def _get_file_id(self, fileId, seed): + mixed = self._get_file_ID_mix_string(seed) + ids = fileId.split('*') + realId = [] + for ch in ids: + if ch is not '': + realId.append(mixed[int(ch)]) + return ''.join(realId) + + def _gen_key(self, key1, key2): + pass + + + + + def _real_extract(self, url): + mobj = re.match(self._VALID_URL, url) + if mobj is None: + self._downloader.trouble(u'ERROR: invalid URL: %s' % url) + return + video_id = mobj.group('ID') + + info_url = 'http://v.youku.com/player/getPlayList/VideoIDS/' + video_id +<<<<<<< HEAD + print info_url +======= +>>>>>>> 51661d86005ffbdd4debd051cd9683cf5f5c2fe9 + + request = urllib2.Request(info_url, None, std_headers) + try: + self.report_download_webpage(video_id) + jsondata = urllib2.urlopen(request).read() + except (urllib2.URLError, httplib.HTTPException, socket.error) as err: + self._downloader.trouble(u'ERROR: Unable to retrieve video webpage: %s' % str(err)) + return + + self.report_extraction(video_id) + try: + config = json.loads(jsondata) + + video_title = config['data'][0]['title'] + seed = config['data'][0]['seed'] + + format = self._downloader.params.get('format', None) + supported_format = config['data'][0]['streamfileids'].keys() + + if format is None or format == 'best': + if 'hd2' in supported_format: + format = 'hd2' + else: + format = 'flv' + ext = u'flv' + elif format == 'worst': + format = 'mp4' + ext = u'mp4' + else: + format = 'flv' + ext = u'flv' + + + fileid = config['data'][0]['streamfileids'][format] + seg_number = len(config['data'][0]['segs'][format]) + + keys=[] + for i in xrange(seg_number): + keys.append(config['data'][0]['segs'][format][i]['k']) + + #TODO check error + #youku only could be viewed from mainland china + except: + self._downloader.trouble(u'ERROR: unable to extract info section') + return + + files_info=[] + sid = self._gen_sid() + fileid = self._get_file_id(fileid, seed) + + #column 8,9 of fileid represent the segment number + #fileid[7:9] should be changed + for index, key in enumerate(keys): + + temp_fileid = '%s%02X%s' % (fileid[0:8], index, fileid[10:]) + download_url = 'http://f.youku.com/player/getFlvPath/sid/%s_%02X/st/flv/fileid/%s?k=%s' % (sid, index, temp_fileid, key) + print download_url + info = { + 'id': '%s_part%02d' % (video_id, index), + 'url': download_url, + 'uploader': None, + 'title': video_title, + 'ext': ext, + 'format': u'NA' + } + files_info.append(info) + + return files_info