X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;f=youtube_dl%2Fextractor%2Fbliptv.py;h=159ea39d9a788bc90304907f75f523a156cebce1;hb=cefcb9fde3bcacc6231f9bba7e025a3a9f21afc3;hp=5e33a69df42fcbaa1b17f1737d66f5841ca50318;hpb=304cbe981e9f3b62288abc54bd84b0eaef286795;p=youtube-dl diff --git a/youtube_dl/extractor/bliptv.py b/youtube_dl/extractor/bliptv.py index 5e33a69df..159ea39d9 100644 --- a/youtube_dl/extractor/bliptv.py +++ b/youtube_dl/extractor/bliptv.py @@ -1,6 +1,7 @@ +from __future__ import unicode_literals + import datetime import json -import os import re import socket @@ -21,29 +22,28 @@ from ..utils import ( class BlipTVIE(InfoExtractor): """Information extractor for blip.tv""" - _VALID_URL = r'^(?:https?://)?(?:\w+\.)?blip\.tv/((.+/)|(play/)|(api\.swf#))(.+)$' - _URL_EXT = r'^.*\.([a-z0-9]+)$' - IE_NAME = u'blip.tv' + _VALID_URL = r'^(?:https?://)?(?:www\.)?blip\.tv/((.+/)|(play/)|(api\.swf#))(.+)$' + IE_NAME = 'blip.tv' _TEST = { - u'url': u'http://blip.tv/cbr/cbr-exclusive-gotham-city-imposters-bats-vs-jokerz-short-3-5796352', - u'file': u'5779306.m4v', - u'md5': u'80baf1ec5c3d2019037c1c707d676b9f', - u'info_dict': { - u"upload_date": u"20111205", - u"description": u"md5:9bc31f227219cde65e47eeec8d2dc596", - u"uploader": u"Comic Book Resources - CBR TV", - u"title": u"CBR EXCLUSIVE: \"Gotham City Imposters\" Bats VS Jokerz Short 3" + 'url': 'http://blip.tv/cbr/cbr-exclusive-gotham-city-imposters-bats-vs-jokerz-short-3-5796352', + 'file': '5779306.mov', + 'md5': 'c6934ad0b6acf2bd920720ec888eb812', + 'info_dict': { + 'upload_date': '20111205', + 'description': 'md5:9bc31f227219cde65e47eeec8d2dc596', + 'uploader': 'Comic Book Resources - CBR TV', + 'title': 'CBR EXCLUSIVE: "Gotham City Imposters" Bats VS Jokerz Short 3', } } def report_direct_download(self, title): """Report information extraction.""" - self.to_screen(u'%s: Direct download detected' % title) + self.to_screen('%s: Direct download detected' % title) def _real_extract(self, url): mobj = re.match(self._VALID_URL, url) if mobj is None: - raise ExtractorError(u'Invalid URL: %s' % url) + raise ExtractorError('Invalid URL: %s' % url) # See https://github.com/rg3/youtube-dl/issues/857 api_mobj = re.match(r'http://a\.blip\.tv/api\.swf#(?P[\d\w]+)', url) @@ -58,7 +58,6 @@ class BlipTVIE(InfoExtractor): url = 'http://blip.tv/a/a-' + file_id return self._real_extract(url) - if '?' in url: cchar = '&' else: @@ -67,67 +66,55 @@ class BlipTVIE(InfoExtractor): request = compat_urllib_request.Request(json_url) request.add_header('User-Agent', 'iTunes/10.6.1') self.report_extraction(mobj.group(1)) - info = None urlh = self._request_webpage(request, None, False, - u'unable to download video info webpage') - if urlh.headers.get('Content-Type', '').startswith('video/'): # Direct download - basename = url.split('/')[-1] - title,ext = os.path.splitext(basename) - title = title.decode('UTF-8') - ext = ext.replace('.', '') - self.report_direct_download(title) - info = { - 'id': title, - 'url': url, - 'uploader': None, - 'upload_date': None, - 'title': title, - 'ext': ext, - 'urlhandle': urlh + 'unable to download video info webpage') + + try: + json_code_bytes = urlh.read() + json_code = json_code_bytes.decode('utf-8') + except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err: + raise ExtractorError('Unable to read video info webpage: %s' % compat_str(err)) + + try: + json_data = json.loads(json_code) + if 'Post' in json_data: + data = json_data['Post'] + else: + data = json_data + + upload_date = datetime.datetime.strptime(data['datestamp'], '%m-%d-%y %H:%M%p').strftime('%Y%m%d') + formats = [] + if 'additionalMedia' in data: + for f in sorted(data['additionalMedia'], key=lambda f: int(f['media_height'])): + if not int(f['media_width']): # filter m3u8 + continue + formats.append({ + 'url': f['url'], + 'format_id': f['role'], + 'width': int(f['media_width']), + 'height': int(f['media_height']), + }) + else: + formats.append({ + 'url': data['media']['url'], + 'width': int(data['media']['width']), + 'height': int(data['media']['height']), + }) + + self._sort_formats(formats) + + return { + 'id': compat_str(data['item_id']), + 'uploader': data['display_name'], + 'upload_date': upload_date, + 'title': data['title'], + 'thumbnail': data['thumbnailUrl'], + 'description': data['description'], + 'user_agent': 'iTunes/10.6.1', + 'formats': formats, } - if info is None: # Regular URL - try: - json_code_bytes = urlh.read() - json_code = json_code_bytes.decode('utf-8') - except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err: - raise ExtractorError(u'Unable to read video info webpage: %s' % compat_str(err)) - - try: - json_data = json.loads(json_code) - if 'Post' in json_data: - data = json_data['Post'] - else: - data = json_data - - upload_date = datetime.datetime.strptime(data['datestamp'], '%m-%d-%y %H:%M%p').strftime('%Y%m%d') - if 'additionalMedia' in data: - formats = sorted(data['additionalMedia'], key=lambda f: int(f['media_height'])) - best_format = formats[-1] - video_url = best_format['url'] - else: - video_url = data['media']['url'] - umobj = re.match(self._URL_EXT, video_url) - if umobj is None: - raise ValueError('Can not determine filename extension') - ext = umobj.group(1) - - info = { - 'id': compat_str(data['item_id']), - 'url': video_url, - 'uploader': data['display_name'], - 'upload_date': upload_date, - 'title': data['title'], - 'ext': ext, - 'format': data['media']['mimeType'], - 'thumbnail': data['thumbnailUrl'], - 'description': data['description'], - 'player_url': data['embedUrl'], - 'user_agent': 'iTunes/10.6.1', - } - except (ValueError,KeyError) as err: - raise ExtractorError(u'Unable to parse video information: %s' % repr(err)) - - return [info] + except (ValueError, KeyError) as err: + raise ExtractorError('Unable to parse video information: %s' % repr(err)) class BlipTVUserIE(InfoExtractor): @@ -135,19 +122,19 @@ class BlipTVUserIE(InfoExtractor): _VALID_URL = r'(?:(?:(?:https?://)?(?:\w+\.)?blip\.tv/)|bliptvuser:)([^/]+)/*$' _PAGE_SIZE = 12 - IE_NAME = u'blip.tv:user' + IE_NAME = 'blip.tv:user' def _real_extract(self, url): # Extract username mobj = re.match(self._VALID_URL, url) if mobj is None: - raise ExtractorError(u'Invalid URL: %s' % url) + raise ExtractorError('Invalid URL: %s' % url) username = mobj.group(1) page_base = 'http://m.blip.tv/pr/show_get_full_episode_list?users_id=%s&lite=0&esi=1' - page = self._download_webpage(url, username, u'Downloading user page') + page = self._download_webpage(url, username, 'Downloading user page') mobj = re.search(r'data-users-id="([^"]+)"', page) page_base = page_base % mobj.group(1) @@ -163,7 +150,7 @@ class BlipTVUserIE(InfoExtractor): while True: url = page_base + "&page=" + str(pagenum) page = self._download_webpage(url, username, - u'Downloading video ids from page %d' % pagenum) + 'Downloading video ids from page %d' % pagenum) # Extract video identifiers ids_in_page = [] @@ -185,6 +172,6 @@ class BlipTVUserIE(InfoExtractor): pagenum += 1 - urls = [u'http://blip.tv/%s' % video_id for video_id in video_ids] + urls = ['http://blip.tv/%s' % video_id for video_id in video_ids] url_entries = [self.url_result(vurl, 'BlipTV') for vurl in urls] return [self.playlist_result(url_entries, playlist_title = username)]