From df5374743615fe54178b3942b115e7168ce0cd97 Mon Sep 17 00:00:00 2001 From: dst Date: Thu, 12 Dec 2013 15:13:45 +0700 Subject: [PATCH 1/4] [channel9] Initial implementation (#1885) --- youtube_dl/extractor/channel9.py | 351 +++++++++++++++++++++++++++++++ 1 file changed, 351 insertions(+) create mode 100644 youtube_dl/extractor/channel9.py diff --git a/youtube_dl/extractor/channel9.py b/youtube_dl/extractor/channel9.py new file mode 100644 index 0000000000..016c4497a3 --- /dev/null +++ b/youtube_dl/extractor/channel9.py @@ -0,0 +1,351 @@ +# encoding: utf-8 + +import re + +from .common import InfoExtractor +from ..utils import ( + format_bytes, + ExtractorError, +) + +class Channel9IE(InfoExtractor): + ''' + Common extractor for channel9.msdn.com. + + The type of provided URL (video or playlist) is determined according to + meta Search.PageType from web page HTML rather than URL itself, as it is + not always possible to do. + ''' + IE_DESC = u'Channel 9' + IE_NAME = u'channel9' + _VALID_URL = r'^https?://(?:www\.)?channel9\.msdn\.com/(?P.+)/?' + + _TESTS = [ + { + u'url': u'http://channel9.msdn.com/Events/TechEd/Australia/2013/KOS002', + u'file': u'Events_TechEd_Australia_2013_KOS002.mp4', + u'md5': u'bbd75296ba47916b754e73c3a4bbdf10', + u'info_dict': { + u'title': u'Developer Kick-Off Session: Stuff We Love', + u'description': u'md5:c08d72240b7c87fcecafe2692f80e35f', + u'duration': 4576, + u'thumbnail': u'http://media.ch9.ms/ch9/9d51/03902f2d-fc97-4d3c-b195-0bfe15a19d51/KOS002_220.jpg', + u'session_code': u'KOS002', + u'session_day': u'Day 1', + u'session_room': u'Arena 1A', + u'session_speakers': [ u'Ed Blankenship', u'Andrew Coates', u'Brady Gaster', u'Patrick Klug', u'Mads Kristensen' ], + }, + }, + { + u'url': u'http://channel9.msdn.com/posts/Self-service-BI-with-Power-BI-nuclear-testing', + u'file': u'posts_Self-service-BI-with-Power-BI-nuclear-testing.mp4', + u'md5': u'b43ee4529d111bc37ba7ee4f34813e68', + u'info_dict': { + u'title': u'Self-service BI with Power BI - nuclear testing', + u'description': u'md5:a6d5cfd9ee46d1851cf6e40ea61cfc10', + u'duration': 1540, + u'thumbnail': u'http://media.ch9.ms/ch9/87e1/0300391f-a455-4c72-bec3-4422f19287e1/selfservicenuk_512.jpg', + u'authors': [ u'Mike Wilmot' ], + }, + } + ] + + _RSS_URL = 'http://channel9.msdn.com/%s/RSS' + _EXTRACT_ENTRY_ITEMS_FROM_RSS = False + + # Sorted by quality + _known_formats = ['MP3', 'MP4', 'Mid Quality WMV', 'Mid Quality MP4', 'High Quality WMV', 'High Quality MP4'] + + def _restore_bytes(self, formatted_size): + if not formatted_size: + return 0 + m = re.match(r'^(?P\d+(?:\.\d+)?)\s+(?P[a-zA-Z]+)', formatted_size) + if not m: + return 0 + units = m.group('units') + try: + exponent = [u'B', u'KB', u'MB', u'GB', u'TB', u'PB', u'EB', u'ZB', u'YB'].index(units.upper()) + except ValueError: + return 0 + size = float(m.group('size')) + return int(size * (1024 ** exponent)) + + def _formats_from_html(self, html): + FORMAT_REGEX = r''' + (?x) + [^"]+)">(?P[^<]+)\s* + \((?P[^\)]+)\)\s* + (?:\s* +

File\s+size

\s*(?P.*?)\s* + )? # File size part may be missing + ''' + # Extract known formats + formats = [{'url': x.group('url'), + 'format_id': x.group('quality'), + 'format_note': x.group('note'), + 'format': '%s (%s)' % (x.group('quality'), x.group('note')), + 'filesize': self._restore_bytes(x.group('filesize')), # File size is approximate + } for x in list(re.finditer(FORMAT_REGEX, html)) if x.group('quality') in self._known_formats] + # Sort according to known formats list + formats.sort(key=lambda fmt: self._known_formats.index(fmt['format_id'])) + return formats + + def _formats_from_rss_item(self, item): + + def process_formats(elem): + formats = [] + for media_content in elem.findall('./{http://search.yahoo.com/mrss/}content'): + url = media_content.attrib['url'] + # Ignore unrelated media + if url.endswith('.ism/manifest'): + continue + format_note = media_content.attrib['type'] + filesize = int(media_content.attrib['fileSize']) + formats.append({'url': url, + 'format_note': format_note, + 'format': '%s %s' % (format_note, format_bytes(filesize)), + 'filesize': filesize, + }) + return formats + + formats = [] + + for media_group in item.findall('./{http://search.yahoo.com/mrss/}group'): + formats.extend(process_formats(media_group)) + + # Sometimes there are no media:groups in item, but there is media:content + # right in item (usually when there is the only media source) + formats.extend(process_formats(item)) + + # Sort by file size + formats.sort(key=lambda fmt: fmt['filesize']) + return formats + + def _extract_title(self, html): + title = self._html_search_meta(u'title', html, u'title') + if title is None: + title = self._og_search_title(html) + TITLE_SUFFIX = u' (Channel 9)' + if title is not None and title.endswith(TITLE_SUFFIX): + title = title[:-len(TITLE_SUFFIX)] + return title + + def _extract_description(self, html): + DESCRIPTION_REGEX = r'''(?sx) + \s* + \s* + (?P.+?)\s* + \s* + + ''' + m = re.search(DESCRIPTION_REGEX, html) + if m is not None: + return m.group('description') + return self._html_search_meta(u'description', html, u'description') + + def _extract_duration(self, html): + m = re.search(r'data-video_duration="(?P\d{2}):(?P\d{2}):(?P\d{2})"', html) + return ((int(m.group('hours')) * 60 * 60) + (int(m.group('minutes')) * 60) + int(m.group('seconds'))) if m else None + + def _extract_slides(self, html): + m = re.search(r'Slides', html) + return m.group('slidesurl') if m is not None else None + + def _extract_zip(self, html): + m = re.search(r'Zip', html) + return m.group('zipurl') if m is not None else None + + def _extract_avg_rating(self, html): + m = re.search(r'

Avg Rating: (?P[^<]+)

', html) + return float(m.group('avgrating')) if m is not None else 0 + + def _extract_rating_count(self, html): + m = re.search(r'
\((?P[^<]+)\)
', html) + return int(self._fix_count(m.group('ratingcount'))) if m is not None else 0 + + def _extract_view_count(self, html): + m = re.search(r'
  • \s*(?P[^<]+) Views\s*
    • ', html) + return int(self._fix_count(m.group('viewcount'))) if m is not None else 0 + + def _extract_comment_count(self, html): + m = re.search(r'
  • \s*\s*(?P[^<]+) Comments\s*\s*
    • ', html) + return int(self._fix_count(m.group('commentcount'))) if m is not None else 0 + + def _fix_count(self, count): + return int(str(count).replace(',', '')) if count is not None else None + + def _extract_authors(self, html): + m = re.search(r'(?s)
  • (.*?)
    • ', html) + if m is None: + return None + return re.findall(r'([^<]+)', m.group(1)) + + def _extract_session_code(self, html): + m = re.search(r'
  • \s*(?P.+?)\s*
    • ', html) + return m.group('code') if m is not None else None + + def _extract_session_day(self, html): + m = re.search(r'
  • \s*(?P[^<]+)\s*
    • ', html) + return m.group('day') if m is not None else None + + def _extract_session_room(self, html): + m = re.search(r'
  • \s*(?P.+?)\s*
    • ', html) + return m.group('room') if m is not None else None + + def _extract_session_speakers(self, html): + return re.findall(r'([^<]+)', html) + + def _extract_content(self, html, content_path): + # Look for downloadable content + formats = self._formats_from_html(html) + slides = self._extract_slides(html) + zip_ = self._extract_zip(html) + + # Nothing to download + if len(formats) == 0 and slides is None and zip_ is None: + self._downloader.report_warning(u'None of recording, slides or zip are available for %s' % content_path) + return + + # Extract meta + title = self._extract_title(html) + description = self._extract_description(html) + thumbnail = self._og_search_thumbnail(html) + duration = self._extract_duration(html) + avg_rating = self._extract_avg_rating(html) + rating_count = self._extract_rating_count(html) + view_count = self._extract_view_count(html) + comment_count = self._extract_comment_count(html) + + common = {'_type': 'video', + 'id': content_path, + 'description': description, + 'thumbnail': thumbnail, + 'duration': duration, + 'avg_rating': avg_rating, + 'rating_count': rating_count, + 'view_count': view_count, + 'comment_count': comment_count, + } + + result = [] + + if slides is not None: + d = common.copy() + d.update({ 'title': title + '-Slides', 'url': slides }) + result.append(d) + + if zip_ is not None: + d = common.copy() + d.update({ 'title': title + '-Zip', 'url': zip_ }) + result.append(d) + + if len(formats) > 0: + d = common.copy() + d.update({ 'title': title, 'formats': formats }) + result.append(d) + + return result + + def _extract_entry_item(self, html, content_path): + contents = self._extract_content(html, content_path) + if contents is None: + return contents + + authors = self._extract_authors(html) + + for content in contents: + content['authors'] = authors + + return contents + + def _extract_session(self, html, content_path): + contents = self._extract_content(html, content_path) + if contents is None: + return contents + + session_meta = {'session_code': self._extract_session_code(html), + 'session_day': self._extract_session_day(html), + 'session_room': self._extract_session_room(html), + 'session_speakers': self._extract_session_speakers(html), + } + + for content in contents: + content.update(session_meta) + + return contents + + def _extract_content_rss(self, rss): + ''' + Extracts links to entry items right out of RSS feed. + This approach is faster than extracting from web pages + one by one, but suffers from some problems. + Pros: + - no need to download additional pages + - provides more media links + - accurate file size + Cons: + - fewer meta data provided + - links to media files have no appropriate data that may be used as format_id + - RSS does not contain links to presentation materials (slides, zip) + ''' + entries = [] + for item in rss.findall('./channel/item'): + url = item.find('./link').text + video_id = url.split('/')[-1] + formats = self._formats_from_rss_item(item) + + if len(formats) == 0: + self._downloader.report_warning(u'The recording for session %s is not yet available' % video_id) + continue + + title = item.find('./title').text + description = item.find('./description').text + + thumbnail = item.find('./{http://search.yahoo.com/mrss/}thumbnail').text + + duration_e = item.find('./{http://www.itunes.com/dtds/podcast-1.0.dtd}duration') + duration = duration_e.text if duration_e is not None else 0 + + speakers_e = item.find('./{http://purl.org/dc/elements/1.1/}creator') + speakers = speakers_e.text.split(', ') if speakers_e is not None and speakers_e.text else [] + + entries.append({'_type': 'video', + 'id': video_id, + 'formats': formats, + 'title': title, + 'description': description, + 'thumbnail': thumbnail, + 'duration': duration, + 'session_speakers': speakers, + }) + return entries + + def _extract_list(self, content_path): + rss = self._download_xml(self._RSS_URL % content_path, content_path, u'Downloading RSS') + if self._EXTRACT_ENTRY_ITEMS_FROM_RSS: + return self._extract_content_rss(rss) + else: + entries = [self.url_result(session_url.text, 'Channel9') + for session_url in rss.findall('./channel/item/link')] + title_text = rss.find('./channel/title').text + return self.playlist_result(entries, content_path, title_text) + + def _real_extract(self, url): + mobj = re.match(self._VALID_URL, url) + content_path = mobj.group('contentpath') + + webpage = self._download_webpage(url, content_path, u'Downloading web page') + + page_type_m = re.search(r'', webpage) + if page_type_m is None: + raise ExtractorError(u'Search.PageType not found, don\'t know how to process this page', expected=True) + + page_type = page_type_m.group('pagetype') + if page_type == 'List': # List page, may contain list of 'item'-like objects + return self._extract_list(content_path) + elif page_type == 'Entry.Item': # Any 'item'-like page, may contain downloadable content + return self._extract_entry_item(webpage, content_path) + elif page_type == 'Session': # Event session page, may contain downloadable content + return self._extract_session(webpage, content_path) + else: + raise ExtractorError(u'Unexpected Search.PageType %s' % page_type, expected=True) \ No newline at end of file From 4d2ebb6bd771edaa26ebb5e76f9ee8e880a4d152 Mon Sep 17 00:00:00 2001 From: dst Date: Thu, 12 Dec 2013 15:19:23 +0700 Subject: [PATCH 2/4] [channel9] Cleanup --- youtube_dl/extractor/channel9.py | 94 ++------------------------------ 1 file changed, 5 insertions(+), 89 deletions(-) diff --git a/youtube_dl/extractor/channel9.py b/youtube_dl/extractor/channel9.py index 016c4497a3..b4b586c122 100644 --- a/youtube_dl/extractor/channel9.py +++ b/youtube_dl/extractor/channel9.py @@ -3,10 +3,7 @@ import re from .common import InfoExtractor -from ..utils import ( - format_bytes, - ExtractorError, -) +from ..utils import ExtractorError class Channel9IE(InfoExtractor): ''' @@ -51,7 +48,6 @@ class Channel9IE(InfoExtractor): ] _RSS_URL = 'http://channel9.msdn.com/%s/RSS' - _EXTRACT_ENTRY_ITEMS_FROM_RSS = False # Sorted by quality _known_formats = ['MP3', 'MP4', 'Mid Quality WMV', 'Mid Quality MP4', 'High Quality WMV', 'High Quality MP4'] @@ -90,37 +86,6 @@ def _formats_from_html(self, html): formats.sort(key=lambda fmt: self._known_formats.index(fmt['format_id'])) return formats - def _formats_from_rss_item(self, item): - - def process_formats(elem): - formats = [] - for media_content in elem.findall('./{http://search.yahoo.com/mrss/}content'): - url = media_content.attrib['url'] - # Ignore unrelated media - if url.endswith('.ism/manifest'): - continue - format_note = media_content.attrib['type'] - filesize = int(media_content.attrib['fileSize']) - formats.append({'url': url, - 'format_note': format_note, - 'format': '%s %s' % (format_note, format_bytes(filesize)), - 'filesize': filesize, - }) - return formats - - formats = [] - - for media_group in item.findall('./{http://search.yahoo.com/mrss/}group'): - formats.extend(process_formats(media_group)) - - # Sometimes there are no media:groups in item, but there is media:content - # right in item (usually when there is the only media source) - formats.extend(process_formats(item)) - - # Sort by file size - formats.sort(key=lambda fmt: fmt['filesize']) - return formats - def _extract_title(self, html): title = self._html_search_meta(u'title', html, u'title') if title is None: @@ -274,61 +239,12 @@ def _extract_session(self, html, content_path): return contents - def _extract_content_rss(self, rss): - ''' - Extracts links to entry items right out of RSS feed. - This approach is faster than extracting from web pages - one by one, but suffers from some problems. - Pros: - - no need to download additional pages - - provides more media links - - accurate file size - Cons: - - fewer meta data provided - - links to media files have no appropriate data that may be used as format_id - - RSS does not contain links to presentation materials (slides, zip) - ''' - entries = [] - for item in rss.findall('./channel/item'): - url = item.find('./link').text - video_id = url.split('/')[-1] - formats = self._formats_from_rss_item(item) - - if len(formats) == 0: - self._downloader.report_warning(u'The recording for session %s is not yet available' % video_id) - continue - - title = item.find('./title').text - description = item.find('./description').text - - thumbnail = item.find('./{http://search.yahoo.com/mrss/}thumbnail').text - - duration_e = item.find('./{http://www.itunes.com/dtds/podcast-1.0.dtd}duration') - duration = duration_e.text if duration_e is not None else 0 - - speakers_e = item.find('./{http://purl.org/dc/elements/1.1/}creator') - speakers = speakers_e.text.split(', ') if speakers_e is not None and speakers_e.text else [] - - entries.append({'_type': 'video', - 'id': video_id, - 'formats': formats, - 'title': title, - 'description': description, - 'thumbnail': thumbnail, - 'duration': duration, - 'session_speakers': speakers, - }) - return entries - def _extract_list(self, content_path): rss = self._download_xml(self._RSS_URL % content_path, content_path, u'Downloading RSS') - if self._EXTRACT_ENTRY_ITEMS_FROM_RSS: - return self._extract_content_rss(rss) - else: - entries = [self.url_result(session_url.text, 'Channel9') - for session_url in rss.findall('./channel/item/link')] - title_text = rss.find('./channel/title').text - return self.playlist_result(entries, content_path, title_text) + entries = [self.url_result(session_url.text, 'Channel9') + for session_url in rss.findall('./channel/item/link')] + title_text = rss.find('./channel/title').text + return self.playlist_result(entries, content_path, title_text) def _real_extract(self, url): mobj = re.match(self._VALID_URL, url) From 211f555d4cb0bc142d5b9304a919e27aba237d40 Mon Sep 17 00:00:00 2001 From: dst Date: Thu, 12 Dec 2013 15:55:31 +0700 Subject: [PATCH 3/4] [channel9] Missing import in __init__ --- youtube_dl/extractor/__init__.py | 1 + 1 file changed, 1 insertion(+) diff --git a/youtube_dl/extractor/__init__.py b/youtube_dl/extractor/__init__.py index 1149dc1ec4..fb206a7425 100644 --- a/youtube_dl/extractor/__init__.py +++ b/youtube_dl/extractor/__init__.py @@ -20,6 +20,7 @@ from .c56 import C56IE from .canalplus import CanalplusIE from .canalc2 import Canalc2IE +from .channel9 import Channel9IE from .cinemassacre import CinemassacreIE from .clipfish import ClipfishIE from .clipsyndicate import ClipsyndicateIE From 9b17ba0fa512a9882a22bb2407ec433038bc8327 Mon Sep 17 00:00:00 2001 From: dst Date: Thu, 12 Dec 2013 16:10:17 +0700 Subject: [PATCH 4/4] [channel9] Fix test description md5 --- youtube_dl/extractor/channel9.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/youtube_dl/extractor/channel9.py b/youtube_dl/extractor/channel9.py index b4b586c122..ae70ea2296 100644 --- a/youtube_dl/extractor/channel9.py +++ b/youtube_dl/extractor/channel9.py @@ -39,7 +39,7 @@ class Channel9IE(InfoExtractor): u'md5': u'b43ee4529d111bc37ba7ee4f34813e68', u'info_dict': { u'title': u'Self-service BI with Power BI - nuclear testing', - u'description': u'md5:a6d5cfd9ee46d1851cf6e40ea61cfc10', + u'description': u'md5:d1e6ecaafa7fb52a2cacdf9599829f5b', u'duration': 1540, u'thumbnail': u'http://media.ch9.ms/ch9/87e1/0300391f-a455-4c72-bec3-4422f19287e1/selfservicenuk_512.jpg', u'authors': [ u'Mike Wilmot' ],