[adobetv] improve extaction

- use OnDemandPagedList for list extractors
- reduce show extraction requests
- extract original video format and subtitles
- add support for adobe tv embeds

2 files changed, 167 insertions, 75 deletions

diff --git a/youtube_dl/extractor/adobetv.py b/youtube_dl/extractor/adobetv.py
index 008c98e51..80060f037 100644
--- a/youtube_dl/extractor/adobetv.py
+++ b/youtube_dl/extractor/adobetv.py
@@ -1,25 +1,119 @@
 from __future__ import unicode_literals
 
+import functools
 import re
 
 from .common import InfoExtractor
 from ..compat import compat_str
 from ..utils import (
-    parse_duration,
-    unified_strdate,
-    str_to_int,
-    int_or_none,
     float_or_none,
+    int_or_none,
     ISO639Utils,
-    determine_ext,
+    OnDemandPagedList,
+    parse_duration,
+    str_or_none,
+    str_to_int,
+    unified_strdate,
 )
 
 
 class AdobeTVBaseIE(InfoExtractor):
-    _API_BASE_URL = 'http://tv.adobe.com/api/v4/'
+    def _call_api(self, path, video_id, query, note=None):
+        return self._download_json(
+            'http://tv.adobe.com/api/v4/' + path,
+            video_id, note, query=query)['data']
+
+    def _parse_subtitles(self, video_data, url_key):
+        subtitles = {}
+        for translation in video_data.get('translations', []):
+            vtt_path = translation.get(url_key)
+            if not vtt_path:
+                continue
+            lang = translation.get('language_w3c') or ISO639Utils.long2short(translation['language_medium'])
+            subtitles.setdefault(lang, []).append({
+                'ext': 'vtt',
+                'url': vtt_path,
+            })
+        return subtitles
+
+    def _parse_video_data(self, video_data):
+        video_id = compat_str(video_data['id'])
+        title = video_data['title']
+
+        s3_extracted = False
+        formats = []
+        for source in video_data.get('videos', []):
+            source_url = source.get('url')
+            if not source_url:
+                continue
+            f = {
+                'format_id': source.get('quality_level'),
+                'fps': int_or_none(source.get('frame_rate')),
+                'height': int_or_none(source.get('height')),
+                'tbr': int_or_none(source.get('video_data_rate')),
+                'width': int_or_none(source.get('width')),
+                'url': source_url,
+            }
+            original_filename = source.get('original_filename')
+            if original_filename:
+                if not (f.get('height') and f.get('width')):
+                    mobj = re.search(r'_(\d+)x(\d+)', original_filename)
+                    if mobj:
+                        f.update({
+                            'height': int(mobj.group(2)),
+                            'width': int(mobj.group(1)),
+                        })
+                if original_filename.startswith('s3://') and not s3_extracted:
+                    formats.append({
+                        'format_id': 'original',
+                        'preference': 1,
+                        'url': original_filename.replace('s3://', 'https://s3.amazonaws.com/'),
+                    })
+                    s3_extracted = True
+            formats.append(f)
+        self._sort_formats(formats)
+
+        return {
+            'id': video_id,
+            'title': title,
+            'description': video_data.get('description'),
+            'thumbnail': video_data.get('thumbnail'),
+            'upload_date': unified_strdate(video_data.get('start_date')),
+            'duration': parse_duration(video_data.get('duration')),
+            'view_count': str_to_int(video_data.get('playcount')),
+            'formats': formats,
+            'subtitles': self._parse_subtitles(video_data, 'vtt'),
+        }
+
+
+class AdobeTVEmbedIE(AdobeTVBaseIE):
+    IE_NAME = 'adobetv:embed'
+    _VALID_URL = r'https?://tv\.adobe\.com/embed/\d+/(?P<id>\d+)'
+    _TEST = {
+        'url': 'https://tv.adobe.com/embed/22/4153',
+        'md5': 'c8c0461bf04d54574fc2b4d07ac6783a',
+        'info_dict': {
+            'id': '4153',
+            'ext': 'flv',
+            'title': 'Creating Graphics Optimized for BlackBerry',
+            'description': 'md5:eac6e8dced38bdaae51cd94447927459',
+            'thumbnail': r're:https?://.*\.jpg$',
+            'upload_date': '20091109',
+            'duration': 377,
+            'view_count': int,
+        },
+    }
+
+    def _real_extract(self, url):
+        video_id = self._match_id(url)
+
+        video_data = self._call_api(
+            'episode/' + video_id, video_id, {'disclosure': 'standard'})[0]
+        return self._parse_video_data(video_data)
 
 
 class AdobeTVIE(AdobeTVBaseIE):
+    IE_NAME = 'adobetv'
     _VALID_URL = r'https?://tv\.adobe\.com/(?:(?P<language>fr|de|es|jp)/)?watch/(?P<show_urlname>[^/]+)/(?P<id>[^/]+)'
 
     _TEST = {
@@ -42,45 +136,33 @@ class AdobeTVIE(AdobeTVBaseIE):
         if not language:
             language = 'en'
 
-        video_data = self._download_json(
-            self._API_BASE_URL + 'episode/get/?language=%s&show_urlname=%s&urlname=%s&disclosure=standard' % (language, show_urlname, urlname),
-            urlname)['data'][0]
-
-        formats = [{
-            'url': source['url'],
-            'format_id': source.get('quality_level') or source['url'].split('-')[-1].split('.')[0] or None,
-            'width': int_or_none(source.get('width')),
-            'height': int_or_none(source.get('height')),
-            'tbr': int_or_none(source.get('video_data_rate')),
-        } for source in video_data['videos']]
-        self._sort_formats(formats)
-
-        return {
-            'id': compat_str(video_data['id']),
-            'title': video_data['title'],
-            'description': video_data.get('description'),
-            'thumbnail': video_data.get('thumbnail'),
-            'upload_date': unified_strdate(video_data.get('start_date')),
-            'duration': parse_duration(video_data.get('duration')),
-            'view_count': str_to_int(video_data.get('playcount')),
-            'formats': formats,
-        }
+        video_data = self._call_api(
+            'episode/get', urlname, {
+                'disclosure': 'standard',
+                'language': language,
+                'show_urlname': show_urlname,
+                'urlname': urlname,
+            })[0]
+        return self._parse_video_data(video_data)
 
 
 class AdobeTVPlaylistBaseIE(AdobeTVBaseIE):
-    def _parse_page_data(self, page_data):
-        return [self.url_result(self._get_element_url(element_data)) for element_data in page_data]
+    _PAGE_SIZE = 25
+
+    def _fetch_page(self, display_id, query, page):
+        page += 1
+        query['page'] = page
+        for element_data in self._call_api(
+                self._RESOURCE, display_id, query, 'Download Page %d' % page):
+            yield self._process_data(element_data)
 
-    def _extract_playlist_entries(self, url, display_id):
-        page = self._download_json(url, display_id)
-        entries = self._parse_page_data(page['data'])
-        for page_num in range(2, page['paging']['pages'] + 1):
-            entries.extend(self._parse_page_data(
-                self._download_json(url + '&page=%d' % page_num, display_id)['data']))
-        return entries
+    def _extract_playlist_entries(self, display_id, query):
+        return OnDemandPagedList(functools.partial(
+            self._fetch_page, display_id, query), self._PAGE_SIZE)
 
 
 class AdobeTVShowIE(AdobeTVPlaylistBaseIE):
+    IE_NAME = 'adobetv:show'
     _VALID_URL = r'https?://tv\.adobe\.com/(?:(?P<language>fr|de|es|jp)/)?show/(?P<id>[^/]+)'
 
     _TEST = {
@@ -92,26 +174,31 @@ class AdobeTVShowIE(AdobeTVPlaylistBaseIE):
         },
         'playlist_mincount': 136,
     }
-
-    def _get_element_url(self, element_data):
-        return element_data['urls'][0]
+    _RESOURCE = 'episode'
+    _process_data = AdobeTVBaseIE._parse_video_data
 
     def _real_extract(self, url):
         language, show_urlname = re.match(self._VALID_URL, url).groups()
         if not language:
             language = 'en'
-        query = 'language=%s&show_urlname=%s' % (language, show_urlname)
+        query = {
+            'disclosure': 'standard',
+            'language': language,
+            'show_urlname': show_urlname,
+        }
 
-        show_data = self._download_json(self._API_BASE_URL + 'show/get/?%s' % query, show_urlname)['data'][0]
+        show_data = self._call_api(
+            'show/get', show_urlname, query)[0]
 
         return self.playlist_result(
-            self._extract_playlist_entries(self._API_BASE_URL + 'episode/?%s' % query, show_urlname),
-            compat_str(show_data['id']),
-            show_data['show_name'],
-            show_data['show_description'])
+            self._extract_playlist_entries(show_urlname, query),
+            str_or_none(show_data.get('id')),
+            show_data.get('show_name'),
+            show_data.get('show_description'))
 
 
 class AdobeTVChannelIE(AdobeTVPlaylistBaseIE):
+    IE_NAME = 'adobetv:channel'
     _VALID_URL = r'https?://tv\.adobe\.com/(?:(?P<language>fr|de|es|jp)/)?channel/(?P<id>[^/]+)(?:/(?P<category_urlname>[^/]+))?'
 
     _TEST = {
@@ -121,24 +208,30 @@ class AdobeTVChannelIE(AdobeTVPlaylistBaseIE):
         },
         'playlist_mincount': 96,
     }
+    _RESOURCE = 'show'
 
-    def _get_element_url(self, element_data):
-        return element_data['url']
+    def _process_data(self, show_data):
+        return self.url_result(
+            show_data['url'], 'AdobeTVShow', str_or_none(show_data.get('id')))
 
     def _real_extract(self, url):
         language, channel_urlname, category_urlname = re.match(self._VALID_URL, url).groups()
         if not language:
             language = 'en'
-        query = 'language=%s&channel_urlname=%s' % (language, channel_urlname)
+        query = {
+            'channel_urlname': channel_urlname,
+            'language': language,
+        }
         if category_urlname:
-            query += '&category_urlname=%s' % category_urlname
+            query['category_urlname'] = category_urlname
 
         return self.playlist_result(
-            self._extract_playlist_entries(self._API_BASE_URL + 'show/?%s' % query, channel_urlname),
+            self._extract_playlist_entries(channel_urlname, query),
             channel_urlname)
 
 
-class AdobeTVVideoIE(InfoExtractor):
+class AdobeTVVideoIE(AdobeTVBaseIE):
+    IE_NAME = 'adobetv:video'
     _VALID_URL = r'https?://video\.tv\.adobe\.com/v/(?P<id>\d+)'
 
     _TEST = {
@@ -160,38 +253,36 @@ class AdobeTVVideoIE(InfoExtractor):
 
         video_data = self._parse_json(self._search_regex(
             r'var\s+bridge\s*=\s*([^;]+);', webpage, 'bridged data'), video_id)
-
-        formats = [{
-            'format_id': '%s-%s' % (determine_ext(source['src']), source.get('height')),
-            'url': source['src'],
-            'width': int_or_none(source.get('width')),
-            'height': int_or_none(source.get('height')),
-            'tbr': int_or_none(source.get('bitrate')),
-        } for source in video_data['sources']]
+        title = video_data['title']
+
+        formats = []
+        sources = video_data.get('sources') or []
+        for source in sources:
+            source_src = source.get('src')
+            if not source_src:
+                continue
+            formats.append({
+                'filesize': int_or_none(source.get('kilobytes') or None, invscale=1000),
+                'format_id': '-'.join(filter(None, [source.get('format'), source.get('label')])),
+                'height': int_or_none(source.get('height') or None),
+                'tbr': int_or_none(source.get('bitrate') or None),
+                'width': int_or_none(source.get('width') or None),
+                'url': source_src,
+            })
         self._sort_formats(formats)
 
         # For both metadata and downloaded files the duration varies among
         # formats. I just pick the max one
         duration = max(filter(None, [
             float_or_none(source.get('duration'), scale=1000)
-            for source in video_data['sources']]))
-
-        subtitles = {}
-        for translation in video_data.get('translations', []):
-            lang_id = translation.get('language_w3c') or ISO639Utils.long2short(translation['language_medium'])
-            if lang_id not in subtitles:
-                subtitles[lang_id] = []
-            subtitles[lang_id].append({
-                'url': translation['vttPath'],
-                'ext': 'vtt',
-            })
+            for source in sources]))
 
         return {
             'id': video_id,
             'formats': formats,
-            'title': video_data['title'],
+            'title': title,
             'description': video_data.get('description'),
-            'thumbnail': video_data['video'].get('poster'),
+            'thumbnail': video_data.get('video', {}).get('poster'),
             'duration': duration,
-            'subtitles': subtitles,
+            'subtitles': self._parse_subtitles(video_data, 'vttPath'),
         }
diff --git a/youtube_dl/extractor/extractors.py b/youtube_dl/extractor/extractors.py
index 0e349b778..0f27c9678 100644
--- a/youtube_dl/extractor/extractors.py
+++ b/youtube_dl/extractor/extractors.py
@@ -21,6 +21,7 @@ from .acast import (
 from .adn import ADNIE
 from .adobeconnect import AdobeConnectIE
 from .adobetv import (
+    AdobeTVEmbedIE,
     AdobeTVIE,
     AdobeTVShowIE,
     AdobeTVChannelIE,