[canalplus] improve extraction(fixes #6301)

- extract data from json instead of xml
- fix http format urls
- extract more metadata
- update tests
- make m3u8 and f4m format extraction non fatal
- use m3u8_native implementation

1 files changed, 34 insertions, 25 deletions

diff --git a/youtube_dl/extractor/canalplus.py b/youtube_dl/extractor/canalplus.py
index 004372f8d..25b2d4efe 100644
--- a/youtube_dl/extractor/canalplus.py
+++ b/youtube_dl/extractor/canalplus.py
@@ -10,13 +10,14 @@ from ..utils import (
     unified_strdate,
     url_basename,
     qualities,
+    int_or_none,
 )
 
 
 class CanalplusIE(InfoExtractor):
     IE_DESC = 'canalplus.fr, piwiplus.fr and d8.tv'
     _VALID_URL = r'https?://(?:www\.(?P<site>canalplus\.fr|piwiplus\.fr|d8\.tv|itele\.fr)/.*?/(?P<path>.*)|player\.canalplus\.fr/#/(?P<id>[0-9]+))'
-    _VIDEO_INFO_TEMPLATE = 'http://service.canal-plus.com/video/rest/getVideosLiees/%s/%s'
+    _VIDEO_INFO_TEMPLATE = 'http://service.canal-plus.com/video/rest/getVideosLiees/%s/%s?format=json'
     _SITE_ID_MAP = {
         'canalplus.fr': 'cplus',
         'piwiplus.fr': 'teletoon',
@@ -26,10 +27,10 @@ class CanalplusIE(InfoExtractor):
 
     _TESTS = [{
         'url': 'http://www.canalplus.fr/c-emissions/pid1830-c-zapping.html?vid=1263092',
-        'md5': 'b3481d7ca972f61e37420798d0a9d934',
+        'md5': '12164a6f14ff6df8bd628e8ba9b10b78',
         'info_dict': {
             'id': '1263092',
-            'ext': 'flv',
+            'ext': 'mp4',
             'title': 'Le Zapping - 13/05/15',
             'description': 'md5:09738c0d06be4b5d06a0940edb0da73f',
             'upload_date': '20150513',
@@ -56,10 +57,10 @@ class CanalplusIE(InfoExtractor):
         'skip': 'videos get deleted after a while',
     }, {
         'url': 'http://www.itele.fr/france/video/aubervilliers-un-lycee-en-colere-111559',
-        'md5': 'f3a46edcdf28006598ffaf5b30e6a2d4',
+        'md5': '38b8f7934def74f0d6f3ba6c036a5f82',
         'info_dict': {
             'id': '1213714',
-            'ext': 'flv',
+            'ext': 'mp4',
             'title': 'Aubervilliers : un lycée en colère - Le 11/02/2015 à 06h45',
             'description': 'md5:8216206ec53426ea6321321f3b3c16db',
             'upload_date': '20150211',
@@ -82,15 +83,16 @@ class CanalplusIE(InfoExtractor):
                 webpage, 'video id', group='id')
 
         info_url = self._VIDEO_INFO_TEMPLATE % (site_id, video_id)
-        doc = self._download_xml(info_url, video_id, 'Downloading video XML')
+        video_data = self._download_json(info_url, video_id, 'Downloading video JSON')
 
-        video_info = [video for video in doc if video.find('ID').text == video_id][0]
-        media = video_info.find('MEDIA')
-        infos = video_info.find('INFOS')
+        if isinstance(video_data, list):
+            video_data = [video for video in video_data if video.get('ID') == video_id][0]
+        media = video_data['MEDIA']
+        infos = video_data['INFOS']
 
-        preference = qualities(['MOBILE', 'BAS_DEBIT', 'HAUT_DEBIT', 'HD', 'HLS', 'HDS'])
+        preference = qualities(['MOBILE', 'BAS_DEBIT', 'HAUT_DEBIT', 'HD'])
 
-        fmt_url = next(iter(media.find('VIDEOS'))).text
+        fmt_url = next(iter(media.get('VIDEOS')))
         if '/geo' in fmt_url.lower():
             response = self._request_webpage(
                 HEADRequest(fmt_url), video_id,
@@ -101,35 +103,42 @@ class CanalplusIE(InfoExtractor):
                     expected=True)
 
         formats = []
-        for fmt in media.find('VIDEOS'):
-            format_url = fmt.text
+        for format_id, format_url in media['VIDEOS'].items():
             if not format_url:
                 continue
-            format_id = fmt.tag
             if format_id == 'HLS':
                 formats.extend(self._extract_m3u8_formats(
-                    format_url, video_id, 'mp4', preference=preference(format_id)))
+                    format_url, video_id, 'mp4', 'm3u8_native', m3u8_id=format_id, fatal=False))
             elif format_id == 'HDS':
                 formats.extend(self._extract_f4m_formats(
-                    format_url + '?hdcore=2.11.3', video_id, preference=preference(format_id)))
+                    format_url + '?hdcore=2.11.3', video_id, f4m_id=format_id, fatal=False))
             else:
                 formats.append({
-                    'url': format_url,
+                    # the secret extracted ya function in http://player.canalplus.fr/common/js/canalPlayer.js
+                    'url': format_url + '?secret=pqzerjlsmdkjfoiuerhsdlfknaes',
                     'format_id': format_id,
                     'preference': preference(format_id),
                 })
         self._sort_formats(formats)
 
+        thumbnails = [{
+            'id': image_id,
+            'url': image_url,
+        } for image_id, image_url in media.get('images', {}).items()]
+
+        titrage = infos['TITRAGE']
+
         return {
             'id': video_id,
             'display_id': display_id,
-            'title': '%s - %s' % (infos.find('TITRAGE/TITRE').text,
-                                  infos.find('TITRAGE/SOUS_TITRE').text),
-            'upload_date': unified_strdate(infos.find('PUBLICATION/DATE').text),
-            'thumbnail': media.find('IMAGES/GRAND').text,
-            'description': infos.find('DESCRIPTION').text,
-            'view_count': int(infos.find('NB_VUES').text),
-            'like_count': int(infos.find('NB_LIKES').text),
-            'comment_count': int(infos.find('NB_COMMENTS').text),
+            'title': '%s - %s' % (titrage['TITRE'],
+                                  titrage['SOUS_TITRE']),
+            'upload_date': unified_strdate(infos.get('PUBLICATION', {}).get('DATE')),
+            'thumbnails': thumbnails,
+            'description': infos.get('DESCRIPTION'),
+            'duration': int_or_none(infos.get('DURATION')),
+            'view_count': int_or_none(infos.get('NB_VUES')),
+            'like_count': int_or_none(infos.get('NB_LIKES')),
+            'comment_count': int_or_none(infos.get('NB_COMMENTS')),
             'formats': formats,
         }