[zdf] fix formats extraction

2 files changed, 161 insertions, 101 deletions

diff --git a/youtube_dl/extractor/phoenix.py b/youtube_dl/extractor/phoenix.py
index 46cebc0d7..6ce2ec19d 100644
--- a/youtube_dl/extractor/phoenix.py
+++ b/youtube_dl/extractor/phoenix.py
@@ -1,10 +1,9 @@
 from __future__ import unicode_literals
 
-from .common import InfoExtractor
-from .zdf import extract_from_xml_url
+from .zdf import ZDFIE
 
 
-class PhoenixIE(InfoExtractor):
+class PhoenixIE(ZDFIE):
     _VALID_URL = r'''(?x)https?://(?:www\.)?phoenix\.de/content/
         (?:
             phoenix/die_sendungen/(?:[^/]+/)?
@@ -41,5 +40,5 @@ class PhoenixIE(InfoExtractor):
             r'<div class="phx_vod" id="phx_vod_([0-9]+)"',
             webpage, 'internal video ID')
 
-        api_url = 'http://www.phoenix.de/php/zdfplayer-v1.3/data/beitragsDetails.php?ak=web&id=%s' % internal_id
-        return extract_from_xml_url(self, video_id, api_url)
+        api_url = 'http://www.phoenix.de/php/mediaplayer/data/beitrags_details.php?ak=web&id=%s' % internal_id
+        return self.extract_from_xml_url(video_id, api_url)
diff --git a/youtube_dl/extractor/zdf.py b/youtube_dl/extractor/zdf.py
index 9a3331a69..92c12bac6 100644
--- a/youtube_dl/extractor/zdf.py
+++ b/youtube_dl/extractor/zdf.py
@@ -10,106 +10,16 @@ from ..utils import (
     unified_strdate,
     OnDemandPagedList,
     xpath_text,
+    determine_ext,
+    qualities,
+    float_or_none,
 )
 
 
-def extract_from_xml_url(ie, video_id, xml_url):
-    doc = ie._download_xml(
-        xml_url, video_id,
-        note='Downloading video info',
-        errnote='Failed to download video info')
-
-    title = doc.find('.//information/title').text
-    description = xpath_text(doc, './/information/detail', 'description')
-    duration = int_or_none(xpath_text(doc, './/details/lengthSec', 'duration'))
-    uploader = xpath_text(doc, './/details/originChannelTitle', 'uploader')
-    uploader_id = xpath_text(doc, './/details/originChannelId', 'uploader id')
-    upload_date = unified_strdate(xpath_text(doc, './/details/airtime', 'upload date'))
-
-    def xml_to_format(fnode):
-        video_url = fnode.find('url').text
-        is_available = 'http://www.metafilegenerator' not in video_url
-
-        format_id = fnode.attrib['basetype']
-        format_m = re.match(r'''(?x)
-            (?P<vcodec>[^_]+)_(?P<acodec>[^_]+)_(?P<container>[^_]+)_
-            (?P<proto>[^_]+)_(?P<index>[^_]+)_(?P<indexproto>[^_]+)
-        ''', format_id)
-
-        ext = format_m.group('container')
-        proto = format_m.group('proto').lower()
-
-        quality = xpath_text(fnode, './quality', 'quality')
-        abr = int_or_none(xpath_text(fnode, './audioBitrate', 'abr'), 1000)
-        vbr = int_or_none(xpath_text(fnode, './videoBitrate', 'vbr'), 1000)
-
-        width = int_or_none(xpath_text(fnode, './width', 'width'))
-        height = int_or_none(xpath_text(fnode, './height', 'height'))
-
-        filesize = int_or_none(xpath_text(fnode, './filesize', 'filesize'))
-
-        format_note = ''
-        if not format_note:
-            format_note = None
-
-        return {
-            'format_id': format_id + '-' + quality,
-            'url': video_url,
-            'ext': ext,
-            'acodec': format_m.group('acodec'),
-            'vcodec': format_m.group('vcodec'),
-            'abr': abr,
-            'vbr': vbr,
-            'width': width,
-            'height': height,
-            'filesize': filesize,
-            'format_note': format_note,
-            'protocol': proto,
-            '_available': is_available,
-        }
-
-    def xml_to_thumbnails(fnode):
-        thumbnails = []
-        for node in fnode:
-            thumbnail_url = node.text
-            if not thumbnail_url:
-                continue
-            thumbnail = {
-                'url': thumbnail_url,
-            }
-            if 'key' in node.attrib:
-                m = re.match('^([0-9]+)x([0-9]+)$', node.attrib['key'])
-                if m:
-                    thumbnail['width'] = int(m.group(1))
-                    thumbnail['height'] = int(m.group(2))
-            thumbnails.append(thumbnail)
-        return thumbnails
-
-    thumbnails = xml_to_thumbnails(doc.findall('.//teaserimages/teaserimage'))
-
-    format_nodes = doc.findall('.//formitaeten/formitaet')
-    formats = list(filter(
-        lambda f: f['_available'],
-        map(xml_to_format, format_nodes)))
-    ie._sort_formats(formats)
-
-    return {
-        'id': video_id,
-        'title': title,
-        'description': description,
-        'duration': duration,
-        'thumbnails': thumbnails,
-        'uploader': uploader,
-        'uploader_id': uploader_id,
-        'upload_date': upload_date,
-        'formats': formats,
-    }
-
-
 class ZDFIE(InfoExtractor):
     _VALID_URL = r'(?:zdf:|zdf:video:|https?://www\.zdf\.de/ZDFmediathek(?:#)?/(.*beitrag/(?:video/)?))(?P<id>[0-9]+)(?:/[^/?]+)?(?:\?.*)?'
 
-    _TEST = {
+    _TESTS = [{
         'url': 'http://www.zdf.de/ZDFmediathek/beitrag/video/2037704/ZDFspezial---Ende-des-Machtpokers--?bc=sts;stt',
         'info_dict': {
             'id': '2037704',
@@ -122,12 +32,163 @@ class ZDFIE(InfoExtractor):
             'upload_date': '20131127',
         },
         'skip': 'Videos on ZDF.de are depublicised in short order',
-    }
+    }]
+
+    def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
+        param_groups = {}
+        for param_group in smil.findall(self._xpath_ns('./head/paramGroup', namespace)):
+            group_id = param_group.attrib.get(self._xpath_ns('id', 'http://www.w3.org/XML/1998/namespace'))
+            params = {}
+            for param in param_group:
+                params[param.get('name')] = param.get('value')
+            param_groups[group_id] = params
+
+        formats = []
+        for video in smil.findall(self._xpath_ns('.//video', namespace)):
+            src = video.get('src')
+            if not src:
+                continue
+            bitrate = float_or_none(video.get('system-bitrate') or video.get('systemBitrate'), 1000)
+            group_id = video.get('paramGroup')
+            param_group = param_groups[group_id]
+            for proto in param_group['protocols'].split(','):
+                formats.append({
+                    'url': '%s://%s' % (proto, param_group['host']),
+                    'app': param_group['app'],
+                    'play_path': src,
+                    'ext': 'flv',
+                    'format_id': '%s-%d' % (proto, bitrate),
+                    'tbr': bitrate,
+                    'protocol': proto,
+                })
+        self._sort_formats(formats)
+        return formats
+
+    def extract_from_xml_url(self, video_id, xml_url):
+        doc = self._download_xml(
+            xml_url, video_id,
+            note='Downloading video info',
+            errnote='Failed to download video info')
+
+        title = doc.find('.//information/title').text
+        description = xpath_text(doc, './/information/detail', 'description')
+        duration = int_or_none(xpath_text(doc, './/details/lengthSec', 'duration'))
+        uploader = xpath_text(doc, './/details/originChannelTitle', 'uploader')
+        uploader_id = xpath_text(doc, './/details/originChannelId', 'uploader id')
+        upload_date = unified_strdate(xpath_text(doc, './/details/airtime', 'upload date'))
+
+        def xml_to_thumbnails(fnode):
+            thumbnails = []
+            for node in fnode:
+                thumbnail_url = node.text
+                if not thumbnail_url:
+                    continue
+                thumbnail = {
+                    'url': thumbnail_url,
+                }
+                if 'key' in node.attrib:
+                    m = re.match('^([0-9]+)x([0-9]+)$', node.attrib['key'])
+                    if m:
+                        thumbnail['width'] = int(m.group(1))
+                        thumbnail['height'] = int(m.group(2))
+                thumbnails.append(thumbnail)
+            return thumbnails
+
+        thumbnails = xml_to_thumbnails(doc.findall('.//teaserimages/teaserimage'))
+
+        format_nodes = doc.findall('.//formitaeten/formitaet')
+        quality = qualities(['veryhigh', 'high', 'med', 'low'])
+
+        def get_quality(elem):
+            return quality(xpath_text(elem, 'quality'))
+        format_nodes.sort(key=get_quality)
+        format_ids = []
+        formats = []
+        for fnode in format_nodes:
+            video_url = fnode.find('url').text
+            is_available = 'http://www.metafilegenerator' not in video_url
+            if not is_available:
+                continue
+            format_id = fnode.attrib['basetype']
+            quality = xpath_text(fnode, './quality', 'quality')
+            format_m = re.match(r'''(?x)
+                (?P<vcodec>[^_]+)_(?P<acodec>[^_]+)_(?P<container>[^_]+)_
+                (?P<proto>[^_]+)_(?P<index>[^_]+)_(?P<indexproto>[^_]+)
+            ''', format_id)
+
+            ext = determine_ext(video_url, None) or format_m.group('container')
+            if ext not in ('smil', 'f4m', 'm3u8'):
+                format_id = format_id + '-' + quality
+            if format_id in format_ids:
+                continue
+
+            if ext == 'meta':
+                continue
+            elif ext == 'smil':
+                smil_formats = self._extract_smil_formats(
+                    video_url, video_id, fatal=False)
+                if smil_formats:
+                    formats.extend(smil_formats)
+            elif ext == 'm3u8':
+                m3u8_formats = self._extract_m3u8_formats(
+                    video_url, video_id, 'mp4', m3u8_id='hls', fatal=False)
+                if m3u8_formats:
+                    formats.extend(m3u8_formats)
+            elif ext == 'f4m':
+                f4m_formats = self._extract_f4m_formats(
+                    video_url, video_id, f4m_id='hds', fatal=False)
+                if f4m_formats:
+                    formats.extend(f4m_formats)
+            else:
+                proto = format_m.group('proto').lower()
+
+                abr = int_or_none(xpath_text(fnode, './audioBitrate', 'abr'), 1000)
+                vbr = int_or_none(xpath_text(fnode, './videoBitrate', 'vbr'), 1000)
+
+                width = int_or_none(xpath_text(fnode, './width', 'width'))
+                height = int_or_none(xpath_text(fnode, './height', 'height'))
+
+                filesize = int_or_none(xpath_text(fnode, './filesize', 'filesize'))
+
+                format_note = ''
+                if not format_note:
+                    format_note = None
+
+                formats.append({
+                    'format_id': format_id,
+                    'url': video_url,
+                    'ext': ext,
+                    'acodec': format_m.group('acodec'),
+                    'vcodec': format_m.group('vcodec'),
+                    'abr': abr,
+                    'vbr': vbr,
+                    'width': width,
+                    'height': height,
+                    'filesize': filesize,
+                    'format_note': format_note,
+                    'protocol': proto,
+                    '_available': is_available,
+                })
+            format_ids.append(format_id)
+
+        self._sort_formats(formats)
+
+        return {
+            'id': video_id,
+            'title': title,
+            'description': description,
+            'duration': duration,
+            'thumbnails': thumbnails,
+            'uploader': uploader,
+            'uploader_id': uploader_id,
+            'upload_date': upload_date,
+            'formats': formats,
+        }
 
     def _real_extract(self, url):
         video_id = self._match_id(url)
         xml_url = 'http://www.zdf.de/ZDFmediathek/xmlservice/web/beitragsDetails?ak=web&id=%s' % video_id
-        return extract_from_xml_url(self, video_id, xml_url)
+        return self.extract_from_xml_url(video_id, xml_url)
 
 
 class ZDFChannelIE(InfoExtractor):