[zdf/common] Use API in ZDF extractor.

This also comes with a lot of extra format fields
Fixes #1518

1 files changed, 70 insertions, 45 deletions

diff --git a/youtube_dl/extractor/zdf.py b/youtube_dl/extractor/zdf.py
index c6a9d06f2..a8d899883 100644
--- a/youtube_dl/extractor/zdf.py
+++ b/youtube_dl/extractor/zdf.py
@@ -1,75 +1,100 @@
+import operator
 import re
 
 from .common import InfoExtractor
 from ..utils import (
-    determine_ext,
-    ExtractorError,
+    parse_xml_doc,
+    unified_strdate,
 )
 
 
 class ZDFIE(InfoExtractor):
     _VALID_URL = r'^http://www\.zdf\.de\/ZDFmediathek(?P<hash>#)?\/(.*beitrag\/video\/)(?P<video_id>[^/\?]+)(?:\?.*)?'
-    _MEDIA_STREAM = r'<a href="(?P<video_url>.+(?P<media_type>.streaming).+/zdf/(?P<quality>[^\/]+)/[^"]*)".+class="play".+>'
 
     def _real_extract(self, url):
         mobj = re.match(self._VALID_URL, url)
-        if mobj is None:
-            raise ExtractorError(u'Invalid URL: %s' % url)
         video_id = mobj.group('video_id')
 
-        if mobj.group('hash'):
-            url = url.replace(u'#', u'', 1)
+        xml_url = u'http://www.zdf.de/ZDFmediathek/xmlservice/web/beitragsDetails?ak=web&id=%s' % video_id
+        info_xml = self._download_webpage(
+            xml_url, video_id, note=u'Downloading video info')
+        doc = parse_xml_doc(info_xml)
 
-        html = self._download_webpage(url, video_id)
-        streams = [m.groupdict() for m in re.finditer(self._MEDIA_STREAM, html)]
-        if streams is None:
-            raise ExtractorError(u'No media url found.')
+        title = doc.find('.//information/title').text
+        description = doc.find('.//information/detail').text
+        uploader_node = doc.find('.//details/originChannelTitle')
+        uploader = None if uploader_node is None else uploader_node.text
+        duration_str = doc.find('.//details/length').text
+        duration_m = re.match(r'''(?x)^
+            (?P<hours>[0-9]{2})
+            :(?P<minutes>[0-9]{2})
+            :(?P<seconds>[0-9]{2})
+            (?:\.(?P<ms>[0-9]+)?)
+            ''', duration_str)
+        duration = (
+            (
+                (int(duration_m.group('hours')) * 60 * 60) +
+                (int(duration_m.group('minutes')) * 60) +
+                int(duration_m.group('seconds'))
+            )
+            if duration_m
+            else None
+        )
+        upload_date = unified_strdate(doc.find('.//details/airtime').text)
 
-        # s['media_type'] == 'wstreaming' -> use 'Windows Media Player' and mms url
-        # s['media_type'] == 'hstreaming' -> use 'Quicktime' and rtsp url
-        # choose first/default media type and highest quality for now
-        def stream_pref(s):
-            TYPE_ORDER = ['ostreaming', 'hstreaming', 'wstreaming']
+        def xml_to_format(fnode):
+            video_url = fnode.find('url').text
+            is_available = u'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)
+
+            PROTO_ORDER = ['http', 'rtmp', 'rtsp']
             try:
-                type_pref = TYPE_ORDER.index(s['media_type'])
+                proto_pref = -PROTO_ORDER.index(format_m.group('proto'))
             except ValueError:
-                type_pref = 999
+                proto_pref = 999
 
-            QUALITY_ORDER = ['veryhigh', '300']
+            quality = fnode.find('./quality').text
+            QUALITY_ORDER = ['veryhigh', '300', 'high', 'med', 'low']
             try:
-                quality_pref = QUALITY_ORDER.index(s['quality'])
+                quality_pref = -QUALITY_ORDER.index(quality)
             except ValueError:
                 quality_pref = 999
 
-            return (type_pref, quality_pref)
-
-        sorted_streams = sorted(streams, key=stream_pref)
-        if not sorted_streams:
-            raise ExtractorError(u'No stream found.')
-        stream = sorted_streams[0]
-
-        media_link = self._download_webpage(
-            stream['video_url'],
-            video_id,
-            u'Get stream URL')
-
-        #MMS_STREAM = r'href="(?P<video_url>mms://[^"]*)"'
-        RTSP_STREAM = r'(?P<video_url>rtsp://[^"]*.mp4)'
+            abr = int(fnode.find('./audioBitrate').text) // 1000
+            vbr = int(fnode.find('./videoBitrate').text) // 1000
+            pref = (is_available, proto_pref, quality_pref, vbr, abr)
 
-        mobj = re.search(self._MEDIA_STREAM, media_link)
-        if mobj is None:
-            mobj = re.search(RTSP_STREAM, media_link)
-            if mobj is None:
-                raise ExtractorError(u'Cannot extract mms:// or rtsp:// URL')
-        video_url = mobj.group('video_url')
+            return {
+                'format_id': format_id,
+                'url': video_url,
+                'ext': format_m.group('container'),
+                'acodec': format_m.group('acodec'),
+                'vcodec': format_m.group('vcodec'),
+                'abr': abr,
+                'vbr': vbr,
+                'width': int(fnode.find('./width').text),
+                'height': int(fnode.find('./height').text),
+                'quality_name': quality,
+                'filesize': int(fnode.find('./filesize').text),
+                'format_note': None if is_available else u'(unavailable)',
+                '_pref': pref,
+            }
 
-        title = self._html_search_regex(
-            r'<h1(?: class="beitragHeadline")?>(.*?)</h1>',
-            html, u'title')
+        format_nodes = doc.findall('.//formitaeten/formitaet')
+        formats = sorted(map(xml_to_format, format_nodes),
+                         key=operator.itemgetter('_pref'))
 
         return {
             'id': video_id,
-            'url': video_url,
             'title': title,
-            'ext': determine_ext(video_url)
+            'formats': formats,
+            'description': description,
+            'uploader': uploader,
+            'duration': duration,
+            'upload_date': upload_date,
         }