Merge remote-tracking branch 'jaimeMF/f4m'

Conflicts:
	youtube_dl/extractor/__init__.py

youtube_dl/downloader/__init__.py

@@ -5,6 +5,7 @@ from .hls import HlsFD
 from .http import HttpFD
 from .mplayer import MplayerFD
 from .rtmp import RtmpFD
+from .f4m import F4mFD
 
 from ..utils import (
     determine_ext,
@@ -22,5 +23,7 @@ def get_suitable_downloader(info_dict):
         return HlsFD
     if url.startswith('mms') or url.startswith('rtsp'):
         return MplayerFD
+    if determine_ext(url) == 'f4m':
+        return F4mFD
     else:
         return HttpFD

youtube_dl/downloader/f4m.py

@@ -0,0 +1,305 @@
+from __future__ import unicode_literals
+
+import base64
+import io
+import itertools
+import os
+from struct import unpack, pack
+import time
+import xml.etree.ElementTree as etree
+
+from .common import FileDownloader
+from .http import HttpFD
+from ..utils import (
+    compat_urllib_request,
+    compat_urlparse,
+    format_bytes,
+    encodeFilename,
+    sanitize_open,
+)
+
+
+class FlvReader(io.BytesIO):
+    """
+    Reader for Flv files
+    The file format is documented in https://www.adobe.com/devnet/f4v.html
+    """
+
+    # Utility functions for reading numbers and strings
+    def read_unsigned_long_long(self):
+        return unpack('!Q', self.read(8))[0]
+
+    def read_unsigned_int(self):
+        return unpack('!I', self.read(4))[0]
+
+    def read_unsigned_char(self):
+        return unpack('!B', self.read(1))[0]
+
+    def read_string(self):
+        res = b''
+        while True:
+            char = self.read(1)
+            if char == b'\x00':
+                break
+            res += char
+        return res
+
+    def read_box_info(self):
+        """
+        Read a box and return the info as a tuple: (box_size, box_type, box_data)
+        """
+        real_size = size = self.read_unsigned_int()
+        box_type = self.read(4)
+        header_end = 8
+        if size == 1:
+            real_size = self.read_unsigned_long_long()
+            header_end = 16
+        return real_size, box_type, self.read(real_size-header_end)
+
+    def read_asrt(self):
+        # version
+        self.read_unsigned_char()
+        # flags
+        self.read(3)
+        quality_entry_count = self.read_unsigned_char()
+        # QualityEntryCount
+        for i in range(quality_entry_count):
+            self.read_string()
+
+        segment_run_count = self.read_unsigned_int()
+        segments = []
+        for i in range(segment_run_count):
+            first_segment = self.read_unsigned_int()
+            fragments_per_segment = self.read_unsigned_int()
+            segments.append((first_segment, fragments_per_segment))
+
+        return {
+            'segment_run': segments,
+        }
+
+    def read_afrt(self):
+        # version
+        self.read_unsigned_char()
+        # flags
+        self.read(3)
+        # time scale
+        self.read_unsigned_int()
+
+        quality_entry_count = self.read_unsigned_char()
+        # QualitySegmentUrlModifiers
+        for i in range(quality_entry_count):
+            self.read_string()
+
+        fragments_count = self.read_unsigned_int()
+        fragments = []
+        for i in range(fragments_count):
+            first = self.read_unsigned_int()
+            first_ts = self.read_unsigned_long_long()
+            duration = self.read_unsigned_int()
+            if duration == 0:
+                discontinuity_indicator = self.read_unsigned_char()
+            else:
+                discontinuity_indicator = None
+            fragments.append({
+                'first': first,
+                'ts': first_ts,
+                'duration': duration,
+                'discontinuity_indicator': discontinuity_indicator,
+            })
+
+        return {
+            'fragments': fragments,
+        }
+
+    def read_abst(self):
+        # version
+        self.read_unsigned_char()
+        # flags
+        self.read(3)
+        # BootstrapinfoVersion
+        bootstrap_info_version = self.read_unsigned_int()
+        # Profile,Live,Update,Reserved
+        self.read(1)
+        # time scale
+        self.read_unsigned_int()
+        # CurrentMediaTime
+        self.read_unsigned_long_long()
+        # SmpteTimeCodeOffset
+        self.read_unsigned_long_long()
+        # MovieIdentifier
+        movie_identifier = self.read_string()
+        server_count = self.read_unsigned_char()
+        # ServerEntryTable
+        for i in range(server_count):
+            self.read_string()
+        quality_count = self.read_unsigned_char()
+        # QualityEntryTable
+        for i in range(server_count):
+            self.read_string()
+        # DrmData
+        self.read_string()
+        # MetaData
+        self.read_string()
+
+        segments_count = self.read_unsigned_char()
+        segments = []
+        for i in range(segments_count):
+            box_size, box_type, box_data = self.read_box_info()
+            assert box_type == b'asrt'
+            segment = FlvReader(box_data).read_asrt()
+            segments.append(segment)
+        fragments_run_count = self.read_unsigned_char()
+        fragments = []
+        for i in range(fragments_run_count):
+            box_size, box_type, box_data = self.read_box_info()
+            assert box_type == b'afrt'
+            fragments.append(FlvReader(box_data).read_afrt())
+
+        return {
+            'segments': segments,
+            'fragments': fragments,
+        }
+
+    def read_bootstrap_info(self):
+        total_size, box_type, box_data = self.read_box_info()
+        assert box_type == b'abst'
+        return FlvReader(box_data).read_abst()
+
+
+def read_bootstrap_info(bootstrap_bytes):
+    return FlvReader(bootstrap_bytes).read_bootstrap_info()
+
+
+def build_fragments_list(boot_info):
+    """ Return a list of (segment, fragment) for each fragment in the video """
+    res = []
+    segment_run_table = boot_info['segments'][0]
+    # I've only found videos with one segment
+    segment_run_entry = segment_run_table['segment_run'][0]
+    n_frags = segment_run_entry[1]
+    fragment_run_entry_table = boot_info['fragments'][0]['fragments']
+    first_frag_number = fragment_run_entry_table[0]['first']
+    for (i, frag_number) in zip(range(1, n_frags+1), itertools.count(first_frag_number)):
+        res.append((1, frag_number))
+    return res
+
+
+def write_flv_header(stream, metadata):
+    """Writes the FLV header and the metadata to stream"""
+    # FLV header
+    stream.write(b'FLV\x01')
+    stream.write(b'\x05')
+    stream.write(b'\x00\x00\x00\x09')
+    # FLV File body
+    stream.write(b'\x00\x00\x00\x00')
+    # FLVTAG
+    # Script data
+    stream.write(b'\x12')
+    # Size of the metadata with 3 bytes
+    stream.write(pack('!L', len(metadata))[1:])
+    stream.write(b'\x00\x00\x00\x00\x00\x00\x00')
+    stream.write(metadata)
+    # Magic numbers extracted from the output files produced by AdobeHDS.php
+    #(https://github.com/K-S-V/Scripts)
+    stream.write(b'\x00\x00\x01\x73')
+
+
+def _add_ns(prop):
+    return '{http://ns.adobe.com/f4m/1.0}%s' % prop
+
+
+class HttpQuietDownloader(HttpFD):
+    def to_screen(self, *args, **kargs):
+        pass
+
+
+class F4mFD(FileDownloader):
+    """
+    A downloader for f4m manifests or AdobeHDS.
+    """
+
+    def real_download(self, filename, info_dict):
+        man_url = info_dict['url']
+        self.to_screen('[download] Downloading f4m manifest')
+        manifest = self.ydl.urlopen(man_url).read()
+        self.report_destination(filename)
+        http_dl = HttpQuietDownloader(self.ydl, {'continuedl': True, 'quiet': True, 'noprogress': True})
+
+        doc = etree.fromstring(manifest)
+        formats = [(int(f.attrib.get('bitrate', -1)), f) for f in doc.findall(_add_ns('media'))]
+        formats = sorted(formats, key=lambda f: f[0])
+        rate, media = formats[-1]
+        base_url = compat_urlparse.urljoin(man_url, media.attrib['url'])
+        bootstrap = base64.b64decode(doc.find(_add_ns('bootstrapInfo')).text)
+        metadata = base64.b64decode(media.find(_add_ns('metadata')).text)
+        boot_info = read_bootstrap_info(bootstrap)
+        fragments_list = build_fragments_list(boot_info)
+        total_frags = len(fragments_list)
+
+        tmpfilename = self.temp_name(filename)
+        (dest_stream, tmpfilename) = sanitize_open(tmpfilename, 'wb')
+        write_flv_header(dest_stream, metadata)
+
+        # This dict stores the download progress, it's updated by the progress
+        # hook
+        state = {
+            'downloaded_bytes': 0,
+            'frag_counter': 0,
+        }
+        start = time.time()
+
+        def frag_progress_hook(status):
+            frag_total_bytes = status.get('total_bytes', 0)
+            estimated_size = (state['downloaded_bytes'] +
+                (total_frags - state['frag_counter']) * frag_total_bytes)
+            if status['status'] == 'finished':
+                state['downloaded_bytes'] += frag_total_bytes
+                state['frag_counter'] += 1
+                progress = self.calc_percent(state['frag_counter'], total_frags)
+                byte_counter = state['downloaded_bytes']
+            else:
+                frag_downloaded_bytes = status['downloaded_bytes']
+                byte_counter = state['downloaded_bytes'] + frag_downloaded_bytes
+                frag_progress = self.calc_percent(frag_downloaded_bytes,
+                    frag_total_bytes)
+                progress = self.calc_percent(state['frag_counter'], total_frags)
+                progress += frag_progress / float(total_frags)
+
+            eta = self.calc_eta(start, time.time(), estimated_size, byte_counter)
+            self.report_progress(progress, format_bytes(estimated_size),
+                status.get('speed'), eta)
+        http_dl.add_progress_hook(frag_progress_hook)
+
+        frags_filenames = []
+        for (seg_i, frag_i) in fragments_list:
+            name = 'Seg%d-Frag%d' % (seg_i, frag_i)
+            url = base_url + name
+            frag_filename = '%s-%s' % (tmpfilename, name)
+            success = http_dl.download(frag_filename, {'url': url})
+            if not success:
+                return False
+            with open(frag_filename, 'rb') as down:
+                down_data = down.read()
+                reader = FlvReader(down_data)
+                while True:
+                    _, box_type, box_data = reader.read_box_info()
+                    if box_type == b'mdat':
+                        dest_stream.write(box_data)
+                        break
+            frags_filenames.append(frag_filename)
+
+        self.report_finish(format_bytes(state['downloaded_bytes']), time.time() - start)
+
+        self.try_rename(tmpfilename, filename)
+        for frag_file in frags_filenames:
+            os.remove(frag_file)
+
+        fsize = os.path.getsize(encodeFilename(filename))
+        self._hook_progress({
+            'downloaded_bytes': fsize,
+            'total_bytes': fsize,
+            'filename': filename,
+            'status': 'finished',
+        })
+
+        return True

youtube_dl/extractor/__init__.py

@@ -209,6 +209,7 @@ from .statigram import StatigramIE
 from .steam import SteamIE
 from .streamcloud import StreamcloudIE
 from .streamcz import StreamCZIE
+from .syfy import SyfyIE
 from .sztvhu import SztvHuIE
 from .teamcoco import TeamcocoIE
 from .techtalks import TechTalksIE

youtube_dl/extractor/syfy.py

@@ -0,0 +1,30 @@
+from __future__ import unicode_literals
+
+import re
+
+from .common import InfoExtractor
+
+
+class SyfyIE(InfoExtractor):
+    _VALID_URL = r'https?://www\.syfy\.com/videos/.+?vid:(?P<id>\d+)'
+
+    _TEST = {
+        'url': 'http://www.syfy.com/videos/Robot%20Combat%20League/Behind%20the%20Scenes/vid:2631458',
+        'info_dict': {
+            'id': 'NmqMrGnXvmO1',
+            'ext': 'flv',
+            'title': 'George Lucas has Advice for his Daughter',
+            'description': 'Listen to what insights George Lucas give his daughter Amanda.',
+        },
+        'params': {
+            # f4m download
+            'skip_download': True,
+        },
+        'add_ie': ['ThePlatform'],
+    }
+
+    def _real_extract(self, url):
+        mobj = re.match(self._VALID_URL, url)
+        video_id = mobj.group('id')
+        webpage = self._download_webpage(url, video_id)
+        return self.url_result(self._og_search_video_url(webpage))

youtube_dl/extractor/theplatform.py

@@ -11,7 +11,10 @@ _x = lambda p: xpath_with_ns(p, {'smil': 'http://www.w3.org/2005/SMIL21/Language
 
 
 class ThePlatformIE(InfoExtractor):
-    _VALID_URL = r'(?:https?://link\.theplatform\.com/s/[^/]+/|theplatform:)(?P<id>[^/\?]+)'
+    _VALID_URL = r'''(?x)
+        (?:https?://(?:link|player)\.theplatform\.com/[sp]/[^/]+/
+           (?P<config>[^/\?]+/(?:swf|config)/select/)?
+         |theplatform:)(?P<id>[^/\?&]+)'''
 
     _TEST = {
         # from http://www.metacafe.com/watch/cb-e9I_cZgTgIPd/blackberrys_big_bold_z30/
@@ -29,9 +32,7 @@ class ThePlatformIE(InfoExtractor):
         },
     }
 
-    def _get_info(self, video_id):
-        smil_url = ('http://link.theplatform.com/s/dJ5BDC/{0}/meta.smil?'
-            'format=smil&mbr=true'.format(video_id))
+    def _get_info(self, video_id, smil_url):
         meta = self._download_xml(smil_url, video_id)
 
         try:
@@ -50,26 +51,34 @@ class ThePlatformIE(InfoExtractor):
 
         head = meta.find(_x('smil:head'))
         body = meta.find(_x('smil:body'))
-        base_url = head.find(_x('smil:meta')).attrib['base']
-        switch = body.find(_x('smil:switch'))
-        formats = []
-        for f in switch.findall(_x('smil:video')):
-            attr = f.attrib
-            width = int(attr['width'])
-            height = int(attr['height'])
-            vbr = int(attr['system-bitrate']) // 1000
-            format_id = '%dx%d_%dk' % (width, height, vbr)
-            formats.append({
-                'format_id': format_id,
-                'url': base_url,
-                'play_path': 'mp4:' + attr['src'],
-                'ext': 'flv',
-                'width': width,
-                'height': height,
-                'vbr': vbr,
-            })
 
-        self._sort_formats(formats)
+        f4m_node = body.find(_x('smil:seq/smil:video'))
+        if f4m_node is not None:
+            formats = [{
+                'ext': 'flv',
+                # the parameters are from syfy.com, other sites may use others
+                'url': f4m_node.attrib['src'] + '?g=UXWGVKRWHFSP&hdcore=3.0.3',
+            }]
+        else:
+            base_url = head.find(_x('smil:meta')).attrib['base']
+            switch = body.find(_x('smil:switch'))
+            formats = []
+            for f in switch.findall(_x('smil:video')):
+                attr = f.attrib
+                width = int(attr['width'])
+                height = int(attr['height'])
+                vbr = int(attr['system-bitrate']) // 1000
+                format_id = '%dx%d_%dk' % (width, height, vbr)
+                formats.append({
+                    'format_id': format_id,
+                    'url': base_url,
+                    'play_path': 'mp4:' + attr['src'],
+                    'ext': 'flv',
+                    'width': width,
+                    'height': height,
+                    'vbr': vbr,
+                })
+            self._sort_formats(formats)
 
         return {
             'id': video_id,
@@ -83,4 +92,13 @@ class ThePlatformIE(InfoExtractor):
     def _real_extract(self, url):
         mobj = re.match(self._VALID_URL, url)
         video_id = mobj.group('id')
-        return self._get_info(video_id)
+        if mobj.group('config'):
+            config_url = url+ '&form=json'
+            config_url = config_url.replace('swf/', 'config/')
+            config_json = self._download_webpage(config_url, video_id, u'Downloading config')
+            config = json.loads(config_json)
+            smil_url = config['releaseUrl'] + '&format=SMIL&formats=MPEG4'
+        else:
+            smil_url = ('http://link.theplatform.com/s/dJ5BDC/{0}/meta.smil?'
+                'format=smil&mbr=true'.format(video_id))
+        return self._get_info(video_id, smil_url)