]> Raphaƫl G. Git Repositories - youtubedl/blobdiff - youtube_dl/extractor/common.py
Set upstream metadata fields: Bug-Database, Bug-Submit, Repository, Repository-Browse.
[youtubedl] / youtube_dl / extractor / common.py
index e4e4feef9ea18787d196b2ca7d3414191409ba97..eaae5e484f99311ccf018301f773c87f9c8cf544 100644 (file)
@@ -1,27 +1,85 @@
+# coding: utf-8
+from __future__ import unicode_literals
+
 import base64
+import datetime
 import hashlib
 import json
+import netrc
 import os
+import random
 import re
 import socket
 import sys
-import netrc
-import xml.etree.ElementTree
-
-from ..utils import (
+import time
+import math
+
+from ..compat import (
+    compat_cookiejar,
+    compat_cookies,
+    compat_etree_Element,
+    compat_etree_fromstring,
+    compat_getpass,
+    compat_integer_types,
     compat_http_client,
-    compat_urllib_error,
-    compat_urllib_parse_urlparse,
+    compat_os_name,
     compat_str,
-
+    compat_urllib_error,
+    compat_urllib_parse_unquote,
+    compat_urllib_parse_urlencode,
+    compat_urllib_request,
+    compat_urlparse,
+    compat_xml_parse_error,
+)
+from ..downloader.f4m import (
+    get_base_url,
+    remove_encrypted_media,
+)
+from ..utils import (
+    NO_DEFAULT,
+    age_restricted,
+    base_url,
+    bug_reports_message,
     clean_html,
     compiled_regex_type,
+    determine_ext,
+    determine_protocol,
+    dict_get,
+    error_to_compat_str,
     ExtractorError,
+    extract_attributes,
+    fix_xml_ampersands,
+    float_or_none,
+    GeoRestrictedError,
+    GeoUtils,
+    int_or_none,
+    js_to_json,
+    JSON_LD_RE,
+    mimetype2ext,
+    orderedSet,
+    parse_bitrate,
+    parse_codecs,
+    parse_duration,
+    parse_iso8601,
+    parse_m3u8_attributes,
+    parse_resolution,
     RegexNotFoundError,
+    sanitized_Request,
     sanitize_filename,
+    str_or_none,
+    strip_or_none,
     unescapeHTML,
+    unified_strdate,
+    unified_timestamp,
+    update_Request,
+    update_url_query,
+    urljoin,
+    url_basename,
+    url_or_none,
+    xpath_element,
+    xpath_text,
+    xpath_with_ns,
 )
-_NO_DEFAULT = object()
 
 
 class InfoExtractor(object):
@@ -31,11 +89,15 @@ class InfoExtractor(object):
     information about the video (or videos) the URL refers to. This
     information includes the real video URL, the video title, author and
     others. The information is stored in a dictionary which is then
-    passed to the FileDownloader. The FileDownloader processes this
+    passed to the YoutubeDL. The YoutubeDL processes this
     information possibly downloading the video to the file system, among
     other possible outcomes.
 
-    The dictionaries must include the following fields:
+    The type field determines the type of the result.
+    By far the most common value (and the default if _type is missing) is
+    "video", which indicates a single video.
+
+    For a video, the dictionaries must include the following fields:
 
     id:             Video identifier.
     title:          Video title, unescaped.
@@ -46,8 +108,27 @@ class InfoExtractor(object):
                     from worst to best quality.
 
                     Potential fields:
-                    * url        Mandatory. The URL of the video file
-                    * ext        Will be calculated from url if missing
+                    * url        The mandatory URL representing the media:
+                                   for plain file media - HTTP URL of this file,
+                                   for RTMP - RTMP URL,
+                                   for HLS - URL of the M3U8 media playlist,
+                                   for HDS - URL of the F4M manifest,
+                                   for DASH
+                                     - HTTP URL to plain file media (in case of
+                                       unfragmented media)
+                                     - URL of the MPD manifest or base URL
+                                       representing the media if MPD manifest
+                                       is parsed from a string (in case of
+                                       fragmented media)
+                                   for MSS - URL of the ISM manifest.
+                    * manifest_url
+                                 The URL of the manifest file in case of
+                                 fragmented media:
+                                   for HLS - URL of the M3U8 master playlist,
+                                   for HDS - URL of the F4M manifest,
+                                   for DASH - URL of the MPD manifest,
+                                   for MSS - URL of the ISM manifest.
+                    * ext        Will be calculated from URL if missing
                     * format     A human-readable description of the format
                                  ("mp4 container with h264/opus").
                                  Calculated from the format_id, width, height.
@@ -65,22 +146,62 @@ class InfoExtractor(object):
                     * acodec     Name of the audio codec in use
                     * asr        Audio sampling rate in Hertz
                     * vbr        Average video bitrate in KBit/s
+                    * fps        Frame rate
                     * vcodec     Name of the video codec in use
                     * container  Name of the container format
                     * filesize   The number of bytes, if known in advance
+                    * filesize_approx  An estimate for the number of bytes
                     * player_url SWF Player URL (used for rtmpdump).
                     * protocol   The protocol that will be used for the actual
                                  download, lower-case.
-                                 "http", "https", "rtsp", "rtmp", "m3u8" or so.
+                                 "http", "https", "rtsp", "rtmp", "rtmpe",
+                                 "m3u8", "m3u8_native" or "http_dash_segments".
+                    * fragment_base_url
+                                 Base URL for fragments. Each fragment's path
+                                 value (if present) will be relative to
+                                 this URL.
+                    * fragments  A list of fragments of a fragmented media.
+                                 Each fragment entry must contain either an url
+                                 or a path. If an url is present it should be
+                                 considered by a client. Otherwise both path and
+                                 fragment_base_url must be present. Here is
+                                 the list of all potential fields:
+                                 * "url" - fragment's URL
+                                 * "path" - fragment's path relative to
+                                            fragment_base_url
+                                 * "duration" (optional, int or float)
+                                 * "filesize" (optional, int)
                     * preference Order number of this format. If this field is
                                  present and not None, the formats get sorted
                                  by this field, regardless of all other values.
                                  -1 for default (order by other properties),
                                  -2 or smaller for less than default.
+                                 < -1000 to hide the format (if there is
+                                    another one which is strictly better)
+                    * language   Language code, e.g. "de" or "en-US".
+                    * language_preference  Is this in the language mentioned in
+                                 the URL?
+                                 10 if it's what the URL is about,
+                                 -1 for default (don't know),
+                                 -10 otherwise, other values reserved for now.
                     * quality    Order number of the video quality of this
                                  format, irrespective of the file format.
                                  -1 for default (order by other properties),
                                  -2 or smaller for less than default.
+                    * source_preference  Order number for this video source
+                                  (quality takes higher priority)
+                                 -1 for default (order by other properties),
+                                 -2 or smaller for less than default.
+                    * http_headers  A dictionary of additional HTTP headers
+                                 to add to the request.
+                    * stretched_ratio  If given and not 1, indicates that the
+                                 video's pixels are not square.
+                                 width : height ratio as float.
+                    * no_resume  The server does not support resuming the
+                                 (HTTP or RTMP) download. Boolean.
+                    * downloader_options  A dictionary of downloader options as
+                                 described in FileDownloader
+
     url:            Final video URL.
     ext:            Video filename extension.
     format:         The video format, defaults to ext (used for --get-format)
@@ -88,55 +209,194 @@ class InfoExtractor(object):
 
     The following fields are optional:
 
+    alt_title:      A secondary title of the video.
     display_id      An alternative identifier for the video, not necessarily
                     unique, but available before title. Typically, id is
                     something like "4234987", title "Dancing naked mole rats",
                     and display_id "dancing-naked-mole-rats"
     thumbnails:     A list of dictionaries, with the following entries:
+                        * "id" (optional, string) - Thumbnail format ID
                         * "url"
+                        * "preference" (optional, int) - quality of the image
                         * "width" (optional, int)
                         * "height" (optional, int)
-                        * "resolution" (optional, string "{width}x{height"},
+                        * "resolution" (optional, string "{width}x{height}",
                                         deprecated)
+                        * "filesize" (optional, int)
     thumbnail:      Full URL to a video thumbnail image.
-    description:    One-line video description.
+    description:    Full video description.
     uploader:       Full name of the video uploader.
+    license:        License name the video is licensed under.
+    creator:        The creator of the video.
+    release_date:   The date (YYYYMMDD) when the video was released.
     timestamp:      UNIX timestamp of the moment the video became available.
     upload_date:    Video upload date (YYYYMMDD).
                     If not explicitly set, calculated from timestamp.
     uploader_id:    Nickname or id of the video uploader.
-    location:       Physical location of the video.
-    subtitles:      The subtitle file contents as a dictionary in the format
-                    {language: subtitles}.
-    duration:       Length of the video in seconds, as an integer.
+    uploader_url:   Full URL to a personal webpage of the video uploader.
+    channel:        Full name of the channel the video is uploaded on.
+                    Note that channel fields may or may not repeat uploader
+                    fields. This depends on a particular extractor.
+    channel_id:     Id of the channel.
+    channel_url:    Full URL to a channel webpage.
+    location:       Physical location where the video was filmed.
+    subtitles:      The available subtitles as a dictionary in the format
+                    {tag: subformats}. "tag" is usually a language code, and
+                    "subformats" is a list sorted from lower to higher
+                    preference, each element is a dictionary with the "ext"
+                    entry and one of:
+                        * "data": The subtitles file contents
+                        * "url": A URL pointing to the subtitles file
+                    "ext" will be calculated from URL if missing
+    automatic_captions: Like 'subtitles', used by the YoutubeIE for
+                    automatically generated captions
+    duration:       Length of the video in seconds, as an integer or float.
     view_count:     How many users have watched the video on the platform.
     like_count:     Number of positive ratings of the video
     dislike_count:  Number of negative ratings of the video
+    repost_count:   Number of reposts of the video
+    average_rating: Average rating give by users, the scale used depends on the webpage
     comment_count:  Number of comments on the video
+    comments:       A list of comments, each with one or more of the following
+                    properties (all but one of text or html optional):
+                        * "author" - human-readable name of the comment author
+                        * "author_id" - user ID of the comment author
+                        * "id" - Comment ID
+                        * "html" - Comment as HTML
+                        * "text" - Plain text of the comment
+                        * "timestamp" - UNIX timestamp of comment
+                        * "parent" - ID of the comment this one is replying to.
+                                     Set to "root" to indicate that this is a
+                                     comment to the original video.
     age_limit:      Age restriction for the video, as an integer (years)
-    webpage_url:    The url to the video webpage, if given to youtube-dl it
+    webpage_url:    The URL to the video webpage, if given to youtube-dl it
                     should allow to get the same result again. (It will be set
                     by YoutubeDL if it's missing)
     categories:     A list of categories that the video falls in, for example
                     ["Sports", "Berlin"]
+    tags:           A list of tags assigned to the video, e.g. ["sweden", "pop music"]
+    is_live:        True, False, or None (=unknown). Whether this video is a
+                    live stream that goes on instead of a fixed-length video.
+    start_time:     Time in seconds where the reproduction should start, as
+                    specified in the URL.
+    end_time:       Time in seconds where the reproduction should end, as
+                    specified in the URL.
+    chapters:       A list of dictionaries, with the following entries:
+                        * "start_time" - The start time of the chapter in seconds
+                        * "end_time" - The end time of the chapter in seconds
+                        * "title" (optional, string)
+
+    The following fields should only be used when the video belongs to some logical
+    chapter or section:
+
+    chapter:        Name or title of the chapter the video belongs to.
+    chapter_number: Number of the chapter the video belongs to, as an integer.
+    chapter_id:     Id of the chapter the video belongs to, as a unicode string.
+
+    The following fields should only be used when the video is an episode of some
+    series, programme or podcast:
+
+    series:         Title of the series or programme the video episode belongs to.
+    season:         Title of the season the video episode belongs to.
+    season_number:  Number of the season the video episode belongs to, as an integer.
+    season_id:      Id of the season the video episode belongs to, as a unicode string.
+    episode:        Title of the video episode. Unlike mandatory video title field,
+                    this field should denote the exact title of the video episode
+                    without any kind of decoration.
+    episode_number: Number of the video episode within a season, as an integer.
+    episode_id:     Id of the video episode, as a unicode string.
+
+    The following fields should only be used when the media is a track or a part of
+    a music album:
+
+    track:          Title of the track.
+    track_number:   Number of the track within an album or a disc, as an integer.
+    track_id:       Id of the track (useful in case of custom indexing, e.g. 6.iii),
+                    as a unicode string.
+    artist:         Artist(s) of the track.
+    genre:          Genre(s) of the track.
+    album:          Title of the album the track belongs to.
+    album_type:     Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
+    album_artist:   List of all artists appeared on the album (e.g.
+                    "Ash Borer / Fell Voices" or "Various Artists", useful for splits
+                    and compilations).
+    disc_number:    Number of the disc or other physical medium the track belongs to,
+                    as an integer.
+    release_year:   Year (YYYY) when the album was released.
 
     Unless mentioned otherwise, the fields should be Unicode strings.
 
+    Unless mentioned otherwise, None is equivalent to absence of information.
+
+
+    _type "playlist" indicates multiple videos.
+    There must be a key "entries", which is a list, an iterable, or a PagedList
+    object, each element of which is a valid dictionary by this specification.
+
+    Additionally, playlists can have "id", "title", "description", "uploader",
+    "uploader_id", "uploader_url" attributes with the same semantics as videos
+    (see above).
+
+
+    _type "multi_video" indicates that there are multiple videos that
+    form a single show, for examples multiple acts of an opera or TV episode.
+    It must have an entries key like a playlist and contain all the keys
+    required for a video at the same time.
+
+
+    _type "url" indicates that the video must be extracted from another
+    location, possibly by a different extractor. Its only required key is:
+    "url" - the next URL to extract.
+    The key "ie_key" can be set to the class name (minus the trailing "IE",
+    e.g. "Youtube") if the extractor class is known in advance.
+    Additionally, the dictionary may have any properties of the resolved entity
+    known in advance, for example "title" if the title of the referred video is
+    known ahead of time.
+
+
+    _type "url_transparent" entities have the same specification as "url", but
+    indicate that the given additional information is more precise than the one
+    associated with the resolved URL.
+    This is useful when a site employs a video service that hosts the video and
+    its technical metadata, but that video service does not embed a useful
+    title, description etc.
+
+
     Subclasses of this one should re-define the _real_initialize() and
     _real_extract() methods and define a _VALID_URL regexp.
     Probably, they should also be added to the list of extractors.
 
+    _GEO_BYPASS attribute may be set to False in order to disable
+    geo restriction bypass mechanisms for a particular extractor.
+    Though it won't disable explicit geo restriction bypass based on
+    country code provided with geo_bypass_country.
+
+    _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
+    countries for this extractor. One of these countries will be used by
+    geo restriction bypass mechanism right away in order to bypass
+    geo restriction, of course, if the mechanism is not disabled.
+
+    _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted
+    IP blocks in CIDR notation for this extractor. One of these IP blocks
+    will be used by geo restriction bypass mechanism similarly
+    to _GEO_COUNTRIES.
+
     Finally, the _WORKING attribute should be set to False for broken IEs
     in order to warn the users and skip the tests.
     """
 
     _ready = False
     _downloader = None
+    _x_forwarded_for_ip = None
+    _GEO_BYPASS = True
+    _GEO_COUNTRIES = None
+    _GEO_IP_BLOCKS = None
     _WORKING = True
 
     def __init__(self, downloader=None):
         """Constructor. Receives an optional downloader."""
         self._ready = False
+        self._x_forwarded_for_ip = None
         self.set_downloader(downloader)
 
     @classmethod
@@ -150,6 +410,14 @@ class InfoExtractor(object):
             cls._VALID_URL_RE = re.compile(cls._VALID_URL)
         return cls._VALID_URL_RE.match(url) is not None
 
+    @classmethod
+    def _match_id(cls, url):
+        if '_VALID_URL_RE' not in cls.__dict__:
+            cls._VALID_URL_RE = re.compile(cls._VALID_URL)
+        m = cls._VALID_URL_RE.match(url)
+        assert m
+        return compat_str(m.group('id'))
+
     @classmethod
     def working(cls):
         """Getter method for _WORKING."""
@@ -157,14 +425,137 @@ class InfoExtractor(object):
 
     def initialize(self):
         """Initializes an instance (authentication, etc)."""
+        self._initialize_geo_bypass({
+            'countries': self._GEO_COUNTRIES,
+            'ip_blocks': self._GEO_IP_BLOCKS,
+        })
         if not self._ready:
             self._real_initialize()
             self._ready = True
 
+    def _initialize_geo_bypass(self, geo_bypass_context):
+        """
+        Initialize geo restriction bypass mechanism.
+
+        This method is used to initialize geo bypass mechanism based on faking
+        X-Forwarded-For HTTP header. A random country from provided country list
+        is selected and a random IP belonging to this country is generated. This
+        IP will be passed as X-Forwarded-For HTTP header in all subsequent
+        HTTP requests.
+
+        This method will be used for initial geo bypass mechanism initialization
+        during the instance initialization with _GEO_COUNTRIES and
+        _GEO_IP_BLOCKS.
+
+        You may also manually call it from extractor's code if geo bypass
+        information is not available beforehand (e.g. obtained during
+        extraction) or due to some other reason. In this case you should pass
+        this information in geo bypass context passed as first argument. It may
+        contain following fields:
+
+        countries:  List of geo unrestricted countries (similar
+                    to _GEO_COUNTRIES)
+        ip_blocks:  List of geo unrestricted IP blocks in CIDR notation
+                    (similar to _GEO_IP_BLOCKS)
+
+        """
+        if not self._x_forwarded_for_ip:
+
+            # Geo bypass mechanism is explicitly disabled by user
+            if not self._downloader.params.get('geo_bypass', True):
+                return
+
+            if not geo_bypass_context:
+                geo_bypass_context = {}
+
+            # Backward compatibility: previously _initialize_geo_bypass
+            # expected a list of countries, some 3rd party code may still use
+            # it this way
+            if isinstance(geo_bypass_context, (list, tuple)):
+                geo_bypass_context = {
+                    'countries': geo_bypass_context,
+                }
+
+            # The whole point of geo bypass mechanism is to fake IP
+            # as X-Forwarded-For HTTP header based on some IP block or
+            # country code.
+
+            # Path 1: bypassing based on IP block in CIDR notation
+
+            # Explicit IP block specified by user, use it right away
+            # regardless of whether extractor is geo bypassable or not
+            ip_block = self._downloader.params.get('geo_bypass_ip_block', None)
+
+            # Otherwise use random IP block from geo bypass context but only
+            # if extractor is known as geo bypassable
+            if not ip_block:
+                ip_blocks = geo_bypass_context.get('ip_blocks')
+                if self._GEO_BYPASS and ip_blocks:
+                    ip_block = random.choice(ip_blocks)
+
+            if ip_block:
+                self._x_forwarded_for_ip = GeoUtils.random_ipv4(ip_block)
+                if self._downloader.params.get('verbose', False):
+                    self._downloader.to_screen(
+                        '[debug] Using fake IP %s as X-Forwarded-For.'
+                        % self._x_forwarded_for_ip)
+                return
+
+            # Path 2: bypassing based on country code
+
+            # Explicit country code specified by user, use it right away
+            # regardless of whether extractor is geo bypassable or not
+            country = self._downloader.params.get('geo_bypass_country', None)
+
+            # Otherwise use random country code from geo bypass context but
+            # only if extractor is known as geo bypassable
+            if not country:
+                countries = geo_bypass_context.get('countries')
+                if self._GEO_BYPASS and countries:
+                    country = random.choice(countries)
+
+            if country:
+                self._x_forwarded_for_ip = GeoUtils.random_ipv4(country)
+                if self._downloader.params.get('verbose', False):
+                    self._downloader.to_screen(
+                        '[debug] Using fake IP %s (%s) as X-Forwarded-For.'
+                        % (self._x_forwarded_for_ip, country.upper()))
+
     def extract(self, url):
         """Extracts URL information and returns it in list of dicts."""
-        self.initialize()
-        return self._real_extract(url)
+        try:
+            for _ in range(2):
+                try:
+                    self.initialize()
+                    ie_result = self._real_extract(url)
+                    if self._x_forwarded_for_ip:
+                        ie_result['__x_forwarded_for_ip'] = self._x_forwarded_for_ip
+                    return ie_result
+                except GeoRestrictedError as e:
+                    if self.__maybe_fake_ip_and_retry(e.countries):
+                        continue
+                    raise
+        except ExtractorError:
+            raise
+        except compat_http_client.IncompleteRead as e:
+            raise ExtractorError('A network error has occurred.', cause=e, expected=True)
+        except (KeyError, StopIteration) as e:
+            raise ExtractorError('An extractor error has occurred.', cause=e)
+
+    def __maybe_fake_ip_and_retry(self, countries):
+        if (not self._downloader.params.get('geo_bypass_country', None)
+                and self._GEO_BYPASS
+                and self._downloader.params.get('geo_bypass', True)
+                and not self._x_forwarded_for_ip
+                and countries):
+            country_code = random.choice(countries)
+            self._x_forwarded_for_ip = GeoUtils.random_ipv4(country_code)
+            if self._x_forwarded_for_ip:
+                self.report_warning(
+                    'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
+                    % (self._x_forwarded_for_ip, country_code.upper()))
+                return True
+        return False
 
     def set_downloader(self, downloader):
         """Sets the downloader for this IE."""
@@ -181,48 +572,100 @@ class InfoExtractor(object):
     @classmethod
     def ie_key(cls):
         """A string for getting the InfoExtractor with get_info_extractor"""
-        return cls.__name__[:-2]
+        return compat_str(cls.__name__[:-2])
 
     @property
     def IE_NAME(self):
-        return type(self).__name__[:-2]
+        return compat_str(type(self).__name__[:-2])
+
+    @staticmethod
+    def __can_accept_status_code(err, expected_status):
+        assert isinstance(err, compat_urllib_error.HTTPError)
+        if expected_status is None:
+            return False
+        if isinstance(expected_status, compat_integer_types):
+            return err.code == expected_status
+        elif isinstance(expected_status, (list, tuple)):
+            return err.code in expected_status
+        elif callable(expected_status):
+            return expected_status(err.code) is True
+        else:
+            assert False
+
+    def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers={}, query={}, expected_status=None):
+        """
+        Return the response handle.
 
-    def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
-        """ Returns the response handle """
+        See _download_webpage docstring for arguments specification.
+        """
         if note is None:
             self.report_download_webpage(video_id)
         elif note is not False:
             if video_id is None:
-                self.to_screen(u'%s' % (note,))
+                self.to_screen('%s' % (note,))
             else:
-                self.to_screen(u'%s: %s' % (video_id, note))
+                self.to_screen('%s: %s' % (video_id, note))
+
+        # Some sites check X-Forwarded-For HTTP header in order to figure out
+        # the origin of the client behind proxy. This allows bypassing geo
+        # restriction by faking this header's value to IP that belongs to some
+        # geo unrestricted country. We will do so once we encounter any
+        # geo restriction error.
+        if self._x_forwarded_for_ip:
+            if 'X-Forwarded-For' not in headers:
+                headers['X-Forwarded-For'] = self._x_forwarded_for_ip
+
+        if isinstance(url_or_request, compat_urllib_request.Request):
+            url_or_request = update_Request(
+                url_or_request, data=data, headers=headers, query=query)
+        else:
+            if query:
+                url_or_request = update_url_query(url_or_request, query)
+            if data is not None or headers:
+                url_or_request = sanitized_Request(url_or_request, data, headers)
         try:
             return self._downloader.urlopen(url_or_request)
         except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
+            if isinstance(err, compat_urllib_error.HTTPError):
+                if self.__can_accept_status_code(err, expected_status):
+                    # Retain reference to error to prevent file object from
+                    # being closed before it can be read. Works around the
+                    # effects of <https://bugs.python.org/issue15002>
+                    # introduced in Python 3.4.1.
+                    err.fp._error = err
+                    return err.fp
+
             if errnote is False:
                 return False
             if errnote is None:
-                errnote = u'Unable to download webpage'
-            errmsg = u'%s: %s' % (errnote, compat_str(err))
+                errnote = 'Unable to download webpage'
+
+            errmsg = '%s: %s' % (errnote, error_to_compat_str(err))
             if fatal:
                 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
             else:
                 self._downloader.report_warning(errmsg)
                 return False
 
-    def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
-        """ Returns a tuple (page content as string, URL handle) """
+    def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None, data=None, headers={}, query={}, expected_status=None):
+        """
+        Return a tuple (page content as string, URL handle).
 
+        See _download_webpage docstring for arguments specification.
+        """
         # Strip hashes from the URL (#1038)
         if isinstance(url_or_request, (compat_str, str)):
             url_or_request = url_or_request.partition('#')[0]
 
-        urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal)
+        urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query, expected_status=expected_status)
         if urlh is False:
             assert not fatal
             return False
-        content_type = urlh.headers.get('Content-Type', '')
-        webpage_bytes = urlh.read()
+        content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
+        return (content, urlh)
+
+    @staticmethod
+    def _guess_encoding_from_content(content_type, webpage_bytes):
         m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
         if m:
             encoding = m.group(1)
@@ -235,26 +678,62 @@ class InfoExtractor(object):
                 encoding = 'utf-16'
             else:
                 encoding = 'utf-8'
+
+        return encoding
+
+    def __check_blocked(self, content):
+        first_block = content[:512]
+        if ('<title>Access to this site is blocked</title>' in content
+                and 'Websense' in first_block):
+            msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
+            blocked_iframe = self._html_search_regex(
+                r'<iframe src="([^"]+)"', content,
+                'Websense information URL', default=None)
+            if blocked_iframe:
+                msg += ' Visit %s for more details' % blocked_iframe
+            raise ExtractorError(msg, expected=True)
+        if '<title>The URL you requested has been blocked</title>' in first_block:
+            msg = (
+                'Access to this webpage has been blocked by Indian censorship. '
+                'Use a VPN or proxy server (with --proxy) to route around it.')
+            block_msg = self._html_search_regex(
+                r'</h1><p>(.*?)</p>',
+                content, 'block message', default=None)
+            if block_msg:
+                msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
+            raise ExtractorError(msg, expected=True)
+        if ('<title>TTK :: Š”Š¾ŃŃ‚ŃƒŠæ Šŗ Ń€ŠµŃŃƒŃ€ŃŃƒ Š¾Š³Ń€Š°Š½ŠøчŠµŠ½</title>' in content
+                and 'blocklist.rkn.gov.ru' in content):
+            raise ExtractorError(
+                'Access to this webpage has been blocked by decision of the Russian government. '
+                'Visit http://blocklist.rkn.gov.ru/ for a block reason.',
+                expected=True)
+
+    def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
+        content_type = urlh.headers.get('Content-Type', '')
+        webpage_bytes = urlh.read()
+        if prefix is not None:
+            webpage_bytes = prefix + webpage_bytes
+        if not encoding:
+            encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
         if self._downloader.params.get('dump_intermediate_pages', False):
-            try:
-                url = url_or_request.get_full_url()
-            except AttributeError:
-                url = url_or_request
-            self.to_screen(u'Dumping request to ' + url)
+            self.to_screen('Dumping request to ' + urlh.geturl())
             dump = base64.b64encode(webpage_bytes).decode('ascii')
             self._downloader.to_screen(dump)
         if self._downloader.params.get('write_pages', False):
-            try:
-                url = url_or_request.get_full_url()
-            except AttributeError:
-                url = url_or_request
-            basen = '%s_%s' % (video_id, url)
+            basen = '%s_%s' % (video_id, urlh.geturl())
             if len(basen) > 240:
-                h = u'___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
+                h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
                 basen = basen[:240 - len(h)] + h
             raw_filename = basen + '.dump'
             filename = sanitize_filename(raw_filename, restricted=True)
-            self.to_screen(u'Saving request to ' + filename)
+            self.to_screen('Saving request to ' + filename)
+            # Working around MAX_PATH limitation on Windows (see
+            # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
+            if compat_os_name == 'nt':
+                absfilepath = os.path.abspath(filename)
+                if len(absfilepath) > 259:
+                    filename = '\\\\?\\' + absfilepath
             with open(filename, 'wb') as outf:
                 outf.write(webpage_bytes)
 
@@ -263,89 +742,226 @@ class InfoExtractor(object):
         except LookupError:
             content = webpage_bytes.decode('utf-8', 'replace')
 
-        if (u'<title>Access to this site is blocked</title>' in content and
-                u'Websense' in content[:512]):
-            msg = u'Access to this webpage has been blocked by Websense filtering software in your network.'
-            blocked_iframe = self._html_search_regex(
-                r'<iframe src="([^"]+)"', content,
-                u'Websense information URL', default=None)
-            if blocked_iframe:
-                msg += u' Visit %s for more details' % blocked_iframe
-            raise ExtractorError(msg, expected=True)
+        self.__check_blocked(content)
 
-        return (content, urlh)
+        return content
 
-    def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
-        """ Returns the data of the page as a string """
-        res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal)
+    def _download_webpage(
+            self, url_or_request, video_id, note=None, errnote=None,
+            fatal=True, tries=1, timeout=5, encoding=None, data=None,
+            headers={}, query={}, expected_status=None):
+        """
+        Return the data of the page as a string.
+
+        Arguments:
+        url_or_request -- plain text URL as a string or
+            a compat_urllib_request.Requestobject
+        video_id -- Video/playlist/item identifier (string)
+
+        Keyword arguments:
+        note -- note printed before downloading (string)
+        errnote -- note printed in case of an error (string)
+        fatal -- flag denoting whether error should be considered fatal,
+            i.e. whether it should cause ExtractionError to be raised,
+            otherwise a warning will be reported and extraction continued
+        tries -- number of tries
+        timeout -- sleep interval between tries
+        encoding -- encoding for a page content decoding, guessed automatically
+            when not explicitly specified
+        data -- POST data (bytes)
+        headers -- HTTP headers (dict)
+        query -- URL query (dict)
+        expected_status -- allows to accept failed HTTP requests (non 2xx
+            status code) by explicitly specifying a set of accepted status
+            codes. Can be any of the following entities:
+                - an integer type specifying an exact failed status code to
+                  accept
+                - a list or a tuple of integer types specifying a list of
+                  failed status codes to accept
+                - a callable accepting an actual failed status code and
+                  returning True if it should be accepted
+            Note that this argument does not affect success status codes (2xx)
+            which are always accepted.
+        """
+
+        success = False
+        try_count = 0
+        while success is False:
+            try:
+                res = self._download_webpage_handle(
+                    url_or_request, video_id, note, errnote, fatal,
+                    encoding=encoding, data=data, headers=headers, query=query,
+                    expected_status=expected_status)
+                success = True
+            except compat_http_client.IncompleteRead as e:
+                try_count += 1
+                if try_count >= tries:
+                    raise e
+                self._sleep(timeout, video_id)
         if res is False:
             return res
         else:
             content, _ = res
             return content
 
-    def _download_xml(self, url_or_request, video_id,
-                      note=u'Downloading XML', errnote=u'Unable to download XML',
-                      transform_source=None, fatal=True):
-        """Return the xml as an xml.etree.ElementTree.Element"""
-        xml_string = self._download_webpage(
-            url_or_request, video_id, note, errnote, fatal=fatal)
-        if xml_string is False:
-            return xml_string
+    def _download_xml_handle(
+            self, url_or_request, video_id, note='Downloading XML',
+            errnote='Unable to download XML', transform_source=None,
+            fatal=True, encoding=None, data=None, headers={}, query={},
+            expected_status=None):
+        """
+        Return a tuple (xml as an compat_etree_Element, URL handle).
+
+        See _download_webpage docstring for arguments specification.
+        """
+        res = self._download_webpage_handle(
+            url_or_request, video_id, note, errnote, fatal=fatal,
+            encoding=encoding, data=data, headers=headers, query=query,
+            expected_status=expected_status)
+        if res is False:
+            return res
+        xml_string, urlh = res
+        return self._parse_xml(
+            xml_string, video_id, transform_source=transform_source,
+            fatal=fatal), urlh
+
+    def _download_xml(
+            self, url_or_request, video_id,
+            note='Downloading XML', errnote='Unable to download XML',
+            transform_source=None, fatal=True, encoding=None,
+            data=None, headers={}, query={}, expected_status=None):
+        """
+        Return the xml as an compat_etree_Element.
+
+        See _download_webpage docstring for arguments specification.
+        """
+        res = self._download_xml_handle(
+            url_or_request, video_id, note=note, errnote=errnote,
+            transform_source=transform_source, fatal=fatal, encoding=encoding,
+            data=data, headers=headers, query=query,
+            expected_status=expected_status)
+        return res if res is False else res[0]
+
+    def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
         if transform_source:
             xml_string = transform_source(xml_string)
-        return xml.etree.ElementTree.fromstring(xml_string.encode('utf-8'))
+        try:
+            return compat_etree_fromstring(xml_string.encode('utf-8'))
+        except compat_xml_parse_error as ve:
+            errmsg = '%s: Failed to parse XML ' % video_id
+            if fatal:
+                raise ExtractorError(errmsg, cause=ve)
+            else:
+                self.report_warning(errmsg + str(ve))
+
+    def _download_json_handle(
+            self, url_or_request, video_id, note='Downloading JSON metadata',
+            errnote='Unable to download JSON metadata', transform_source=None,
+            fatal=True, encoding=None, data=None, headers={}, query={},
+            expected_status=None):
+        """
+        Return a tuple (JSON object, URL handle).
 
-    def _download_json(self, url_or_request, video_id,
-                       note=u'Downloading JSON metadata',
-                       errnote=u'Unable to download JSON metadata',
-                       transform_source=None):
-        json_string = self._download_webpage(url_or_request, video_id, note, errnote)
+        See _download_webpage docstring for arguments specification.
+        """
+        res = self._download_webpage_handle(
+            url_or_request, video_id, note, errnote, fatal=fatal,
+            encoding=encoding, data=data, headers=headers, query=query,
+            expected_status=expected_status)
+        if res is False:
+            return res
+        json_string, urlh = res
+        return self._parse_json(
+            json_string, video_id, transform_source=transform_source,
+            fatal=fatal), urlh
+
+    def _download_json(
+            self, url_or_request, video_id, note='Downloading JSON metadata',
+            errnote='Unable to download JSON metadata', transform_source=None,
+            fatal=True, encoding=None, data=None, headers={}, query={},
+            expected_status=None):
+        """
+        Return the JSON object as a dict.
+
+        See _download_webpage docstring for arguments specification.
+        """
+        res = self._download_json_handle(
+            url_or_request, video_id, note=note, errnote=errnote,
+            transform_source=transform_source, fatal=fatal, encoding=encoding,
+            data=data, headers=headers, query=query,
+            expected_status=expected_status)
+        return res if res is False else res[0]
+
+    def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
         if transform_source:
             json_string = transform_source(json_string)
         try:
             return json.loads(json_string)
         except ValueError as ve:
-            raise ExtractorError('Failed to download JSON', cause=ve)
+            errmsg = '%s: Failed to parse JSON ' % video_id
+            if fatal:
+                raise ExtractorError(errmsg, cause=ve)
+            else:
+                self.report_warning(errmsg + str(ve))
 
     def report_warning(self, msg, video_id=None):
-        idstr = u'' if video_id is None else u'%s: ' % video_id
+        idstr = '' if video_id is None else '%s: ' % video_id
         self._downloader.report_warning(
-            u'[%s] %s%s' % (self.IE_NAME, idstr, msg))
+            '[%s] %s%s' % (self.IE_NAME, idstr, msg))
 
     def to_screen(self, msg):
         """Print msg to screen, prefixing it with '[ie_name]'"""
-        self._downloader.to_screen(u'[%s] %s' % (self.IE_NAME, msg))
+        self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
 
     def report_extraction(self, id_or_name):
         """Report information extraction."""
-        self.to_screen(u'%s: Extracting information' % id_or_name)
+        self.to_screen('%s: Extracting information' % id_or_name)
 
     def report_download_webpage(self, video_id):
         """Report webpage download."""
-        self.to_screen(u'%s: Downloading webpage' % video_id)
+        self.to_screen('%s: Downloading webpage' % video_id)
 
     def report_age_confirmation(self):
         """Report attempt to confirm age."""
-        self.to_screen(u'Confirming age')
+        self.to_screen('Confirming age')
 
     def report_login(self):
         """Report attempt to log in."""
-        self.to_screen(u'Logging in')
+        self.to_screen('Logging in')
+
+    @staticmethod
+    def raise_login_required(msg='This video is only available for registered users'):
+        raise ExtractorError(
+            '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
+            expected=True)
 
-    #Methods for following #608
     @staticmethod
-    def url_result(url, ie=None, video_id=None):
-        """Returns a url that points to a page that should be processed"""
-        #TODO: ie should be the class used for getting the info
+    def raise_geo_restricted(msg='This video is not available from your location due to geo restriction', countries=None):
+        raise GeoRestrictedError(msg, countries=countries)
+
+    # Methods for following #608
+    @staticmethod
+    def url_result(url, ie=None, video_id=None, video_title=None):
+        """Returns a URL that points to a page that should be processed"""
+        # TODO: ie should be the class used for getting the info
         video_info = {'_type': 'url',
                       'url': url,
                       'ie_key': ie}
         if video_id is not None:
             video_info['id'] = video_id
+        if video_title is not None:
+            video_info['title'] = video_title
         return video_info
+
+    def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None):
+        urls = orderedSet(
+            self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
+            for m in matches)
+        return self.playlist_result(
+            urls, playlist_id=playlist_id, playlist_title=playlist_title)
+
     @staticmethod
-    def playlist_result(entries, playlist_id=None, playlist_title=None):
+    def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
         """Returns a playlist"""
         video_info = {'_type': 'playlist',
                       'entries': entries}
@@ -353,9 +969,11 @@ class InfoExtractor(object):
             video_info['id'] = playlist_id
         if playlist_title:
             video_info['title'] = playlist_title
+        if playlist_description:
+            video_info['description'] = playlist_description
         return video_info
 
-    def _search_regex(self, pattern, string, name, default=_NO_DEFAULT, fatal=True, flags=0):
+    def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
         """
         Perform a regex search on the given string, using a single or a list of
         patterns returning the first matching group.
@@ -367,86 +985,130 @@ class InfoExtractor(object):
         else:
             for p in pattern:
                 mobj = re.search(p, string, flags)
-                if mobj: break
+                if mobj:
+                    break
 
-        if os.name != 'nt' and sys.stderr.isatty():
-            _name = u'\033[0;34m%s\033[0m' % name
+        if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty():
+            _name = '\033[0;34m%s\033[0m' % name
         else:
             _name = name
 
         if mobj:
-            # return the first matching group
-            return next(g for g in mobj.groups() if g is not None)
-        elif default is not _NO_DEFAULT:
+            if group is None:
+                # return the first matching group
+                return next(g for g in mobj.groups() if g is not None)
+            else:
+                return mobj.group(group)
+        elif default is not NO_DEFAULT:
             return default
         elif fatal:
-            raise RegexNotFoundError(u'Unable to extract %s' % _name)
+            raise RegexNotFoundError('Unable to extract %s' % _name)
         else:
-            self._downloader.report_warning(u'unable to extract %s; '
-                u'please report this issue on http://yt-dl.org/bug' % _name)
+            self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
             return None
 
-    def _html_search_regex(self, pattern, string, name, default=_NO_DEFAULT, fatal=True, flags=0):
+    def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
         """
         Like _search_regex, but strips HTML tags and unescapes entities.
         """
-        res = self._search_regex(pattern, string, name, default, fatal, flags)
+        res = self._search_regex(pattern, string, name, default, fatal, flags, group)
         if res:
             return clean_html(res).strip()
         else:
             return res
 
-    def _get_login_info(self):
-        """
-        Get the the login info as (username, password)
-        It will look in the netrc file using the _NETRC_MACHINE value
-        If there's no info available, return (None, None)
-        """
-        if self._downloader is None:
-            return (None, None)
-
+    def _get_netrc_login_info(self, netrc_machine=None):
         username = None
         password = None
-        downloader_params = self._downloader.params
+        netrc_machine = netrc_machine or self._NETRC_MACHINE
 
-        # Attempt to use provided username and password or .netrc data
-        if downloader_params.get('username', None) is not None:
-            username = downloader_params['username']
-            password = downloader_params['password']
-        elif downloader_params.get('usenetrc', False):
+        if self._downloader.params.get('usenetrc', False):
             try:
-                info = netrc.netrc().authenticators(self._NETRC_MACHINE)
+                info = netrc.netrc().authenticators(netrc_machine)
                 if info is not None:
                     username = info[0]
                     password = info[2]
                 else:
-                    raise netrc.NetrcParseError('No authenticators for %s' % self._NETRC_MACHINE)
+                    raise netrc.NetrcParseError(
+                        'No authenticators for %s' % netrc_machine)
             except (IOError, netrc.NetrcParseError) as err:
-                self._downloader.report_warning(u'parsing .netrc: %s' % compat_str(err))
-        
-        return (username, password)
+                self._downloader.report_warning(
+                    'parsing .netrc: %s' % error_to_compat_str(err))
+
+        return username, password
+
+    def _get_login_info(self, username_option='username', password_option='password', netrc_machine=None):
+        """
+        Get the login info as (username, password)
+        First look for the manually specified credentials using username_option
+        and password_option as keys in params dictionary. If no such credentials
+        available look in the netrc file using the netrc_machine or _NETRC_MACHINE
+        value.
+        If there's no info available, return (None, None)
+        """
+        if self._downloader is None:
+            return (None, None)
+
+        downloader_params = self._downloader.params
+
+        # Attempt to use provided username and password or .netrc data
+        if downloader_params.get(username_option) is not None:
+            username = downloader_params[username_option]
+            password = downloader_params[password_option]
+        else:
+            username, password = self._get_netrc_login_info(netrc_machine)
+
+        return username, password
+
+    def _get_tfa_info(self, note='two-factor verification code'):
+        """
+        Get the two-factor authentication info
+        TODO - asking the user will be required for sms/phone verify
+        currently just uses the command line option
+        If there's no info available, return None
+        """
+        if self._downloader is None:
+            return None
+        downloader_params = self._downloader.params
+
+        if downloader_params.get('twofactor') is not None:
+            return downloader_params['twofactor']
+
+        return compat_getpass('Type %s and press [Return]: ' % note)
 
     # Helper functions for extracting OpenGraph info
     @staticmethod
     def _og_regexes(prop):
-        content_re = r'content=(?:"([^>]+?)"|\'([^>]+?)\')'
-        property_re = r'(?:name|property)=[\'"]og:%s[\'"]' % re.escape(prop)
+        content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
+        property_re = (r'(?:name|property)=(?:\'og[:-]%(prop)s\'|"og[:-]%(prop)s"|\s*og[:-]%(prop)s\b)'
+                       % {'prop': re.escape(prop)})
         template = r'<meta[^>]+?%s[^>]+?%s'
         return [
             template % (property_re, content_re),
             template % (content_re, property_re),
         ]
 
+    @staticmethod
+    def _meta_regex(prop):
+        return r'''(?isx)<meta
+                    (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
+                    [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
+
     def _og_search_property(self, prop, html, name=None, **kargs):
+        if not isinstance(prop, (list, tuple)):
+            prop = [prop]
         if name is None:
-            name = 'OpenGraph %s' % prop
-        escaped = self._search_regex(self._og_regexes(prop), html, name, flags=re.DOTALL, **kargs)
+            name = 'OpenGraph %s' % prop[0]
+        og_regexes = []
+        for p in prop:
+            og_regexes.extend(self._og_regexes(p))
+        escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
         if escaped is None:
             return None
         return unescapeHTML(escaped)
 
     def _og_search_thumbnail(self, html, **kargs):
-        return self._og_search_property('image', html, u'thumbnail url', fatal=False, **kargs)
+        return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
 
     def _og_search_description(self, html, **kargs):
         return self._og_search_property('description', html, fatal=False, **kargs)
@@ -455,21 +1117,22 @@ class InfoExtractor(object):
         return self._og_search_property('title', html, **kargs)
 
     def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
-        regexes = self._og_regexes('video')
-        if secure: regexes = self._og_regexes('video:secure_url') + regexes
+        regexes = self._og_regexes('video') + self._og_regexes('video:url')
+        if secure:
+            regexes = self._og_regexes('video:secure_url') + regexes
         return self._html_search_regex(regexes, html, name, **kargs)
 
     def _og_search_url(self, html, **kargs):
         return self._og_search_property('url', html, **kargs)
 
-    def _html_search_meta(self, name, html, display_name=None, fatal=False):
+    def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
+        if not isinstance(name, (list, tuple)):
+            name = [name]
         if display_name is None:
-            display_name = name
+            display_name = name[0]
         return self._html_search_regex(
-            r'''(?ix)<meta
-                    (?=[^>]+(?:itemprop|name|property)=["\']%s["\'])
-                    [^>]+content=["\']([^"\']+)["\']''' % re.escape(name),
-            html, display_name, fatal=fatal)
+            [self._meta_regex(n) for n in name],
+            html, display_name, fatal=fatal, group='content', **kwargs)
 
     def _dc_search_uploader(self, html):
         return self._html_search_meta('dc.creator', html, 'uploader')
@@ -496,15 +1159,178 @@ class InfoExtractor(object):
             'mature': 17,
             'restricted': 19,
         }
-        return RATING_TABLE.get(rating.lower(), None)
+        return RATING_TABLE.get(rating.lower())
+
+    def _family_friendly_search(self, html):
+        # See http://schema.org/VideoObject
+        family_friendly = self._html_search_meta(
+            'isFamilyFriendly', html, default=None)
+
+        if not family_friendly:
+            return None
+
+        RATING_TABLE = {
+            '1': 0,
+            'true': 0,
+            '0': 18,
+            'false': 18,
+        }
+        return RATING_TABLE.get(family_friendly.lower())
 
     def _twitter_search_player(self, html):
         return self._html_search_meta('twitter:player', html,
-            'twitter card player')
+                                      'twitter card player')
+
+    def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
+        json_ld = self._search_regex(
+            JSON_LD_RE, html, 'JSON-LD', group='json_ld', **kwargs)
+        default = kwargs.get('default', NO_DEFAULT)
+        if not json_ld:
+            return default if default is not NO_DEFAULT else {}
+        # JSON-LD may be malformed and thus `fatal` should be respected.
+        # At the same time `default` may be passed that assumes `fatal=False`
+        # for _search_regex. Let's simulate the same behavior here as well.
+        fatal = kwargs.get('fatal', True) if default == NO_DEFAULT else False
+        return self._json_ld(json_ld, video_id, fatal=fatal, expected_type=expected_type)
+
+    def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None):
+        if isinstance(json_ld, compat_str):
+            json_ld = self._parse_json(json_ld, video_id, fatal=fatal)
+        if not json_ld:
+            return {}
+        info = {}
+        if not isinstance(json_ld, (list, tuple, dict)):
+            return info
+        if isinstance(json_ld, dict):
+            json_ld = [json_ld]
+
+        INTERACTION_TYPE_MAP = {
+            'CommentAction': 'comment',
+            'AgreeAction': 'like',
+            'DisagreeAction': 'dislike',
+            'LikeAction': 'like',
+            'DislikeAction': 'dislike',
+            'ListenAction': 'view',
+            'WatchAction': 'view',
+            'ViewAction': 'view',
+        }
 
-    def _sort_formats(self, formats):
+        def extract_interaction_statistic(e):
+            interaction_statistic = e.get('interactionStatistic')
+            if not isinstance(interaction_statistic, list):
+                return
+            for is_e in interaction_statistic:
+                if not isinstance(is_e, dict):
+                    continue
+                if is_e.get('@type') != 'InteractionCounter':
+                    continue
+                interaction_type = is_e.get('interactionType')
+                if not isinstance(interaction_type, compat_str):
+                    continue
+                interaction_count = int_or_none(is_e.get('userInteractionCount'))
+                if interaction_count is None:
+                    continue
+                count_kind = INTERACTION_TYPE_MAP.get(interaction_type.split('/')[-1])
+                if not count_kind:
+                    continue
+                count_key = '%s_count' % count_kind
+                if info.get(count_key) is not None:
+                    continue
+                info[count_key] = interaction_count
+
+        def extract_video_object(e):
+            assert e['@type'] == 'VideoObject'
+            info.update({
+                'url': url_or_none(e.get('contentUrl')),
+                'title': unescapeHTML(e.get('name')),
+                'description': unescapeHTML(e.get('description')),
+                'thumbnail': url_or_none(e.get('thumbnailUrl') or e.get('thumbnailURL')),
+                'duration': parse_duration(e.get('duration')),
+                'timestamp': unified_timestamp(e.get('uploadDate')),
+                'filesize': float_or_none(e.get('contentSize')),
+                'tbr': int_or_none(e.get('bitrate')),
+                'width': int_or_none(e.get('width')),
+                'height': int_or_none(e.get('height')),
+                'view_count': int_or_none(e.get('interactionCount')),
+            })
+            extract_interaction_statistic(e)
+
+        for e in json_ld:
+            if isinstance(e.get('@context'), compat_str) and re.match(r'^https?://schema.org/?$', e.get('@context')):
+                item_type = e.get('@type')
+                if expected_type is not None and expected_type != item_type:
+                    return info
+                if item_type in ('TVEpisode', 'Episode'):
+                    episode_name = unescapeHTML(e.get('name'))
+                    info.update({
+                        'episode': episode_name,
+                        'episode_number': int_or_none(e.get('episodeNumber')),
+                        'description': unescapeHTML(e.get('description')),
+                    })
+                    if not info.get('title') and episode_name:
+                        info['title'] = episode_name
+                    part_of_season = e.get('partOfSeason')
+                    if isinstance(part_of_season, dict) and part_of_season.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'):
+                        info.update({
+                            'season': unescapeHTML(part_of_season.get('name')),
+                            'season_number': int_or_none(part_of_season.get('seasonNumber')),
+                        })
+                    part_of_series = e.get('partOfSeries') or e.get('partOfTVSeries')
+                    if isinstance(part_of_series, dict) and part_of_series.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
+                        info['series'] = unescapeHTML(part_of_series.get('name'))
+                elif item_type == 'Movie':
+                    info.update({
+                        'title': unescapeHTML(e.get('name')),
+                        'description': unescapeHTML(e.get('description')),
+                        'duration': parse_duration(e.get('duration')),
+                        'timestamp': unified_timestamp(e.get('dateCreated')),
+                    })
+                elif item_type in ('Article', 'NewsArticle'):
+                    info.update({
+                        'timestamp': parse_iso8601(e.get('datePublished')),
+                        'title': unescapeHTML(e.get('headline')),
+                        'description': unescapeHTML(e.get('articleBody')),
+                    })
+                elif item_type == 'VideoObject':
+                    extract_video_object(e)
+                    continue
+                video = e.get('video')
+                if isinstance(video, dict) and video.get('@type') == 'VideoObject':
+                    extract_video_object(video)
+                break
+        return dict((k, v) for k, v in info.items() if v is not None)
+
+    @staticmethod
+    def _hidden_inputs(html):
+        html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
+        hidden_inputs = {}
+        for input in re.findall(r'(?i)(<input[^>]+>)', html):
+            attrs = extract_attributes(input)
+            if not input:
+                continue
+            if attrs.get('type') not in ('hidden', 'submit'):
+                continue
+            name = attrs.get('name') or attrs.get('id')
+            value = attrs.get('value')
+            if name and value is not None:
+                hidden_inputs[name] = value
+        return hidden_inputs
+
+    def _form_hidden_inputs(self, form_id, html):
+        form = self._search_regex(
+            r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
+            html, '%s form' % form_id, group='form')
+        return self._hidden_inputs(form)
+
+    def _sort_formats(self, formats, field_preference=None):
         if not formats:
-            raise ExtractorError(u'No video formats found')
+            raise ExtractorError('No video formats found')
+
+        for f in formats:
+            # Automatically determine tbr when missing based on abr and vbr (improves
+            # formats sorting in some cases)
+            if 'tbr' not in f and f.get('abr') is not None and f.get('vbr') is not None:
+                f['tbr'] = f['abr'] + f['vbr']
 
         def _formats_key(f):
             # TODO remove the following workaround
@@ -512,31 +1338,40 @@ class InfoExtractor(object):
             if not f.get('ext') and 'url' in f:
                 f['ext'] = determine_ext(f['url'])
 
+            if isinstance(field_preference, (list, tuple)):
+                return tuple(
+                    f.get(field)
+                    if f.get(field) is not None
+                    else ('' if field == 'format_id' else -1)
+                    for field in field_preference)
+
             preference = f.get('preference')
             if preference is None:
-                proto = f.get('protocol')
-                if proto is None:
-                    proto = compat_urllib_parse_urlparse(f.get('url', '')).scheme
-
-                preference = 0 if proto in ['http', 'https'] else -0.1
+                preference = 0
                 if f.get('ext') in ['f4f', 'f4m']:  # Not yet supported
                     preference -= 0.5
 
+            protocol = f.get('protocol') or determine_protocol(f)
+            proto_preference = 0 if protocol in ['http', 'https'] else (-0.5 if protocol == 'rtsp' else -0.1)
+
             if f.get('vcodec') == 'none':  # audio only
+                preference -= 50
                 if self._downloader.params.get('prefer_free_formats'):
-                    ORDER = [u'aac', u'mp3', u'm4a', u'webm', u'ogg', u'opus']
+                    ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
                 else:
-                    ORDER = [u'webm', u'opus', u'ogg', u'mp3', u'aac', u'm4a']
+                    ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
                 ext_preference = 0
                 try:
                     audio_ext_preference = ORDER.index(f['ext'])
                 except ValueError:
                     audio_ext_preference = -1
             else:
+                if f.get('acodec') == 'none':  # video only
+                    preference -= 40
                 if self._downloader.params.get('prefer_free_formats'):
-                    ORDER = [u'flv', u'mp4', u'webm']
+                    ORDER = ['flv', 'mp4', 'webm']
                 else:
-                    ORDER = [u'webm', u'flv', u'mp4']
+                    ORDER = ['webm', 'flv', 'mp4']
                 try:
                     ext_preference = ORDER.index(f['ext'])
                 except ValueError:
@@ -545,21 +1380,57 @@ class InfoExtractor(object):
 
             return (
                 preference,
+                f.get('language_preference') if f.get('language_preference') is not None else -1,
                 f.get('quality') if f.get('quality') is not None else -1,
+                f.get('tbr') if f.get('tbr') is not None else -1,
+                f.get('filesize') if f.get('filesize') is not None else -1,
+                f.get('vbr') if f.get('vbr') is not None else -1,
                 f.get('height') if f.get('height') is not None else -1,
                 f.get('width') if f.get('width') is not None else -1,
+                proto_preference,
                 ext_preference,
-                f.get('tbr') if f.get('tbr') is not None else -1,
-                f.get('vbr') if f.get('vbr') is not None else -1,
                 f.get('abr') if f.get('abr') is not None else -1,
                 audio_ext_preference,
-                f.get('filesize') if f.get('filesize') is not None else -1,
-                f.get('format_id'),
+                f.get('fps') if f.get('fps') is not None else -1,
+                f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
+                f.get('source_preference') if f.get('source_preference') is not None else -1,
+                f.get('format_id') if f.get('format_id') is not None else '',
             )
         formats.sort(key=_formats_key)
 
+    def _check_formats(self, formats, video_id):
+        if formats:
+            formats[:] = filter(
+                lambda f: self._is_valid_url(
+                    f['url'], video_id,
+                    item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
+                formats)
+
+    @staticmethod
+    def _remove_duplicate_formats(formats):
+        format_urls = set()
+        unique_formats = []
+        for f in formats:
+            if f['url'] not in format_urls:
+                format_urls.add(f['url'])
+                unique_formats.append(f)
+        formats[:] = unique_formats
+
+    def _is_valid_url(self, url, video_id, item='video', headers={}):
+        url = self._proto_relative_url(url, scheme='http:')
+        # For now assume non HTTP(S) URLs always valid
+        if not (url.startswith('http://') or url.startswith('https://')):
+            return True
+        try:
+            self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
+            return True
+        except ExtractorError:
+            self.to_screen(
+                '%s: %s URL is invalid, skipping' % (video_id, item))
+            return False
+
     def http_scheme(self):
-        """ Either "https:" or "https:", depending on the user's preferences """
+        """ Either "http:" or "https:", depending on the user's preferences """
         return (
             'http:'
             if self._downloader.params.get('prefer_insecure', False)
@@ -575,11 +1446,1508 @@ class InfoExtractor(object):
         else:
             return url
 
+    def _sleep(self, timeout, video_id, msg_template=None):
+        if msg_template is None:
+            msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
+        msg = msg_template % {'video_id': video_id, 'timeout': timeout}
+        self.to_screen(msg)
+        time.sleep(timeout)
+
+    def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
+                             transform_source=lambda s: fix_xml_ampersands(s).strip(),
+                             fatal=True, m3u8_id=None, data=None, headers={}, query={}):
+        manifest = self._download_xml(
+            manifest_url, video_id, 'Downloading f4m manifest',
+            'Unable to download f4m manifest',
+            # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
+            # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244)
+            transform_source=transform_source,
+            fatal=fatal, data=data, headers=headers, query=query)
+
+        if manifest is False:
+            return []
+
+        return self._parse_f4m_formats(
+            manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
+            transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
+
+    def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
+                           transform_source=lambda s: fix_xml_ampersands(s).strip(),
+                           fatal=True, m3u8_id=None):
+        if not isinstance(manifest, compat_etree_Element) and not fatal:
+            return []
+
+        # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
+        akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
+        if akamai_pv is not None and ';' in akamai_pv.text:
+            playerVerificationChallenge = akamai_pv.text.split(';')[0]
+            if playerVerificationChallenge.strip() != '':
+                return []
+
+        formats = []
+        manifest_version = '1.0'
+        media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
+        if not media_nodes:
+            manifest_version = '2.0'
+            media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
+        # Remove unsupported DRM protected media from final formats
+        # rendition (see https://github.com/ytdl-org/youtube-dl/issues/8573).
+        media_nodes = remove_encrypted_media(media_nodes)
+        if not media_nodes:
+            return formats
+
+        manifest_base_url = get_base_url(manifest)
+
+        bootstrap_info = xpath_element(
+            manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
+            'bootstrap info', default=None)
+
+        vcodec = None
+        mime_type = xpath_text(
+            manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
+            'base URL', default=None)
+        if mime_type and mime_type.startswith('audio/'):
+            vcodec = 'none'
+
+        for i, media_el in enumerate(media_nodes):
+            tbr = int_or_none(media_el.attrib.get('bitrate'))
+            width = int_or_none(media_el.attrib.get('width'))
+            height = int_or_none(media_el.attrib.get('height'))
+            format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
+            # If <bootstrapInfo> is present, the specified f4m is a
+            # stream-level manifest, and only set-level manifests may refer to
+            # external resources.  See section 11.4 and section 4 of F4M spec
+            if bootstrap_info is None:
+                media_url = None
+                # @href is introduced in 2.0, see section 11.6 of F4M spec
+                if manifest_version == '2.0':
+                    media_url = media_el.attrib.get('href')
+                if media_url is None:
+                    media_url = media_el.attrib.get('url')
+                if not media_url:
+                    continue
+                manifest_url = (
+                    media_url if media_url.startswith('http://') or media_url.startswith('https://')
+                    else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
+                # If media_url is itself a f4m manifest do the recursive extraction
+                # since bitrates in parent manifest (this one) and media_url manifest
+                # may differ leading to inability to resolve the format by requested
+                # bitrate in f4m downloader
+                ext = determine_ext(manifest_url)
+                if ext == 'f4m':
+                    f4m_formats = self._extract_f4m_formats(
+                        manifest_url, video_id, preference=preference, f4m_id=f4m_id,
+                        transform_source=transform_source, fatal=fatal)
+                    # Sometimes stream-level manifest contains single media entry that
+                    # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
+                    # At the same time parent's media entry in set-level manifest may
+                    # contain it. We will copy it from parent in such cases.
+                    if len(f4m_formats) == 1:
+                        f = f4m_formats[0]
+                        f.update({
+                            'tbr': f.get('tbr') or tbr,
+                            'width': f.get('width') or width,
+                            'height': f.get('height') or height,
+                            'format_id': f.get('format_id') if not tbr else format_id,
+                            'vcodec': vcodec,
+                        })
+                    formats.extend(f4m_formats)
+                    continue
+                elif ext == 'm3u8':
+                    formats.extend(self._extract_m3u8_formats(
+                        manifest_url, video_id, 'mp4', preference=preference,
+                        m3u8_id=m3u8_id, fatal=fatal))
+                    continue
+            formats.append({
+                'format_id': format_id,
+                'url': manifest_url,
+                'manifest_url': manifest_url,
+                'ext': 'flv' if bootstrap_info is not None else None,
+                'protocol': 'f4m',
+                'tbr': tbr,
+                'width': width,
+                'height': height,
+                'vcodec': vcodec,
+                'preference': preference,
+            })
+        return formats
+
+    def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
+        return {
+            'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
+            'url': m3u8_url,
+            'ext': ext,
+            'protocol': 'm3u8',
+            'preference': preference - 100 if preference else -100,
+            'resolution': 'multiple',
+            'format_note': 'Quality selection URL',
+        }
+
+    def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
+                              entry_protocol='m3u8', preference=None,
+                              m3u8_id=None, note=None, errnote=None,
+                              fatal=True, live=False, data=None, headers={},
+                              query={}):
+        res = self._download_webpage_handle(
+            m3u8_url, video_id,
+            note=note or 'Downloading m3u8 information',
+            errnote=errnote or 'Failed to download m3u8 information',
+            fatal=fatal, data=data, headers=headers, query=query)
+
+        if res is False:
+            return []
+
+        m3u8_doc, urlh = res
+        m3u8_url = urlh.geturl()
+
+        return self._parse_m3u8_formats(
+            m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol,
+            preference=preference, m3u8_id=m3u8_id, live=live)
+
+    def _parse_m3u8_formats(self, m3u8_doc, m3u8_url, ext=None,
+                            entry_protocol='m3u8', preference=None,
+                            m3u8_id=None, live=False):
+        if '#EXT-X-FAXS-CM:' in m3u8_doc:  # Adobe Flash Access
+            return []
+
+        if re.search(r'#EXT-X-SESSION-KEY:.*?URI="skd://', m3u8_doc):  # Apple FairPlay
+            return []
+
+        formats = []
+
+        format_url = lambda u: (
+            u
+            if re.match(r'^https?://', u)
+            else compat_urlparse.urljoin(m3u8_url, u))
+
+        # References:
+        # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21
+        # 2. https://github.com/ytdl-org/youtube-dl/issues/12211
+        # 3. https://github.com/ytdl-org/youtube-dl/issues/18923
+
+        # We should try extracting formats only from master playlists [1, 4.3.4],
+        # i.e. playlists that describe available qualities. On the other hand
+        # media playlists [1, 4.3.3] should be returned as is since they contain
+        # just the media without qualities renditions.
+        # Fortunately, master playlist can be easily distinguished from media
+        # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4]
+        # master playlist tags MUST NOT appear in a media playist and vice versa.
+        # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every
+        # media playlist and MUST NOT appear in master playlist thus we can
+        # clearly detect media playlist with this criterion.
+
+        if '#EXT-X-TARGETDURATION' in m3u8_doc:  # media playlist, return as is
+            return [{
+                'url': m3u8_url,
+                'format_id': m3u8_id,
+                'ext': ext,
+                'protocol': entry_protocol,
+                'preference': preference,
+            }]
+
+        groups = {}
+        last_stream_inf = {}
+
+        def extract_media(x_media_line):
+            media = parse_m3u8_attributes(x_media_line)
+            # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED
+            media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME')
+            if not (media_type and group_id and name):
+                return
+            groups.setdefault(group_id, []).append(media)
+            if media_type not in ('VIDEO', 'AUDIO'):
+                return
+            media_url = media.get('URI')
+            if media_url:
+                format_id = []
+                for v in (m3u8_id, group_id, name):
+                    if v:
+                        format_id.append(v)
+                f = {
+                    'format_id': '-'.join(format_id),
+                    'url': format_url(media_url),
+                    'manifest_url': m3u8_url,
+                    'language': media.get('LANGUAGE'),
+                    'ext': ext,
+                    'protocol': entry_protocol,
+                    'preference': preference,
+                }
+                if media_type == 'AUDIO':
+                    f['vcodec'] = 'none'
+                formats.append(f)
+
+        def build_stream_name():
+            # Despite specification does not mention NAME attribute for
+            # EXT-X-STREAM-INF tag it still sometimes may be present (see [1]
+            # or vidio test in TestInfoExtractor.test_parse_m3u8_formats)
+            # 1. http://www.vidio.com/watch/165683-dj_ambred-booyah-live-2015
+            stream_name = last_stream_inf.get('NAME')
+            if stream_name:
+                return stream_name
+            # If there is no NAME in EXT-X-STREAM-INF it will be obtained
+            # from corresponding rendition group
+            stream_group_id = last_stream_inf.get('VIDEO')
+            if not stream_group_id:
+                return
+            stream_group = groups.get(stream_group_id)
+            if not stream_group:
+                return stream_group_id
+            rendition = stream_group[0]
+            return rendition.get('NAME') or stream_group_id
+
+        # parse EXT-X-MEDIA tags before EXT-X-STREAM-INF in order to have the
+        # chance to detect video only formats when EXT-X-STREAM-INF tags
+        # precede EXT-X-MEDIA tags in HLS manifest such as [3].
+        for line in m3u8_doc.splitlines():
+            if line.startswith('#EXT-X-MEDIA:'):
+                extract_media(line)
+
+        for line in m3u8_doc.splitlines():
+            if line.startswith('#EXT-X-STREAM-INF:'):
+                last_stream_inf = parse_m3u8_attributes(line)
+            elif line.startswith('#') or not line.strip():
+                continue
+            else:
+                tbr = float_or_none(
+                    last_stream_inf.get('AVERAGE-BANDWIDTH')
+                    or last_stream_inf.get('BANDWIDTH'), scale=1000)
+                format_id = []
+                if m3u8_id:
+                    format_id.append(m3u8_id)
+                stream_name = build_stream_name()
+                # Bandwidth of live streams may differ over time thus making
+                # format_id unpredictable. So it's better to keep provided
+                # format_id intact.
+                if not live:
+                    format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
+                manifest_url = format_url(line.strip())
+                f = {
+                    'format_id': '-'.join(format_id),
+                    'url': manifest_url,
+                    'manifest_url': m3u8_url,
+                    'tbr': tbr,
+                    'ext': ext,
+                    'fps': float_or_none(last_stream_inf.get('FRAME-RATE')),
+                    'protocol': entry_protocol,
+                    'preference': preference,
+                }
+                resolution = last_stream_inf.get('RESOLUTION')
+                if resolution:
+                    mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
+                    if mobj:
+                        f['width'] = int(mobj.group('width'))
+                        f['height'] = int(mobj.group('height'))
+                # Unified Streaming Platform
+                mobj = re.search(
+                    r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
+                if mobj:
+                    abr, vbr = mobj.groups()
+                    abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
+                    f.update({
+                        'vbr': vbr,
+                        'abr': abr,
+                    })
+                codecs = parse_codecs(last_stream_inf.get('CODECS'))
+                f.update(codecs)
+                audio_group_id = last_stream_inf.get('AUDIO')
+                # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which
+                # references a rendition group MUST have a CODECS attribute.
+                # However, this is not always respected, for example, [2]
+                # contains EXT-X-STREAM-INF tag which references AUDIO
+                # rendition group but does not have CODECS and despite
+                # referencing an audio group it represents a complete
+                # (with audio and video) format. So, for such cases we will
+                # ignore references to rendition groups and treat them
+                # as complete formats.
+                if audio_group_id and codecs and f.get('vcodec') != 'none':
+                    audio_group = groups.get(audio_group_id)
+                    if audio_group and audio_group[0].get('URI'):
+                        # TODO: update acodec for audio only formats with
+                        # the same GROUP-ID
+                        f['acodec'] = 'none'
+                formats.append(f)
+
+                # for DailyMotion
+                progressive_uri = last_stream_inf.get('PROGRESSIVE-URI')
+                if progressive_uri:
+                    http_f = f.copy()
+                    del http_f['manifest_url']
+                    http_f.update({
+                        'format_id': f['format_id'].replace('hls-', 'http-'),
+                        'protocol': 'http',
+                        'url': progressive_uri,
+                    })
+                    formats.append(http_f)
+
+                last_stream_inf = {}
+        return formats
+
+    @staticmethod
+    def _xpath_ns(path, namespace=None):
+        if not namespace:
+            return path
+        out = []
+        for c in path.split('/'):
+            if not c or c == '.':
+                out.append(c)
+            else:
+                out.append('{%s}%s' % (namespace, c))
+        return '/'.join(out)
+
+    def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
+        smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
+
+        if smil is False:
+            assert not fatal
+            return []
+
+        namespace = self._parse_smil_namespace(smil)
+
+        return self._parse_smil_formats(
+            smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
+
+    def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
+        smil = self._download_smil(smil_url, video_id, fatal=fatal)
+        if smil is False:
+            return {}
+        return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
+
+    def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
+        return self._download_xml(
+            smil_url, video_id, 'Downloading SMIL file',
+            'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
+
+    def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
+        namespace = self._parse_smil_namespace(smil)
+
+        formats = self._parse_smil_formats(
+            smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
+        subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
+
+        video_id = os.path.splitext(url_basename(smil_url))[0]
+        title = None
+        description = None
+        upload_date = None
+        for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
+            name = meta.attrib.get('name')
+            content = meta.attrib.get('content')
+            if not name or not content:
+                continue
+            if not title and name == 'title':
+                title = content
+            elif not description and name in ('description', 'abstract'):
+                description = content
+            elif not upload_date and name == 'date':
+                upload_date = unified_strdate(content)
+
+        thumbnails = [{
+            'id': image.get('type'),
+            'url': image.get('src'),
+            'width': int_or_none(image.get('width')),
+            'height': int_or_none(image.get('height')),
+        } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
+
+        return {
+            'id': video_id,
+            'title': title or video_id,
+            'description': description,
+            'upload_date': upload_date,
+            'thumbnails': thumbnails,
+            'formats': formats,
+            'subtitles': subtitles,
+        }
+
+    def _parse_smil_namespace(self, smil):
+        return self._search_regex(
+            r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
+
+    def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
+        base = smil_url
+        for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
+            b = meta.get('base') or meta.get('httpBase')
+            if b:
+                base = b
+                break
+
+        formats = []
+        rtmp_count = 0
+        http_count = 0
+        m3u8_count = 0
+
+        srcs = []
+        media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
+        for medium in media:
+            src = medium.get('src')
+            if not src or src in srcs:
+                continue
+            srcs.append(src)
+
+            bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
+            filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
+            width = int_or_none(medium.get('width'))
+            height = int_or_none(medium.get('height'))
+            proto = medium.get('proto')
+            ext = medium.get('ext')
+            src_ext = determine_ext(src)
+            streamer = medium.get('streamer') or base
+
+            if proto == 'rtmp' or streamer.startswith('rtmp'):
+                rtmp_count += 1
+                formats.append({
+                    'url': streamer,
+                    'play_path': src,
+                    'ext': 'flv',
+                    'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
+                    'tbr': bitrate,
+                    'filesize': filesize,
+                    'width': width,
+                    'height': height,
+                })
+                if transform_rtmp_url:
+                    streamer, src = transform_rtmp_url(streamer, src)
+                    formats[-1].update({
+                        'url': streamer,
+                        'play_path': src,
+                    })
+                continue
+
+            src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
+            src_url = src_url.strip()
+
+            if proto == 'm3u8' or src_ext == 'm3u8':
+                m3u8_formats = self._extract_m3u8_formats(
+                    src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
+                if len(m3u8_formats) == 1:
+                    m3u8_count += 1
+                    m3u8_formats[0].update({
+                        'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
+                        'tbr': bitrate,
+                        'width': width,
+                        'height': height,
+                    })
+                formats.extend(m3u8_formats)
+            elif src_ext == 'f4m':
+                f4m_url = src_url
+                if not f4m_params:
+                    f4m_params = {
+                        'hdcore': '3.2.0',
+                        'plugin': 'flowplayer-3.2.0.1',
+                    }
+                f4m_url += '&' if '?' in f4m_url else '?'
+                f4m_url += compat_urllib_parse_urlencode(f4m_params)
+                formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
+            elif src_ext == 'mpd':
+                formats.extend(self._extract_mpd_formats(
+                    src_url, video_id, mpd_id='dash', fatal=False))
+            elif re.search(r'\.ism/[Mm]anifest', src_url):
+                formats.extend(self._extract_ism_formats(
+                    src_url, video_id, ism_id='mss', fatal=False))
+            elif src_url.startswith('http') and self._is_valid_url(src, video_id):
+                http_count += 1
+                formats.append({
+                    'url': src_url,
+                    'ext': ext or src_ext or 'flv',
+                    'format_id': 'http-%d' % (bitrate or http_count),
+                    'tbr': bitrate,
+                    'filesize': filesize,
+                    'width': width,
+                    'height': height,
+                })
+
+        return formats
+
+    def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
+        urls = []
+        subtitles = {}
+        for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
+            src = textstream.get('src')
+            if not src or src in urls:
+                continue
+            urls.append(src)
+            ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
+            lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
+            subtitles.setdefault(lang, []).append({
+                'url': src,
+                'ext': ext,
+            })
+        return subtitles
+
+    def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True):
+        xspf = self._download_xml(
+            xspf_url, playlist_id, 'Downloading xpsf playlist',
+            'Unable to download xspf manifest', fatal=fatal)
+        if xspf is False:
+            return []
+        return self._parse_xspf(
+            xspf, playlist_id, xspf_url=xspf_url,
+            xspf_base_url=base_url(xspf_url))
+
+    def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None):
+        NS_MAP = {
+            'xspf': 'http://xspf.org/ns/0/',
+            's1': 'http://static.streamone.nl/player/ns/0',
+        }
+
+        entries = []
+        for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
+            title = xpath_text(
+                track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
+            description = xpath_text(
+                track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
+            thumbnail = xpath_text(
+                track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
+            duration = float_or_none(
+                xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
+
+            formats = []
+            for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)):
+                format_url = urljoin(xspf_base_url, location.text)
+                if not format_url:
+                    continue
+                formats.append({
+                    'url': format_url,
+                    'manifest_url': xspf_url,
+                    'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
+                    'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
+                    'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
+                })
+            self._sort_formats(formats)
+
+            entries.append({
+                'id': playlist_id,
+                'title': title,
+                'description': description,
+                'thumbnail': thumbnail,
+                'duration': duration,
+                'formats': formats,
+            })
+        return entries
+
+    def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}, data=None, headers={}, query={}):
+        res = self._download_xml_handle(
+            mpd_url, video_id,
+            note=note or 'Downloading MPD manifest',
+            errnote=errnote or 'Failed to download MPD manifest',
+            fatal=fatal, data=data, headers=headers, query=query)
+        if res is False:
+            return []
+        mpd_doc, urlh = res
+        if mpd_doc is None:
+            return []
+        mpd_base_url = base_url(urlh.geturl())
+
+        return self._parse_mpd_formats(
+            mpd_doc, mpd_id=mpd_id, mpd_base_url=mpd_base_url,
+            formats_dict=formats_dict, mpd_url=mpd_url)
+
+    def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None):
+        """
+        Parse formats from MPD manifest.
+        References:
+         1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
+            http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
+         2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
+        """
+        if mpd_doc.get('type') == 'dynamic':
+            return []
+
+        namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
+
+        def _add_ns(path):
+            return self._xpath_ns(path, namespace)
+
+        def is_drm_protected(element):
+            return element.find(_add_ns('ContentProtection')) is not None
+
+        def extract_multisegment_info(element, ms_parent_info):
+            ms_info = ms_parent_info.copy()
+
+            # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
+            # common attributes and elements.  We will only extract relevant
+            # for us.
+            def extract_common(source):
+                segment_timeline = source.find(_add_ns('SegmentTimeline'))
+                if segment_timeline is not None:
+                    s_e = segment_timeline.findall(_add_ns('S'))
+                    if s_e:
+                        ms_info['total_number'] = 0
+                        ms_info['s'] = []
+                        for s in s_e:
+                            r = int(s.get('r', 0))
+                            ms_info['total_number'] += 1 + r
+                            ms_info['s'].append({
+                                't': int(s.get('t', 0)),
+                                # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
+                                'd': int(s.attrib['d']),
+                                'r': r,
+                            })
+                start_number = source.get('startNumber')
+                if start_number:
+                    ms_info['start_number'] = int(start_number)
+                timescale = source.get('timescale')
+                if timescale:
+                    ms_info['timescale'] = int(timescale)
+                segment_duration = source.get('duration')
+                if segment_duration:
+                    ms_info['segment_duration'] = float(segment_duration)
+
+            def extract_Initialization(source):
+                initialization = source.find(_add_ns('Initialization'))
+                if initialization is not None:
+                    ms_info['initialization_url'] = initialization.attrib['sourceURL']
+
+            segment_list = element.find(_add_ns('SegmentList'))
+            if segment_list is not None:
+                extract_common(segment_list)
+                extract_Initialization(segment_list)
+                segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
+                if segment_urls_e:
+                    ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
+            else:
+                segment_template = element.find(_add_ns('SegmentTemplate'))
+                if segment_template is not None:
+                    extract_common(segment_template)
+                    media = segment_template.get('media')
+                    if media:
+                        ms_info['media'] = media
+                    initialization = segment_template.get('initialization')
+                    if initialization:
+                        ms_info['initialization'] = initialization
+                    else:
+                        extract_Initialization(segment_template)
+            return ms_info
+
+        mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
+        formats = []
+        for period in mpd_doc.findall(_add_ns('Period')):
+            period_duration = parse_duration(period.get('duration')) or mpd_duration
+            period_ms_info = extract_multisegment_info(period, {
+                'start_number': 1,
+                'timescale': 1,
+            })
+            for adaptation_set in period.findall(_add_ns('AdaptationSet')):
+                if is_drm_protected(adaptation_set):
+                    continue
+                adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
+                for representation in adaptation_set.findall(_add_ns('Representation')):
+                    if is_drm_protected(representation):
+                        continue
+                    representation_attrib = adaptation_set.attrib.copy()
+                    representation_attrib.update(representation.attrib)
+                    # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
+                    mime_type = representation_attrib['mimeType']
+                    content_type = mime_type.split('/')[0]
+                    if content_type == 'text':
+                        # TODO implement WebVTT downloading
+                        pass
+                    elif content_type in ('video', 'audio'):
+                        base_url = ''
+                        for element in (representation, adaptation_set, period, mpd_doc):
+                            base_url_e = element.find(_add_ns('BaseURL'))
+                            if base_url_e is not None:
+                                base_url = base_url_e.text + base_url
+                                if re.match(r'^https?://', base_url):
+                                    break
+                        if mpd_base_url and not re.match(r'^https?://', base_url):
+                            if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
+                                mpd_base_url += '/'
+                            base_url = mpd_base_url + base_url
+                        representation_id = representation_attrib.get('id')
+                        lang = representation_attrib.get('lang')
+                        url_el = representation.find(_add_ns('BaseURL'))
+                        filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
+                        bandwidth = int_or_none(representation_attrib.get('bandwidth'))
+                        f = {
+                            'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
+                            'manifest_url': mpd_url,
+                            'ext': mimetype2ext(mime_type),
+                            'width': int_or_none(representation_attrib.get('width')),
+                            'height': int_or_none(representation_attrib.get('height')),
+                            'tbr': float_or_none(bandwidth, 1000),
+                            'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
+                            'fps': int_or_none(representation_attrib.get('frameRate')),
+                            'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
+                            'format_note': 'DASH %s' % content_type,
+                            'filesize': filesize,
+                            'container': mimetype2ext(mime_type) + '_dash',
+                        }
+                        f.update(parse_codecs(representation_attrib.get('codecs')))
+                        representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
+
+                        def prepare_template(template_name, identifiers):
+                            tmpl = representation_ms_info[template_name]
+                            # First of, % characters outside $...$ templates
+                            # must be escaped by doubling for proper processing
+                            # by % operator string formatting used further (see
+                            # https://github.com/ytdl-org/youtube-dl/issues/16867).
+                            t = ''
+                            in_template = False
+                            for c in tmpl:
+                                t += c
+                                if c == '$':
+                                    in_template = not in_template
+                                elif c == '%' and not in_template:
+                                    t += c
+                            # Next, $...$ templates are translated to their
+                            # %(...) counterparts to be used with % operator
+                            t = t.replace('$RepresentationID$', representation_id)
+                            t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
+                            t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
+                            t.replace('$$', '$')
+                            return t
+
+                        # @initialization is a regular template like @media one
+                        # so it should be handled just the same way (see
+                        # https://github.com/ytdl-org/youtube-dl/issues/11605)
+                        if 'initialization' in representation_ms_info:
+                            initialization_template = prepare_template(
+                                'initialization',
+                                # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
+                                # $Time$ shall not be included for @initialization thus
+                                # only $Bandwidth$ remains
+                                ('Bandwidth', ))
+                            representation_ms_info['initialization_url'] = initialization_template % {
+                                'Bandwidth': bandwidth,
+                            }
+
+                        def location_key(location):
+                            return 'url' if re.match(r'^https?://', location) else 'path'
+
+                        if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
+
+                            media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
+                            media_location_key = location_key(media_template)
+
+                            # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
+                            # can't be used at the same time
+                            if '%(Number' in media_template and 's' not in representation_ms_info:
+                                segment_duration = None
+                                if 'total_number' not in representation_ms_info and 'segment_duration' in representation_ms_info:
+                                    segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
+                                    representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
+                                representation_ms_info['fragments'] = [{
+                                    media_location_key: media_template % {
+                                        'Number': segment_number,
+                                        'Bandwidth': bandwidth,
+                                    },
+                                    'duration': segment_duration,
+                                } for segment_number in range(
+                                    representation_ms_info['start_number'],
+                                    representation_ms_info['total_number'] + representation_ms_info['start_number'])]
+                            else:
+                                # $Number*$ or $Time$ in media template with S list available
+                                # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
+                                # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
+                                representation_ms_info['fragments'] = []
+                                segment_time = 0
+                                segment_d = None
+                                segment_number = representation_ms_info['start_number']
+
+                                def add_segment_url():
+                                    segment_url = media_template % {
+                                        'Time': segment_time,
+                                        'Bandwidth': bandwidth,
+                                        'Number': segment_number,
+                                    }
+                                    representation_ms_info['fragments'].append({
+                                        media_location_key: segment_url,
+                                        'duration': float_or_none(segment_d, representation_ms_info['timescale']),
+                                    })
+
+                                for num, s in enumerate(representation_ms_info['s']):
+                                    segment_time = s.get('t') or segment_time
+                                    segment_d = s['d']
+                                    add_segment_url()
+                                    segment_number += 1
+                                    for r in range(s.get('r', 0)):
+                                        segment_time += segment_d
+                                        add_segment_url()
+                                        segment_number += 1
+                                    segment_time += segment_d
+                        elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
+                            # No media template
+                            # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
+                            # or any YouTube dashsegments video
+                            fragments = []
+                            segment_index = 0
+                            timescale = representation_ms_info['timescale']
+                            for s in representation_ms_info['s']:
+                                duration = float_or_none(s['d'], timescale)
+                                for r in range(s.get('r', 0) + 1):
+                                    segment_uri = representation_ms_info['segment_urls'][segment_index]
+                                    fragments.append({
+                                        location_key(segment_uri): segment_uri,
+                                        'duration': duration,
+                                    })
+                                    segment_index += 1
+                            representation_ms_info['fragments'] = fragments
+                        elif 'segment_urls' in representation_ms_info:
+                            # Segment URLs with no SegmentTimeline
+                            # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
+                            # https://github.com/ytdl-org/youtube-dl/pull/14844
+                            fragments = []
+                            segment_duration = float_or_none(
+                                representation_ms_info['segment_duration'],
+                                representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
+                            for segment_url in representation_ms_info['segment_urls']:
+                                fragment = {
+                                    location_key(segment_url): segment_url,
+                                }
+                                if segment_duration:
+                                    fragment['duration'] = segment_duration
+                                fragments.append(fragment)
+                            representation_ms_info['fragments'] = fragments
+                        # If there is a fragments key available then we correctly recognized fragmented media.
+                        # Otherwise we will assume unfragmented media with direct access. Technically, such
+                        # assumption is not necessarily correct since we may simply have no support for
+                        # some forms of fragmented media renditions yet, but for now we'll use this fallback.
+                        if 'fragments' in representation_ms_info:
+                            f.update({
+                                # NB: mpd_url may be empty when MPD manifest is parsed from a string
+                                'url': mpd_url or base_url,
+                                'fragment_base_url': base_url,
+                                'fragments': [],
+                                'protocol': 'http_dash_segments',
+                            })
+                            if 'initialization_url' in representation_ms_info:
+                                initialization_url = representation_ms_info['initialization_url']
+                                if not f.get('url'):
+                                    f['url'] = initialization_url
+                                f['fragments'].append({location_key(initialization_url): initialization_url})
+                            f['fragments'].extend(representation_ms_info['fragments'])
+                        else:
+                            # Assuming direct URL to unfragmented media.
+                            f['url'] = base_url
+
+                        # According to [1, 5.3.5.2, Table 7, page 35] @id of Representation
+                        # is not necessarily unique within a Period thus formats with
+                        # the same `format_id` are quite possible. There are numerous examples
+                        # of such manifests (see https://github.com/ytdl-org/youtube-dl/issues/15111,
+                        # https://github.com/ytdl-org/youtube-dl/issues/13919)
+                        full_info = formats_dict.get(representation_id, {}).copy()
+                        full_info.update(f)
+                        formats.append(full_info)
+                    else:
+                        self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
+        return formats
+
+    def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
+        res = self._download_xml_handle(
+            ism_url, video_id,
+            note=note or 'Downloading ISM manifest',
+            errnote=errnote or 'Failed to download ISM manifest',
+            fatal=fatal, data=data, headers=headers, query=query)
+        if res is False:
+            return []
+        ism_doc, urlh = res
+
+        return self._parse_ism_formats(ism_doc, urlh.geturl(), ism_id)
+
+    def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None):
+        """
+        Parse formats from ISM manifest.
+        References:
+         1. [MS-SSTR]: Smooth Streaming Protocol,
+            https://msdn.microsoft.com/en-us/library/ff469518.aspx
+        """
+        if ism_doc.get('IsLive') == 'TRUE' or ism_doc.find('Protection') is not None:
+            return []
+
+        duration = int(ism_doc.attrib['Duration'])
+        timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
+
+        formats = []
+        for stream in ism_doc.findall('StreamIndex'):
+            stream_type = stream.get('Type')
+            if stream_type not in ('video', 'audio'):
+                continue
+            url_pattern = stream.attrib['Url']
+            stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
+            stream_name = stream.get('Name')
+            for track in stream.findall('QualityLevel'):
+                fourcc = track.get('FourCC', 'AACL' if track.get('AudioTag') == '255' else None)
+                # TODO: add support for WVC1 and WMAP
+                if fourcc not in ('H264', 'AVC1', 'AACL'):
+                    self.report_warning('%s is not a supported codec' % fourcc)
+                    continue
+                tbr = int(track.attrib['Bitrate']) // 1000
+                # [1] does not mention Width and Height attributes. However,
+                # they're often present while MaxWidth and MaxHeight are
+                # missing, so should be used as fallbacks
+                width = int_or_none(track.get('MaxWidth') or track.get('Width'))
+                height = int_or_none(track.get('MaxHeight') or track.get('Height'))
+                sampling_rate = int_or_none(track.get('SamplingRate'))
+
+                track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
+                track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
+
+                fragments = []
+                fragment_ctx = {
+                    'time': 0,
+                }
+                stream_fragments = stream.findall('c')
+                for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
+                    fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
+                    fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
+                    fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
+                    if not fragment_ctx['duration']:
+                        try:
+                            next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
+                        except IndexError:
+                            next_fragment_time = duration
+                        fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
+                    for _ in range(fragment_repeat):
+                        fragments.append({
+                            'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
+                            'duration': fragment_ctx['duration'] / stream_timescale,
+                        })
+                        fragment_ctx['time'] += fragment_ctx['duration']
+
+                format_id = []
+                if ism_id:
+                    format_id.append(ism_id)
+                if stream_name:
+                    format_id.append(stream_name)
+                format_id.append(compat_str(tbr))
+
+                formats.append({
+                    'format_id': '-'.join(format_id),
+                    'url': ism_url,
+                    'manifest_url': ism_url,
+                    'ext': 'ismv' if stream_type == 'video' else 'isma',
+                    'width': width,
+                    'height': height,
+                    'tbr': tbr,
+                    'asr': sampling_rate,
+                    'vcodec': 'none' if stream_type == 'audio' else fourcc,
+                    'acodec': 'none' if stream_type == 'video' else fourcc,
+                    'protocol': 'ism',
+                    'fragments': fragments,
+                    '_download_params': {
+                        'duration': duration,
+                        'timescale': stream_timescale,
+                        'width': width or 0,
+                        'height': height or 0,
+                        'fourcc': fourcc,
+                        'codec_private_data': track.get('CodecPrivateData'),
+                        'sampling_rate': sampling_rate,
+                        'channels': int_or_none(track.get('Channels', 2)),
+                        'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
+                        'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
+                    },
+                })
+        return formats
+
+    def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None, preference=None):
+        def absolute_url(item_url):
+            return urljoin(base_url, item_url)
+
+        def parse_content_type(content_type):
+            if not content_type:
+                return {}
+            ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
+            if ctr:
+                mimetype, codecs = ctr.groups()
+                f = parse_codecs(codecs)
+                f['ext'] = mimetype2ext(mimetype)
+                return f
+            return {}
+
+        def _media_formats(src, cur_media_type, type_info={}):
+            full_url = absolute_url(src)
+            ext = type_info.get('ext') or determine_ext(full_url)
+            if ext == 'm3u8':
+                is_plain_url = False
+                formats = self._extract_m3u8_formats(
+                    full_url, video_id, ext='mp4',
+                    entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
+                    preference=preference, fatal=False)
+            elif ext == 'mpd':
+                is_plain_url = False
+                formats = self._extract_mpd_formats(
+                    full_url, video_id, mpd_id=mpd_id, fatal=False)
+            else:
+                is_plain_url = True
+                formats = [{
+                    'url': full_url,
+                    'vcodec': 'none' if cur_media_type == 'audio' else None,
+                }]
+            return is_plain_url, formats
+
+        entries = []
+        # amp-video and amp-audio are very similar to their HTML5 counterparts
+        # so we wll include them right here (see
+        # https://www.ampproject.org/docs/reference/components/amp-video)
+        media_tags = [(media_tag, media_type, '')
+                      for media_tag, media_type
+                      in re.findall(r'(?s)(<(?:amp-)?(video|audio)[^>]*/>)', webpage)]
+        media_tags.extend(re.findall(
+            # We only allow video|audio followed by a whitespace or '>'.
+            # Allowing more characters may end up in significant slow down (see
+            # https://github.com/ytdl-org/youtube-dl/issues/11979, example URL:
+            # http://www.porntrex.com/maps/videositemap.xml).
+            r'(?s)(<(?P<tag>(?:amp-)?(?:video|audio))(?:\s+[^>]*)?>)(.*?)</(?P=tag)>', webpage))
+        for media_tag, media_type, media_content in media_tags:
+            media_info = {
+                'formats': [],
+                'subtitles': {},
+            }
+            media_attributes = extract_attributes(media_tag)
+            src = strip_or_none(media_attributes.get('src'))
+            if src:
+                _, formats = _media_formats(src, media_type)
+                media_info['formats'].extend(formats)
+            media_info['thumbnail'] = absolute_url(media_attributes.get('poster'))
+            if media_content:
+                for source_tag in re.findall(r'<source[^>]+>', media_content):
+                    s_attr = extract_attributes(source_tag)
+                    # data-video-src and data-src are non standard but seen
+                    # several times in the wild
+                    src = strip_or_none(dict_get(s_attr, ('src', 'data-video-src', 'data-src')))
+                    if not src:
+                        continue
+                    f = parse_content_type(s_attr.get('type'))
+                    is_plain_url, formats = _media_formats(src, media_type, f)
+                    if is_plain_url:
+                        # width, height, res, label and title attributes are
+                        # all not standard but seen several times in the wild
+                        labels = [
+                            s_attr.get(lbl)
+                            for lbl in ('label', 'title')
+                            if str_or_none(s_attr.get(lbl))
+                        ]
+                        width = int_or_none(s_attr.get('width'))
+                        height = (int_or_none(s_attr.get('height'))
+                                  or int_or_none(s_attr.get('res')))
+                        if not width or not height:
+                            for lbl in labels:
+                                resolution = parse_resolution(lbl)
+                                if not resolution:
+                                    continue
+                                width = width or resolution.get('width')
+                                height = height or resolution.get('height')
+                        for lbl in labels:
+                            tbr = parse_bitrate(lbl)
+                            if tbr:
+                                break
+                        else:
+                            tbr = None
+                        f.update({
+                            'width': width,
+                            'height': height,
+                            'tbr': tbr,
+                            'format_id': s_attr.get('label') or s_attr.get('title'),
+                        })
+                        f.update(formats[0])
+                        media_info['formats'].append(f)
+                    else:
+                        media_info['formats'].extend(formats)
+                for track_tag in re.findall(r'<track[^>]+>', media_content):
+                    track_attributes = extract_attributes(track_tag)
+                    kind = track_attributes.get('kind')
+                    if not kind or kind in ('subtitles', 'captions'):
+                        src = strip_or_none(track_attributes.get('src'))
+                        if not src:
+                            continue
+                        lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
+                        media_info['subtitles'].setdefault(lang, []).append({
+                            'url': absolute_url(src),
+                        })
+            for f in media_info['formats']:
+                f.setdefault('http_headers', {})['Referer'] = base_url
+            if media_info['formats'] or media_info['subtitles']:
+                entries.append(media_info)
+        return entries
+
+    def _extract_akamai_formats(self, manifest_url, video_id, hosts={}):
+        formats = []
+        hdcore_sign = 'hdcore=3.7.0'
+        f4m_url = re.sub(r'(https?://[^/]+)/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
+        hds_host = hosts.get('hds')
+        if hds_host:
+            f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
+        if 'hdcore=' not in f4m_url:
+            f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
+        f4m_formats = self._extract_f4m_formats(
+            f4m_url, video_id, f4m_id='hds', fatal=False)
+        for entry in f4m_formats:
+            entry.update({'extra_param_to_segment_url': hdcore_sign})
+        formats.extend(f4m_formats)
+        m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
+        hls_host = hosts.get('hls')
+        if hls_host:
+            m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
+        formats.extend(self._extract_m3u8_formats(
+            m3u8_url, video_id, 'mp4', 'm3u8_native',
+            m3u8_id='hls', fatal=False))
+        return formats
+
+    def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
+        query = compat_urlparse.urlparse(url).query
+        url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
+        mobj = re.search(
+            r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url)
+        url_base = mobj.group('url')
+        http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base)
+        formats = []
+
+        def manifest_url(manifest):
+            m_url = '%s/%s' % (http_base_url, manifest)
+            if query:
+                m_url += '?%s' % query
+            return m_url
+
+        if 'm3u8' not in skip_protocols:
+            formats.extend(self._extract_m3u8_formats(
+                manifest_url('playlist.m3u8'), video_id, 'mp4',
+                m3u8_entry_protocol, m3u8_id='hls', fatal=False))
+        if 'f4m' not in skip_protocols:
+            formats.extend(self._extract_f4m_formats(
+                manifest_url('manifest.f4m'),
+                video_id, f4m_id='hds', fatal=False))
+        if 'dash' not in skip_protocols:
+            formats.extend(self._extract_mpd_formats(
+                manifest_url('manifest.mpd'),
+                video_id, mpd_id='dash', fatal=False))
+        if re.search(r'(?:/smil:|\.smil)', url_base):
+            if 'smil' not in skip_protocols:
+                rtmp_formats = self._extract_smil_formats(
+                    manifest_url('jwplayer.smil'),
+                    video_id, fatal=False)
+                for rtmp_format in rtmp_formats:
+                    rtsp_format = rtmp_format.copy()
+                    rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
+                    del rtsp_format['play_path']
+                    del rtsp_format['ext']
+                    rtsp_format.update({
+                        'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
+                        'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
+                        'protocol': 'rtsp',
+                    })
+                    formats.extend([rtmp_format, rtsp_format])
+        else:
+            for protocol in ('rtmp', 'rtsp'):
+                if protocol not in skip_protocols:
+                    formats.append({
+                        'url': '%s:%s' % (protocol, url_base),
+                        'format_id': protocol,
+                        'protocol': protocol,
+                    })
+        return formats
+
+    def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
+        mobj = re.search(
+            r'(?s)jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)(?!</script>).*?\.setup\s*\((?P<options>[^)]+)\)',
+            webpage)
+        if mobj:
+            try:
+                jwplayer_data = self._parse_json(mobj.group('options'),
+                                                 video_id=video_id,
+                                                 transform_source=transform_source)
+            except ExtractorError:
+                pass
+            else:
+                if isinstance(jwplayer_data, dict):
+                    return jwplayer_data
+
+    def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
+        jwplayer_data = self._find_jwplayer_data(
+            webpage, video_id, transform_source=js_to_json)
+        return self._parse_jwplayer_data(
+            jwplayer_data, video_id, *args, **kwargs)
+
+    def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
+                             m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
+        # JWPlayer backward compatibility: flattened playlists
+        # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
+        if 'playlist' not in jwplayer_data:
+            jwplayer_data = {'playlist': [jwplayer_data]}
+
+        entries = []
+
+        # JWPlayer backward compatibility: single playlist item
+        # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
+        if not isinstance(jwplayer_data['playlist'], list):
+            jwplayer_data['playlist'] = [jwplayer_data['playlist']]
+
+        for video_data in jwplayer_data['playlist']:
+            # JWPlayer backward compatibility: flattened sources
+            # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
+            if 'sources' not in video_data:
+                video_data['sources'] = [video_data]
+
+            this_video_id = video_id or video_data['mediaid']
+
+            formats = self._parse_jwplayer_formats(
+                video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
+                mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
+
+            subtitles = {}
+            tracks = video_data.get('tracks')
+            if tracks and isinstance(tracks, list):
+                for track in tracks:
+                    if not isinstance(track, dict):
+                        continue
+                    track_kind = track.get('kind')
+                    if not track_kind or not isinstance(track_kind, compat_str):
+                        continue
+                    if track_kind.lower() not in ('captions', 'subtitles'):
+                        continue
+                    track_url = urljoin(base_url, track.get('file'))
+                    if not track_url:
+                        continue
+                    subtitles.setdefault(track.get('label') or 'en', []).append({
+                        'url': self._proto_relative_url(track_url)
+                    })
+
+            entry = {
+                'id': this_video_id,
+                'title': unescapeHTML(video_data['title'] if require_title else video_data.get('title')),
+                'description': clean_html(video_data.get('description')),
+                'thumbnail': urljoin(base_url, self._proto_relative_url(video_data.get('image'))),
+                'timestamp': int_or_none(video_data.get('pubdate')),
+                'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
+                'subtitles': subtitles,
+            }
+            # https://github.com/jwplayer/jwplayer/blob/master/src/js/utils/validator.js#L32
+            if len(formats) == 1 and re.search(r'^(?:http|//).*(?:youtube\.com|youtu\.be)/.+', formats[0]['url']):
+                entry.update({
+                    '_type': 'url_transparent',
+                    'url': formats[0]['url'],
+                })
+            else:
+                self._sort_formats(formats)
+                entry['formats'] = formats
+            entries.append(entry)
+        if len(entries) == 1:
+            return entries[0]
+        else:
+            return self.playlist_result(entries)
+
+    def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
+                                m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
+        urls = []
+        formats = []
+        for source in jwplayer_sources_data:
+            if not isinstance(source, dict):
+                continue
+            source_url = urljoin(
+                base_url, self._proto_relative_url(source.get('file')))
+            if not source_url or source_url in urls:
+                continue
+            urls.append(source_url)
+            source_type = source.get('type') or ''
+            ext = mimetype2ext(source_type) or determine_ext(source_url)
+            if source_type == 'hls' or ext == 'm3u8':
+                formats.extend(self._extract_m3u8_formats(
+                    source_url, video_id, 'mp4', entry_protocol='m3u8_native',
+                    m3u8_id=m3u8_id, fatal=False))
+            elif source_type == 'dash' or ext == 'mpd':
+                formats.extend(self._extract_mpd_formats(
+                    source_url, video_id, mpd_id=mpd_id, fatal=False))
+            elif ext == 'smil':
+                formats.extend(self._extract_smil_formats(
+                    source_url, video_id, fatal=False))
+            # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
+            elif source_type.startswith('audio') or ext in (
+                    'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
+                formats.append({
+                    'url': source_url,
+                    'vcodec': 'none',
+                    'ext': ext,
+                })
+            else:
+                height = int_or_none(source.get('height'))
+                if height is None:
+                    # Often no height is provided but there is a label in
+                    # format like "1080p", "720p SD", or 1080.
+                    height = int_or_none(self._search_regex(
+                        r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
+                        'height', default=None))
+                a_format = {
+                    'url': source_url,
+                    'width': int_or_none(source.get('width')),
+                    'height': height,
+                    'tbr': int_or_none(source.get('bitrate')),
+                    'ext': ext,
+                }
+                if source_url.startswith('rtmp'):
+                    a_format['ext'] = 'flv'
+                    # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
+                    # of jwplayer.flash.swf
+                    rtmp_url_parts = re.split(
+                        r'((?:mp4|mp3|flv):)', source_url, 1)
+                    if len(rtmp_url_parts) == 3:
+                        rtmp_url, prefix, play_path = rtmp_url_parts
+                        a_format.update({
+                            'url': rtmp_url,
+                            'play_path': prefix + play_path,
+                        })
+                    if rtmp_params:
+                        a_format.update(rtmp_params)
+                formats.append(a_format)
+        return formats
+
+    def _live_title(self, name):
+        """ Generate the title for a live video """
+        now = datetime.datetime.now()
+        now_str = now.strftime('%Y-%m-%d %H:%M')
+        return name + ' ' + now_str
+
+    def _int(self, v, name, fatal=False, **kwargs):
+        res = int_or_none(v, **kwargs)
+        if 'get_attr' in kwargs:
+            print(getattr(v, kwargs['get_attr']))
+        if res is None:
+            msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
+            if fatal:
+                raise ExtractorError(msg)
+            else:
+                self._downloader.report_warning(msg)
+        return res
+
+    def _float(self, v, name, fatal=False, **kwargs):
+        res = float_or_none(v, **kwargs)
+        if res is None:
+            msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
+            if fatal:
+                raise ExtractorError(msg)
+            else:
+                self._downloader.report_warning(msg)
+        return res
+
+    def _set_cookie(self, domain, name, value, expire_time=None, port=None,
+                    path='/', secure=False, discard=False, rest={}, **kwargs):
+        cookie = compat_cookiejar.Cookie(
+            0, name, value, port, port is not None, domain, True,
+            domain.startswith('.'), path, True, secure, expire_time,
+            discard, None, None, rest)
+        self._downloader.cookiejar.set_cookie(cookie)
+
+    def _get_cookies(self, url):
+        """ Return a compat_cookies.SimpleCookie with the cookies for the url """
+        req = sanitized_Request(url)
+        self._downloader.cookiejar.add_cookie_header(req)
+        return compat_cookies.SimpleCookie(req.get_header('Cookie'))
+
+    def _apply_first_set_cookie_header(self, url_handle, cookie):
+        """
+        Apply first Set-Cookie header instead of the last. Experimental.
+
+        Some sites (e.g. [1-3]) may serve two cookies under the same name
+        in Set-Cookie header and expect the first (old) one to be set rather
+        than second (new). However, as of RFC6265 the newer one cookie
+        should be set into cookie store what actually happens.
+        We will workaround this issue by resetting the cookie to
+        the first one manually.
+        1. https://new.vk.com/
+        2. https://github.com/ytdl-org/youtube-dl/issues/9841#issuecomment-227871201
+        3. https://learning.oreilly.com/
+        """
+        for header, cookies in url_handle.headers.items():
+            if header.lower() != 'set-cookie':
+                continue
+            if sys.version_info[0] >= 3:
+                cookies = cookies.encode('iso-8859-1')
+            cookies = cookies.decode('utf-8')
+            cookie_value = re.search(
+                r'%s=(.+?);.*?\b[Dd]omain=(.+?)(?:[,;]|$)' % cookie, cookies)
+            if cookie_value:
+                value, domain = cookie_value.groups()
+                self._set_cookie(domain, cookie, value)
+                break
+
+    def get_testcases(self, include_onlymatching=False):
+        t = getattr(self, '_TEST', None)
+        if t:
+            assert not hasattr(self, '_TESTS'), \
+                '%s has _TEST and _TESTS' % type(self).__name__
+            tests = [t]
+        else:
+            tests = getattr(self, '_TESTS', [])
+        for t in tests:
+            if not include_onlymatching and t.get('only_matching', False):
+                continue
+            t['name'] = type(self).__name__[:-len('IE')]
+            yield t
+
+    def is_suitable(self, age_limit):
+        """ Test whether the extractor is generally suitable for the given
+        age limit (i.e. pornographic sites are not, all others usually are) """
+
+        any_restricted = False
+        for tc in self.get_testcases(include_onlymatching=False):
+            if tc.get('playlist', []):
+                tc = tc['playlist'][0]
+            is_restricted = age_restricted(
+                tc.get('info_dict', {}).get('age_limit'), age_limit)
+            if not is_restricted:
+                return True
+            any_restricted = any_restricted or is_restricted
+        return not any_restricted
+
+    def extract_subtitles(self, *args, **kwargs):
+        if (self._downloader.params.get('writesubtitles', False)
+                or self._downloader.params.get('listsubtitles')):
+            return self._get_subtitles(*args, **kwargs)
+        return {}
+
+    def _get_subtitles(self, *args, **kwargs):
+        raise NotImplementedError('This method must be implemented by subclasses')
+
+    @staticmethod
+    def _merge_subtitle_items(subtitle_list1, subtitle_list2):
+        """ Merge subtitle items for one language. Items with duplicated URLs
+        will be dropped. """
+        list1_urls = set([item['url'] for item in subtitle_list1])
+        ret = list(subtitle_list1)
+        ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
+        return ret
+
+    @classmethod
+    def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
+        """ Merge two subtitle dictionaries, language by language. """
+        ret = dict(subtitle_dict1)
+        for lang in subtitle_dict2:
+            ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
+        return ret
+
+    def extract_automatic_captions(self, *args, **kwargs):
+        if (self._downloader.params.get('writeautomaticsub', False)
+                or self._downloader.params.get('listsubtitles')):
+            return self._get_automatic_captions(*args, **kwargs)
+        return {}
+
+    def _get_automatic_captions(self, *args, **kwargs):
+        raise NotImplementedError('This method must be implemented by subclasses')
+
+    def mark_watched(self, *args, **kwargs):
+        if (self._downloader.params.get('mark_watched', False)
+                and (self._get_login_info()[0] is not None
+                     or self._downloader.params.get('cookiefile') is not None)):
+            self._mark_watched(*args, **kwargs)
+
+    def _mark_watched(self, *args, **kwargs):
+        raise NotImplementedError('This method must be implemented by subclasses')
+
+    def geo_verification_headers(self):
+        headers = {}
+        geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
+        if geo_verification_proxy:
+            headers['Ytdl-request-proxy'] = geo_verification_proxy
+        return headers
+
+    def _generic_id(self, url):
+        return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
+
+    def _generic_title(self, url):
+        return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
+
 
 class SearchInfoExtractor(InfoExtractor):
     """
     Base class for paged search queries extractors.
-    They accept urls in the format _SEARCH_KEY(|all|[0-9]):{query}
+    They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
     Instances should define _SEARCH_KEY and _MAX_RESULTS.
     """
 
@@ -594,7 +2962,7 @@ class SearchInfoExtractor(InfoExtractor):
     def _real_extract(self, query):
         mobj = re.match(self._make_valid_url(), query)
         if mobj is None:
-            raise ExtractorError(u'Invalid search query "%s"' % query)
+            raise ExtractorError('Invalid search query "%s"' % query)
 
         prefix = mobj.group('prefix')
         query = mobj.group('query')
@@ -605,17 +2973,16 @@ class SearchInfoExtractor(InfoExtractor):
         else:
             n = int(prefix)
             if n <= 0:
-                raise ExtractorError(u'invalid download number %s for query "%s"' % (n, query))
+                raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
             elif n > self._MAX_RESULTS:
-                self._downloader.report_warning(u'%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
+                self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
                 n = self._MAX_RESULTS
             return self._get_n_results(query, n)
 
     def _get_n_results(self, query, n):
         """Get a specified number of results for a query"""
-        raise NotImplementedError("This method must be implemented by subclasses")
+        raise NotImplementedError('This method must be implemented by subclasses')
 
     @property
     def SEARCH_KEY(self):
         return self._SEARCH_KEY
-