X-Git-Url: https://git.rapsys.eu/youtubedl/blobdiff_plain/467f1716f3b46ffc038848b9b08a833a7811be09..0fc91537a0580721211b68e8cc4367a33cd6d4f2:/youtube_dl/extractor/common.py?ds=sidebyside diff --git a/youtube_dl/extractor/common.py b/youtube_dl/extractor/common.py index da50abf..a2548db 100644 --- a/youtube_dl/extractor/common.py +++ b/youtube_dl/extractor/common.py @@ -1,22 +1,78 @@ +# 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 time +import math -from ..utils import ( +from ..compat import ( + compat_cookiejar, + compat_cookies, + compat_etree_fromstring, + compat_getpass, compat_http_client, + compat_os_name, + compat_str, compat_urllib_error, + compat_urllib_parse_unquote, + compat_urllib_parse_urlencode, compat_urllib_request, - compat_str, - + 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, + error_to_compat_str, ExtractorError, + extract_attributes, + fix_xml_ampersands, + float_or_none, + GeoRestrictedError, + GeoUtils, + int_or_none, + js_to_json, + mimetype2ext, + orderedSet, + parse_codecs, + parse_duration, + parse_iso8601, + parse_m3u8_attributes, + RegexNotFoundError, + sanitized_Request, + sanitize_filename, unescapeHTML, + unified_strdate, + unified_timestamp, + update_Request, + update_url_query, + urljoin, + url_basename, + xpath_element, + xpath_text, + xpath_with_ns, ) + class InfoExtractor(object): """Information Extractor class. @@ -24,42 +80,276 @@ 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. - url: Final video URL. title: Video title, unescaped. + + Additionally, it must contain either a formats entry or a url one: + + formats: A list of dictionaries for each format available, ordered + from worst to best quality. + + Potential fields: + * url Mandatory. The URL of the video file + * manifest_url + The URL of the manifest file in case of + fragmented media (DASH, hls, hds) + * 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. + and format_note fields if missing. + * format_id A short description of the format + ("mp4_h264_opus" or "19"). + Technically optional, but strongly recommended. + * format_note Additional info about the format + ("3D" or "DASH video") + * width Width of the video, if known + * height Height of the video, if known + * resolution Textual description of width and height + * tbr Average bitrate of audio and video in KBit/s + * abr Average audio bitrate in KBit/s + * 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", "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) + player_url: SWF Player URL (used for rtmpdump). The following fields are optional: - format: The video format, defaults to ext (used for --get-format) - thumbnails: A list of dictionaries (with the entries "resolution" and - "url") for the varying thumbnails + 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"}, + 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. - player_url: SWF Player URL (used for rtmpdump). - subtitles: The subtitle file contents. + uploader_url: Full URL to a personal webpage of the video uploader. + 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. - urlhandle: [internal] The urlHandle to be used to download the file, - like returned by urllib.request.urlopen + 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 + 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. - The fields should all be Unicode strings. 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. - _real_extract() must return a *list* of information dictionaries as - described above. + _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. @@ -67,17 +357,36 @@ class InfoExtractor(object): _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 def suitable(cls, url): """Receives a URL and returns True if suitable for this IE.""" - return re.match(cls._VALID_URL, url) is not None + + # This does not use has/getattr intentionally - we want to know whether + # we have cached the regexp for *this* class, whereas getattr would also + # match the superclass + if '_VALID_URL_RE' not in cls.__dict__: + 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): @@ -86,14 +395,137 @@ 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.""" @@ -107,82 +539,304 @@ class InfoExtractor(object): """Real extraction process. Redefine in subclasses.""" pass + @classmethod + def ie_key(cls): + """A string for getting the InfoExtractor with get_info_extractor""" + return compat_str(cls.__name__[:-2]) + @property def IE_NAME(self): - return type(self).__name__[:-2] + return compat_str(type(self).__name__[:-2]) - def _request_webpage(self, url_or_request, video_id, note=None, errnote=None): + def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers={}, query={}): """ Returns the response handle """ if note is None: self.report_download_webpage(video_id) elif note is not False: - self.to_screen(u'%s: %s' % (video_id, note)) + if video_id is None: + self.to_screen('%s' % (note,)) + else: + 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 compat_urllib_request.urlopen(url_or_request) + return self._downloader.urlopen(url_or_request) except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err: + if errnote is False: + return False if errnote is None: - errnote = u'Unable to download webpage' - raise ExtractorError(u'%s: %s' % (errnote, compat_str(err)), sys.exc_info()[2]) + errnote = 'Unable to download webpage' - def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None): - """ Returns a tuple (page content as string, URL handle) """ + 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, encoding=None, data=None, headers={}, query={}): + """ Returns a tuple (page content as string, URL handle) """ # 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) - content_type = urlh.headers.get('Content-Type', '') + urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query) + if urlh is False: + assert not fatal + return False + 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) else: - encoding = 'utf-8' + m = re.search(br']+charset=[\'"]?([^\'")]+)[ /\'">]', + webpage_bytes[:1024]) + if m: + encoding = m.group(1).decode('ascii') + elif webpage_bytes.startswith(b'\xff\xfe'): + encoding = 'utf-16' + else: + encoding = 'utf-8' + + return encoding + + def __check_blocked(self, content): + first_block = content[:512] + if ('Access to this site is blocked' 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'