]> Raphaƫl G. Git Repositories - youtubedl/blobdiff - youtube_dl/extractor/common.py
debian/changelog: Start new release.
[youtubedl] / youtube_dl / extractor / common.py
index ba46a7bc77d17ed4bcf4dcf7764b1d39f4799958..e4e4feef9ea18787d196b2ca7d3414191409ba97 100644 (file)
@@ -1,4 +1,6 @@
 import base64
 import base64
+import hashlib
+import json
 import os
 import re
 import socket
 import os
 import re
 import socket
@@ -9,6 +11,7 @@ import xml.etree.ElementTree
 from ..utils import (
     compat_http_client,
     compat_urllib_error,
 from ..utils import (
     compat_http_client,
     compat_urllib_error,
+    compat_urllib_parse_urlparse,
     compat_str,
 
     clean_html,
     compat_str,
 
     clean_html,
@@ -37,10 +40,12 @@ class InfoExtractor(object):
     id:             Video identifier.
     title:          Video title, unescaped.
 
     id:             Video identifier.
     title:          Video title, unescaped.
 
-    Additionally, it must contain either a formats entry or url and ext:
+    Additionally, it must contain either a formats entry or a url one:
 
 
-    formats:        A list of dictionaries for each format available, it must
-                    be ordered from worst to best quality. Potential fields:
+    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
                     * ext        Will be calculated from url if missing
                     * format     A human-readable description of the format
                     * url        Mandatory. The URL of the video file
                     * ext        Will be calculated from url if missing
                     * format     A human-readable description of the format
@@ -48,32 +53,57 @@ class InfoExtractor(object):
                                  Calculated from the format_id, width, height.
                                  and format_note fields if missing.
                     * format_id  A short description of the format
                                  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")
+                                 ("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
                     * 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
                     * 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
                     * vcodec     Name of the video codec in use
                     * vbr        Average video bitrate in KBit/s
                     * vcodec     Name of the video codec in use
+                    * container  Name of the container format
                     * filesize   The number of bytes, if known in advance
                     * player_url SWF Player URL (used for rtmpdump).
                     * filesize   The number of bytes, if known in advance
                     * 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.
+                    * 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.
+                    * 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.
     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).
     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).
-    urlhandle:      [internal] The urlHandle to be used to download the file,
-                    like returned by urllib.request.urlopen
 
     The following fields are optional:
 
 
     The following fields are optional:
 
-    thumbnails:     A list of dictionaries (with the entries "resolution" and
-                    "url") for the varying thumbnails
+    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:
+                        * "url"
+                        * "width" (optional, int)
+                        * "height" (optional, int)
+                        * "resolution" (optional, string "{width}x{height"},
+                                        deprecated)
     thumbnail:      Full URL to a video thumbnail image.
     description:    One-line video description.
     uploader:       Full name of the video uploader.
     thumbnail:      Full URL to a video thumbnail image.
     description:    One-line video description.
     uploader:       Full name of the video uploader.
+    timestamp:      UNIX timestamp of the moment the video became available.
     upload_date:    Video upload date (YYYYMMDD).
     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
     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
@@ -87,6 +117,8 @@ class InfoExtractor(object):
     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)
     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"]
 
     Unless mentioned otherwise, the fields should be Unicode strings.
 
 
     Unless mentioned otherwise, the fields should be Unicode strings.
 
@@ -94,9 +126,6 @@ class InfoExtractor(object):
     _real_extract() methods and define a _VALID_URL regexp.
     Probably, they should also be added to the list of extractors.
 
     _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.
-
     Finally, the _WORKING attribute should be set to False for broken IEs
     in order to warn the users and skip the tests.
     """
     Finally, the _WORKING attribute should be set to False for broken IEs
     in order to warn the users and skip the tests.
     """
@@ -202,6 +231,8 @@ class InfoExtractor(object):
                           webpage_bytes[:1024])
             if m:
                 encoding = m.group(1).decode('ascii')
                           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'
         if self._downloader.params.get('dump_intermediate_pages', False):
             else:
                 encoding = 'utf-8'
         if self._downloader.params.get('dump_intermediate_pages', False):
@@ -217,13 +248,31 @@ class InfoExtractor(object):
                 url = url_or_request.get_full_url()
             except AttributeError:
                 url = url_or_request
                 url = url_or_request.get_full_url()
             except AttributeError:
                 url = url_or_request
-            raw_filename = ('%s_%s.dump' % (video_id, url))
+            basen = '%s_%s' % (video_id, url)
+            if len(basen) > 240:
+                h = u'___' + 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)
             with open(filename, 'wb') as outf:
                 outf.write(webpage_bytes)
 
             filename = sanitize_filename(raw_filename, restricted=True)
             self.to_screen(u'Saving request to ' + filename)
             with open(filename, 'wb') as outf:
                 outf.write(webpage_bytes)
 
-        content = webpage_bytes.decode(encoding, 'replace')
+        try:
+            content = webpage_bytes.decode(encoding, 'replace')
+        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)
+
         return (content, urlh)
 
     def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
         return (content, urlh)
 
     def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True):
@@ -237,13 +286,33 @@ class InfoExtractor(object):
 
     def _download_xml(self, url_or_request, video_id,
                       note=u'Downloading XML', errnote=u'Unable to download XML',
 
     def _download_xml(self, url_or_request, video_id,
                       note=u'Downloading XML', errnote=u'Unable to download XML',
-                      transform_source=None):
+                      transform_source=None, fatal=True):
         """Return the xml as an xml.etree.ElementTree.Element"""
         """Return the xml as an xml.etree.ElementTree.Element"""
-        xml_string = self._download_webpage(url_or_request, video_id, note, errnote)
+        xml_string = self._download_webpage(
+            url_or_request, video_id, note, errnote, fatal=fatal)
+        if xml_string is False:
+            return xml_string
         if transform_source:
             xml_string = transform_source(xml_string)
         return xml.etree.ElementTree.fromstring(xml_string.encode('utf-8'))
 
         if transform_source:
             xml_string = transform_source(xml_string)
         return xml.etree.ElementTree.fromstring(xml_string.encode('utf-8'))
 
+    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)
+        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)
+
+    def report_warning(self, msg, video_id=None):
+        idstr = u'' if video_id is None else u'%s: ' % video_id
+        self._downloader.report_warning(
+            u'[%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))
     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))
@@ -360,8 +429,8 @@ class InfoExtractor(object):
     # Helper functions for extracting OpenGraph info
     @staticmethod
     def _og_regexes(prop):
     # Helper functions for extracting OpenGraph info
     @staticmethod
     def _og_regexes(prop):
-        content_re = r'content=(?:"([^>]+?)"|\'(.+?)\')'
-        property_re = r'property=[\'"]og:%s[\'"]' % re.escape(prop)
+        content_re = r'content=(?:"([^>]+?)"|\'([^>]+?)\')'
+        property_re = r'(?:name|property)=[\'"]og:%s[\'"]' % re.escape(prop)
         template = r'<meta[^>]+?%s[^>]+?%s'
         return [
             template % (property_re, content_re),
         template = r'<meta[^>]+?%s[^>]+?%s'
         return [
             template % (property_re, content_re),
@@ -390,14 +459,17 @@ class InfoExtractor(object):
         if secure: regexes = self._og_regexes('video:secure_url') + regexes
         return self._html_search_regex(regexes, html, name, **kargs)
 
         if secure: regexes = self._og_regexes('video:secure_url') + regexes
         return self._html_search_regex(regexes, html, name, **kargs)
 
-    def _html_search_meta(self, name, html, display_name=None):
+    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):
         if display_name is None:
             display_name = name
         return self._html_search_regex(
             r'''(?ix)<meta
                     (?=[^>]+(?:itemprop|name|property)=["\']%s["\'])
                     [^>]+content=["\']([^"\']+)["\']''' % re.escape(name),
         if display_name is None:
             display_name = name
         return self._html_search_regex(
             r'''(?ix)<meta
                     (?=[^>]+(?:itemprop|name|property)=["\']%s["\'])
                     [^>]+content=["\']([^"\']+)["\']''' % re.escape(name),
-            html, display_name, fatal=False)
+            html, display_name, fatal=fatal)
 
     def _dc_search_uploader(self, html):
         return self._html_search_meta('dc.creator', html, 'uploader')
 
     def _dc_search_uploader(self, html):
         return self._html_search_meta('dc.creator', html, 'uploader')
@@ -426,6 +498,82 @@ class InfoExtractor(object):
         }
         return RATING_TABLE.get(rating.lower(), None)
 
         }
         return RATING_TABLE.get(rating.lower(), None)
 
+    def _twitter_search_player(self, html):
+        return self._html_search_meta('twitter:player', html,
+            'twitter card player')
+
+    def _sort_formats(self, formats):
+        if not formats:
+            raise ExtractorError(u'No video formats found')
+
+        def _formats_key(f):
+            # TODO remove the following workaround
+            from ..utils import determine_ext
+            if not f.get('ext') and 'url' in f:
+                f['ext'] = determine_ext(f['url'])
+
+            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
+                if f.get('ext') in ['f4f', 'f4m']:  # Not yet supported
+                    preference -= 0.5
+
+            if f.get('vcodec') == 'none':  # audio only
+                if self._downloader.params.get('prefer_free_formats'):
+                    ORDER = [u'aac', u'mp3', u'm4a', u'webm', u'ogg', u'opus']
+                else:
+                    ORDER = [u'webm', u'opus', u'ogg', u'mp3', u'aac', u'm4a']
+                ext_preference = 0
+                try:
+                    audio_ext_preference = ORDER.index(f['ext'])
+                except ValueError:
+                    audio_ext_preference = -1
+            else:
+                if self._downloader.params.get('prefer_free_formats'):
+                    ORDER = [u'flv', u'mp4', u'webm']
+                else:
+                    ORDER = [u'webm', u'flv', u'mp4']
+                try:
+                    ext_preference = ORDER.index(f['ext'])
+                except ValueError:
+                    ext_preference = -1
+                audio_ext_preference = 0
+
+            return (
+                preference,
+                f.get('quality') if f.get('quality') 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,
+                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'),
+            )
+        formats.sort(key=_formats_key)
+
+    def http_scheme(self):
+        """ Either "https:" or "https:", depending on the user's preferences """
+        return (
+            'http:'
+            if self._downloader.params.get('prefer_insecure', False)
+            else 'https:')
+
+    def _proto_relative_url(self, url, scheme=None):
+        if url is None:
+            return url
+        if url.startswith('//'):
+            if scheme is None:
+                scheme = self.http_scheme()
+            return scheme + url
+        else:
+            return url
 
 
 class SearchInfoExtractor(InfoExtractor):
 
 
 class SearchInfoExtractor(InfoExtractor):
@@ -470,3 +618,4 @@ class SearchInfoExtractor(InfoExtractor):
     @property
     def SEARCH_KEY(self):
         return self._SEARCH_KEY
     @property
     def SEARCH_KEY(self):
         return self._SEARCH_KEY
+