]> Raphaƫl G. Git Repositories - youtubedl/blobdiff - youtube_dl/YoutubeDL.py
Prepare to release.
[youtubedl] / youtube_dl / YoutubeDL.py
index 523dd1f7daf80839839d7b8a014f86897640b425..19370f62b0d3ddb91c74ae4b6cf6c569341fbdc5 100755 (executable)
@@ -82,14 +82,17 @@ from .utils import (
     sanitize_url,
     sanitized_Request,
     std_headers,
+    str_or_none,
     subtitles_filename,
     UnavailableVideoError,
     url_basename,
     version_tuple,
     write_json_file,
     write_string,
+    YoutubeDLCookieJar,
     YoutubeDLCookieProcessor,
     YoutubeDLHandler,
+    YoutubeDLRedirectHandler,
 )
 from .cache import Cache
 from .extractor import get_info_extractor, gen_extractor_classes, _LAZY_LOADER
@@ -211,7 +214,7 @@ class YoutubeDL(object):
                        At the moment, this is only supported by YouTube.
     proxy:             URL of the proxy server to use
     geo_verification_proxy:  URL of the proxy to use for IP address verification
-                       on geo-restricted sites. (Experimental)
+                       on geo-restricted sites.
     socket_timeout:    Time to wait for unresponsive hosts, in seconds
     bidi_workaround:   Work around buggy terminals without bidirectional text
                        support, using fridibi
@@ -259,7 +262,7 @@ class YoutubeDL(object):
                        - "warn": only emit a warning
                        - "detect_or_warn": check whether we can do anything
                                            about it, warn otherwise (default)
-    source_address:    (Experimental) Client-side IP address to bind to.
+    source_address:    Client-side IP address to bind to.
     call_home:         Boolean, true iff we are allowed to contact the
                        youtube-dl servers for debugging.
     sleep_interval:    Number of seconds to sleep before each download when
@@ -281,11 +284,14 @@ class YoutubeDL(object):
                        match_filter_func in utils.py is one example for this.
     no_color:          Do not emit color codes in output.
     geo_bypass:        Bypass geographic restriction via faking X-Forwarded-For
-                       HTTP header (experimental)
+                       HTTP header
     geo_bypass_country:
                        Two-letter ISO 3166-2 country code that will be used for
                        explicit geographic restriction bypassing via faking
-                       X-Forwarded-For HTTP header (experimental)
+                       X-Forwarded-For HTTP header
+    geo_bypass_ip_block:
+                       IP range in CIDR notation that will be used similarly to
+                       geo_bypass_country
 
     The following options determine which downloader is picked:
     external_downloader: Executable of the external downloader to call.
@@ -302,8 +308,10 @@ class YoutubeDL(object):
     http_chunk_size.
 
     The following options are used by the post processors:
-    prefer_ffmpeg:     If True, use ffmpeg instead of avconv if both are available,
-                       otherwise prefer avconv.
+    prefer_ffmpeg:     If False, use avconv instead of ffmpeg if both are available,
+                       otherwise prefer ffmpeg.
+    ffmpeg_location:   Location of the ffmpeg/avconv binary; either the path
+                       to the binary or its containing directory.
     postprocessor_args: A list of additional command-line arguments for the
                         postprocessor.
 
@@ -393,9 +401,9 @@ class YoutubeDL(object):
                 else:
                     raise
 
-        if (sys.platform != 'win32' and
-                sys.getfilesystemencoding() in ['ascii', 'ANSI_X3.4-1968'] and
-                not params.get('restrictfilenames', False)):
+        if (sys.platform != 'win32'
+                and sys.getfilesystemencoding() in ['ascii', 'ANSI_X3.4-1968']
+                and not params.get('restrictfilenames', False)):
             # Unicode filesystem API will throw errors (#1474, #13027)
             self.report_warning(
                 'Assuming --restrict-filenames since file system encoding '
@@ -433,9 +441,9 @@ class YoutubeDL(object):
             if re.match(r'^-[0-9A-Za-z_-]{10}$', a)]
         if idxs:
             correct_argv = (
-                ['youtube-dl'] +
-                [a for i, a in enumerate(argv) if i not in idxs] +
-                ['--'] + [argv[i] for i in idxs]
+                ['youtube-dl']
+                + [a for i, a in enumerate(argv) if i not in idxs]
+                ['--'] + [argv[i] for i in idxs]
             )
             self.report_warning(
                 'Long argument string detected. '
@@ -532,6 +540,8 @@ class YoutubeDL(object):
     def save_console_title(self):
         if not self.params.get('consoletitle', False):
             return
+        if self.params.get('simulate', False):
+            return
         if compat_os_name != 'nt' and 'TERM' in os.environ:
             # Save the title on stack
             self._write_string('\033[22;0t', self._screen_file)
@@ -539,6 +549,8 @@ class YoutubeDL(object):
     def restore_console_title(self):
         if not self.params.get('consoletitle', False):
             return
+        if self.params.get('simulate', False):
+            return
         if compat_os_name != 'nt' and 'TERM' in os.environ:
             # Restore the title from stack
             self._write_string('\033[23;0t', self._screen_file)
@@ -551,7 +563,7 @@ class YoutubeDL(object):
         self.restore_console_title()
 
         if self.params.get('cookiefile') is not None:
-            self.cookiejar.save()
+            self.cookiejar.save(ignore_discard=True, ignore_expires=True)
 
     def trouble(self, message=None, tb=None):
         """Determine action to take when a download problem appears.
@@ -839,10 +851,11 @@ class YoutubeDL(object):
         if result_type in ('url', 'url_transparent'):
             ie_result['url'] = sanitize_url(ie_result['url'])
             extract_flat = self.params.get('extract_flat', False)
-            if ((extract_flat == 'in_playlist' and 'playlist' in extra_info) or
-                    extract_flat is True):
-                if self.params.get('forcejson', False):
-                    self.to_stdout(json.dumps(ie_result))
+            if ((extract_flat == 'in_playlist' and 'playlist' in extra_info)
+                    or extract_flat is True):
+                self.__forced_printings(
+                    ie_result, self.prepare_filename(ie_result),
+                    incomplete=True)
                 return ie_result
 
         if result_type == 'video':
@@ -880,7 +893,7 @@ class YoutubeDL(object):
             # url_transparent. In such cases outer metadata (from ie_result)
             # should be propagated to inner one (info). For this to happen
             # _type of info should be overridden with url_transparent. This
-            # fixes issue from https://github.com/rg3/youtube-dl/pull/11163.
+            # fixes issue from https://github.com/ytdl-org/youtube-dl/pull/11163.
             if new_result.get('_type') == 'url':
                 new_result['_type'] = 'url_transparent'
 
@@ -978,7 +991,7 @@ class YoutubeDL(object):
                     'playlist_title': ie_result.get('title'),
                     'playlist_uploader': ie_result.get('uploader'),
                     'playlist_uploader_id': ie_result.get('uploader_id'),
-                    'playlist_index': i + playliststart,
+                    'playlist_index': playlistitems[i - 1] if playlistitems else i + playliststart,
                     'extractor': ie_result['extractor'],
                     'webpage_url': ie_result['webpage_url'],
                     'webpage_url_basename': url_basename(ie_result['webpage_url']),
@@ -1055,21 +1068,24 @@ class YoutubeDL(object):
         if not m:
             STR_OPERATORS = {
                 '=': operator.eq,
-                '!=': operator.ne,
                 '^=': lambda attr, value: attr.startswith(value),
                 '$=': lambda attr, value: attr.endswith(value),
                 '*=': lambda attr, value: value in attr,
             }
             str_operator_rex = re.compile(r'''(?x)
                 \s*(?P<key>ext|acodec|vcodec|container|protocol|format_id)
-                \s*(?P<op>%s)(?P<none_inclusive>\s*\?)?
+                \s*(?P<negation>!\s*)?(?P<op>%s)(?P<none_inclusive>\s*\?)?
                 \s*(?P<value>[a-zA-Z0-9._-]+)
                 \s*$
                 ''' % '|'.join(map(re.escape, STR_OPERATORS.keys())))
             m = str_operator_rex.search(filter_spec)
             if m:
                 comparison_value = m.group('value')
-                op = STR_OPERATORS[m.group('op')]
+                str_op = STR_OPERATORS[m.group('op')]
+                if m.group('negation'):
+                    op = lambda attr, value: not str_op(attr, value)
+                else:
+                    op = str_op
 
         if not m:
             raise ValueError('Invalid filter specification %r' % filter_spec)
@@ -1475,23 +1491,28 @@ class YoutubeDL(object):
             if info_dict.get('%s_number' % field) is not None and not info_dict.get(field):
                 info_dict[field] = '%s %d' % (field.capitalize(), info_dict['%s_number' % field])
 
+        for cc_kind in ('subtitles', 'automatic_captions'):
+            cc = info_dict.get(cc_kind)
+            if cc:
+                for _, subtitle in cc.items():
+                    for subtitle_format in subtitle:
+                        if subtitle_format.get('url'):
+                            subtitle_format['url'] = sanitize_url(subtitle_format['url'])
+                        if subtitle_format.get('ext') is None:
+                            subtitle_format['ext'] = determine_ext(subtitle_format['url']).lower()
+
+        automatic_captions = info_dict.get('automatic_captions')
         subtitles = info_dict.get('subtitles')
-        if subtitles:
-            for _, subtitle in subtitles.items():
-                for subtitle_format in subtitle:
-                    if subtitle_format.get('url'):
-                        subtitle_format['url'] = sanitize_url(subtitle_format['url'])
-                    if subtitle_format.get('ext') is None:
-                        subtitle_format['ext'] = determine_ext(subtitle_format['url']).lower()
 
         if self.params.get('listsubtitles', False):
             if 'automatic_captions' in info_dict:
-                self.list_subtitles(info_dict['id'], info_dict.get('automatic_captions'), 'automatic captions')
+                self.list_subtitles(
+                    info_dict['id'], automatic_captions, 'automatic captions')
             self.list_subtitles(info_dict['id'], subtitles, 'subtitles')
             return
+
         info_dict['requested_subtitles'] = self.process_subtitles(
-            info_dict['id'], subtitles,
-            info_dict.get('automatic_captions'))
+            info_dict['id'], subtitles, automatic_captions)
 
         # We now pick which formats have to be downloaded
         if info_dict.get('formats') is None:
@@ -1589,7 +1610,7 @@ class YoutubeDL(object):
         # by extractor are incomplete or not (i.e. whether extractor provides only
         # video-only or audio-only formats) for proper formats selection for
         # extractors with such incomplete formats (see
-        # https://github.com/rg3/youtube-dl/pull/5556).
+        # https://github.com/ytdl-org/youtube-dl/pull/5556).
         # Since formats may be filtered during format selection and may not match
         # the original formats the results may be incorrect. Thus original formats
         # or pre-calculated metrics should be passed to format selection routines
@@ -1597,12 +1618,12 @@ class YoutubeDL(object):
         # We will pass a context object containing all necessary additional data
         # instead of just formats.
         # This fixes incorrect format selection issue (see
-        # https://github.com/rg3/youtube-dl/issues/10083).
+        # https://github.com/ytdl-org/youtube-dl/issues/10083).
         incomplete_formats = (
             # All formats are video-only or
-            all(f.get('vcodec') != 'none' and f.get('acodec') == 'none' for f in formats) or
+            all(f.get('vcodec') != 'none' and f.get('acodec') == 'none' for f in formats)
             # all formats are audio-only
-            all(f.get('vcodec') == 'none' and f.get('acodec') != 'none' for f in formats))
+            or all(f.get('vcodec') == 'none' and f.get('acodec') != 'none' for f in formats))
 
         ctx = {
             'formats': formats,
@@ -1674,6 +1695,36 @@ class YoutubeDL(object):
             subs[lang] = f
         return subs
 
+    def __forced_printings(self, info_dict, filename, incomplete):
+        def print_mandatory(field):
+            if (self.params.get('force%s' % field, False)
+                    and (not incomplete or info_dict.get(field) is not None)):
+                self.to_stdout(info_dict[field])
+
+        def print_optional(field):
+            if (self.params.get('force%s' % field, False)
+                    and info_dict.get(field) is not None):
+                self.to_stdout(info_dict[field])
+
+        print_mandatory('title')
+        print_mandatory('id')
+        if self.params.get('forceurl', False) and not incomplete:
+            if info_dict.get('requested_formats') is not None:
+                for f in info_dict['requested_formats']:
+                    self.to_stdout(f['url'] + f.get('play_path', ''))
+            else:
+                # For RTMP URLs, also include the playpath
+                self.to_stdout(info_dict['url'] + info_dict.get('play_path', ''))
+        print_optional('thumbnail')
+        print_optional('description')
+        if self.params.get('forcefilename', False) and filename is not None:
+            self.to_stdout(filename)
+        if self.params.get('forceduration', False) and info_dict.get('duration') is not None:
+            self.to_stdout(formatSeconds(info_dict['duration']))
+        print_mandatory('format')
+        if self.params.get('forcejson', False):
+            self.to_stdout(json.dumps(info_dict))
+
     def process_info(self, info_dict):
         """Process a single resolved IE result."""
 
@@ -1684,9 +1735,8 @@ class YoutubeDL(object):
             if self._num_downloads >= int(max_downloads):
                 raise MaxDownloadsReached()
 
+        # TODO: backward compatibility, to be removed
         info_dict['fulltitle'] = info_dict['title']
-        if len(info_dict['title']) > 200:
-            info_dict['title'] = info_dict['title'][:197] + '...'
 
         if 'format' not in info_dict:
             info_dict['format'] = info_dict['ext']
@@ -1701,29 +1751,7 @@ class YoutubeDL(object):
         info_dict['_filename'] = filename = self.prepare_filename(info_dict)
 
         # Forced printings
-        if self.params.get('forcetitle', False):
-            self.to_stdout(info_dict['fulltitle'])
-        if self.params.get('forceid', False):
-            self.to_stdout(info_dict['id'])
-        if self.params.get('forceurl', False):
-            if info_dict.get('requested_formats') is not None:
-                for f in info_dict['requested_formats']:
-                    self.to_stdout(f['url'] + f.get('play_path', ''))
-            else:
-                # For RTMP URLs, also include the playpath
-                self.to_stdout(info_dict['url'] + info_dict.get('play_path', ''))
-        if self.params.get('forcethumbnail', False) and info_dict.get('thumbnail') is not None:
-            self.to_stdout(info_dict['thumbnail'])
-        if self.params.get('forcedescription', False) and info_dict.get('description') is not None:
-            self.to_stdout(info_dict['description'])
-        if self.params.get('forcefilename', False) and filename is not None:
-            self.to_stdout(filename)
-        if self.params.get('forceduration', False) and info_dict.get('duration') is not None:
-            self.to_stdout(formatSeconds(info_dict['duration']))
-        if self.params.get('forceformat', False):
-            self.to_stdout(info_dict['format'])
-        if self.params.get('forcejson', False):
-            self.to_stdout(json.dumps(info_dict))
+        self.__forced_printings(info_dict, filename, incomplete=False)
 
         # Do nothing else if in simulate mode
         if self.params.get('simulate', False):
@@ -1764,6 +1792,8 @@ class YoutubeDL(object):
             annofn = replace_extension(filename, 'annotations.xml', info_dict.get('ext'))
             if self.params.get('nooverwrites', False) and os.path.exists(encodeFilename(annofn)):
                 self.to_screen('[info] Video annotations are already present')
+            elif not info_dict.get('annotations'):
+                self.report_warning('There are no annotations to write.')
             else:
                 try:
                     self.to_screen('[info] Writing video annotations to: ' + annofn)
@@ -1785,7 +1815,7 @@ class YoutubeDL(object):
             ie = self.get_info_extractor(info_dict['extractor_key'])
             for sub_lang, sub_info in subtitles.items():
                 sub_format = sub_info['ext']
-                sub_filename = subtitles_filename(filename, sub_lang, sub_format)
+                sub_filename = subtitles_filename(filename, sub_lang, sub_format, info_dict.get('ext'))
                 if self.params.get('nooverwrites', False) and os.path.exists(encodeFilename(sub_filename)):
                     self.to_screen('[info] Video subtitle %s.%s is already present' % (sub_lang, sub_format))
                 else:
@@ -1793,7 +1823,7 @@ class YoutubeDL(object):
                     if sub_info.get('data') is not None:
                         try:
                             # Use newline='' to prevent conversion of newline characters
-                            # See https://github.com/rg3/youtube-dl/issues/10268
+                            # See https://github.com/ytdl-org/youtube-dl/issues/10268
                             with io.open(encodeFilename(sub_filename), 'w', encoding='utf-8', newline='') as subfile:
                                 subfile.write(sub_info['data'])
                         except (OSError, IOError):
@@ -1849,7 +1879,7 @@ class YoutubeDL(object):
                     def compatible_formats(formats):
                         video, audio = formats
                         # Check extension
-                        video_ext, audio_ext = audio.get('ext'), video.get('ext')
+                        video_ext, audio_ext = video.get('ext'), audio.get('ext')
                         if video_ext and audio_ext:
                             COMPATIBLE_EXTS = (
                                 ('mp3', 'mp4', 'm4a', 'm4p', 'm4b', 'm4r', 'm4v', 'ismv', 'isma'),
@@ -1928,8 +1958,8 @@ class YoutubeDL(object):
                     else:
                         assert fixup_policy in ('ignore', 'never')
 
-                if (info_dict.get('requested_formats') is None and
-                        info_dict.get('container') == 'm4a_dash'):
+                if (info_dict.get('requested_formats') is None
+                        and info_dict.get('container') == 'm4a_dash'):
                     if fixup_policy == 'warn':
                         self.report_warning(
                             '%s: writing DASH m4a. '
@@ -1948,9 +1978,9 @@ class YoutubeDL(object):
                     else:
                         assert fixup_policy in ('ignore', 'never')
 
-                if (info_dict.get('protocol') == 'm3u8_native' or
-                        info_dict.get('protocol') == 'm3u8' and
-                        self.params.get('hls_prefer_native')):
+                if (info_dict.get('protocol') == 'm3u8_native'
+                        or info_dict.get('protocol') == 'm3u8'
+                        and self.params.get('hls_prefer_native')):
                     if fixup_policy == 'warn':
                         self.report_warning('%s: malformed AAC bitstream detected.' % (
                             info_dict['id']))
@@ -1976,10 +2006,10 @@ class YoutubeDL(object):
     def download(self, url_list):
         """Download a given list of URLs."""
         outtmpl = self.params.get('outtmpl', DEFAULT_OUTTMPL)
-        if (len(url_list) > 1 and
-                outtmpl != '-' and
-                '%' not in outtmpl and
-                self.params.get('max_downloads') != 1):
+        if (len(url_list) > 1
+                and outtmpl != '-'
+                and '%' not in outtmpl
+                and self.params.get('max_downloads') != 1):
             raise SameFileError(outtmpl)
 
         for url in url_list:
@@ -2044,15 +2074,24 @@ class YoutubeDL(object):
                         self.report_warning('Unable to remove downloaded original file')
 
     def _make_archive_id(self, info_dict):
+        video_id = info_dict.get('id')
+        if not video_id:
+            return
         # Future-proof against any change in case
         # and backwards compatibility with prior versions
-        extractor = info_dict.get('extractor_key')
-        if extractor is None:
-            if 'id' in info_dict:
-                extractor = info_dict.get('ie_key')  # key in a playlist
+        extractor = info_dict.get('extractor_key') or info_dict.get('ie_key')  # key in a playlist
         if extractor is None:
-            return None  # Incomplete video information
-        return extractor.lower() + ' ' + info_dict['id']
+            url = str_or_none(info_dict.get('url'))
+            if not url:
+                return
+            # Try to find matching extractor for the URL and take its ie_key
+            for ie in self._ies:
+                if ie.suitable(url):
+                    extractor = ie.ie_key()
+                    break
+            else:
+                return
+        return extractor.lower() + ' ' + video_id
 
     def in_download_archive(self, info_dict):
         fn = self.params.get('download_archive')
@@ -2060,7 +2099,7 @@ class YoutubeDL(object):
             return False
 
         vid_id = self._make_archive_id(info_dict)
-        if vid_id is None:
+        if not vid_id:
             return False  # Incomplete video information
 
         try:
@@ -2115,8 +2154,8 @@ class YoutubeDL(object):
             if res:
                 res += ', '
             res += '%s container' % fdict['container']
-        if (fdict.get('vcodec') is not None and
-                fdict.get('vcodec') != 'none'):
+        if (fdict.get('vcodec') is not None
+                and fdict.get('vcodec') != 'none'):
             if res:
                 res += ', '
             res += fdict['vcodec']
@@ -2203,7 +2242,7 @@ class YoutubeDL(object):
             return
 
         if type('') is not compat_str:
-            # Python 2.6 on SLES11 SP1 (https://github.com/rg3/youtube-dl/issues/3326)
+            # Python 2.6 on SLES11 SP1 (https://github.com/ytdl-org/youtube-dl/issues/3326)
             self.report_warning(
                 'Your Python is broken! Update to a newer and supported version')
 
@@ -2285,10 +2324,9 @@ class YoutubeDL(object):
             self.cookiejar = compat_cookiejar.CookieJar()
         else:
             opts_cookiefile = expand_path(opts_cookiefile)
-            self.cookiejar = compat_cookiejar.MozillaCookieJar(
-                opts_cookiefile)
+            self.cookiejar = YoutubeDLCookieJar(opts_cookiefile)
             if os.access(opts_cookiefile, os.R_OK):
-                self.cookiejar.load()
+                self.cookiejar.load(ignore_discard=True, ignore_expires=True)
 
         cookie_processor = YoutubeDLCookieProcessor(self.cookiejar)
         if opts_proxy is not None:
@@ -2298,7 +2336,7 @@ class YoutubeDL(object):
                 proxies = {'http': opts_proxy, 'https': opts_proxy}
         else:
             proxies = compat_urllib_request.getproxies()
-            # Set HTTPS proxy to HTTP one if given (https://github.com/rg3/youtube-dl/issues/805)
+            # Set HTTPS proxy to HTTP one if given (https://github.com/ytdl-org/youtube-dl/issues/805)
             if 'http' in proxies and 'https' not in proxies:
                 proxies['https'] = proxies['http']
         proxy_handler = PerRequestProxyHandler(proxies)
@@ -2306,12 +2344,13 @@ class YoutubeDL(object):
         debuglevel = 1 if self.params.get('debug_printtraffic') else 0
         https_handler = make_HTTPS_handler(self.params, debuglevel=debuglevel)
         ydlh = YoutubeDLHandler(self.params, debuglevel=debuglevel)
+        redirect_handler = YoutubeDLRedirectHandler()
         data_handler = compat_urllib_request_DataHandler()
 
         # When passing our own FileHandler instance, build_opener won't add the
         # default FileHandler and allows us to disable the file protocol, which
         # can be used for malicious purposes (see
-        # https://github.com/rg3/youtube-dl/issues/8227)
+        # https://github.com/ytdl-org/youtube-dl/issues/8227)
         file_handler = compat_urllib_request.FileHandler()
 
         def file_open(*args, **kwargs):
@@ -2319,11 +2358,11 @@ class YoutubeDL(object):
         file_handler.file_open = file_open
 
         opener = compat_urllib_request.build_opener(
-            proxy_handler, https_handler, cookie_processor, ydlh, data_handler, file_handler)
+            proxy_handler, https_handler, cookie_processor, ydlh, redirect_handler, data_handler, file_handler)
 
         # Delete the default user-agent header, which would otherwise apply in
         # cases where our custom HTTP handler doesn't come into play
-        # (See https://github.com/rg3/youtube-dl/issues/1309 for details)
+        # (See https://github.com/ytdl-org/youtube-dl/issues/1309 for details)
         opener.addheaders = []
         self._opener = opener