X-Git-Url: https://git.rapsys.eu/youtubedl/blobdiff_plain/bddc9fc577d16b1428924bf8a5c37ef1d9295f14..3cdde5eb40ad4ca4809bafa3bb7bf12462f486b7:/youtube_dl/__init__.py diff --git a/youtube_dl/__init__.py b/youtube_dl/__init__.py index a8b62a6..431460c 100644 --- a/youtube_dl/__init__.py +++ b/youtube_dl/__init__.py @@ -25,7 +25,10 @@ __authors__ = ( 'M. Yasoob Ullah Khalid', 'Julien Fraichard', 'Johny Mo Swag', - ) + 'Axel Noack', + 'Albert Kim', + 'Pierre Rudloff', +) __license__ = 'Public Domain' @@ -33,6 +36,7 @@ import codecs import getpass import optparse import os +import random import re import shlex import socket @@ -41,11 +45,13 @@ import sys import warnings import platform + from .utils import * from .update import update_self from .version import __version__ from .FileDownloader import * from .extractor import gen_extractors +from .YoutubeDL import YoutubeDL from .PostProcessor import * def parseOpts(overrideArguments=None): @@ -78,6 +84,9 @@ def parseOpts(overrideArguments=None): return "".join(opts) + def _comma_separated_values_options_callback(option, opt_str, value, parser): + setattr(parser.values, option.dest, value.split(',')) + def _find_term_columns(): columns = os.environ.get('COLUMNS', None) if columns: @@ -91,6 +100,16 @@ def parseOpts(overrideArguments=None): pass return None + def _hide_login_info(opts): + opts = list(opts) + for private_opt in ['-p', '--password', '-u', '--username']: + try: + i = opts.index(private_opt) + opts[i+1] = '' + except ValueError: + pass + return opts + max_width = 80 max_help_position = 80 @@ -115,6 +134,8 @@ def parseOpts(overrideArguments=None): selection = optparse.OptionGroup(parser, 'Video Selection') authentication = optparse.OptionGroup(parser, 'Authentication Options') video_format = optparse.OptionGroup(parser, 'Video Format Options') + subtitles = optparse.OptionGroup(parser, 'Subtitle Options') + downloader = optparse.OptionGroup(parser, 'Download Options') postproc = optparse.OptionGroup(parser, 'Post-processing Options') filesystem = optparse.OptionGroup(parser, 'Filesystem Options') verbosity = optparse.OptionGroup(parser, 'Verbosity / Simulation Options') @@ -124,18 +145,9 @@ def parseOpts(overrideArguments=None): general.add_option('-v', '--version', action='version', help='print program version and exit') general.add_option('-U', '--update', - action='store_true', dest='update_self', help='update this program to latest version') + action='store_true', dest='update_self', help='update this program to latest version. Make sure that you have sufficient permissions (run with sudo if needed)') general.add_option('-i', '--ignore-errors', action='store_true', dest='ignoreerrors', help='continue on download errors', default=False) - general.add_option('-r', '--rate-limit', - dest='ratelimit', metavar='LIMIT', help='maximum download rate (e.g. 50k or 44.6m)') - general.add_option('-R', '--retries', - dest='retries', metavar='RETRIES', help='number of retries (default is %default)', default=10) - general.add_option('--buffer-size', - dest='buffersize', metavar='SIZE', help='size of download buffer (e.g. 1024 or 16k) (default is %default)', default="1024") - general.add_option('--no-resize-buffer', - action='store_true', dest='noresizebuffer', - help='do not automatically adjust the buffer size. By default, the buffer size is automatically resized from an initial value of SIZE.', default=False) general.add_option('--dump-user-agent', action='store_true', dest='dump_user_agent', help='display the current browser identification', default=False) @@ -147,9 +159,12 @@ def parseOpts(overrideArguments=None): general.add_option('--list-extractors', action='store_true', dest='list_extractors', help='List all supported extractors and the URLs they would handle', default=False) + general.add_option('--extractor-descriptions', + action='store_true', dest='list_extractor_descriptions', + help='Output descriptions of all supported extractors', default=False) general.add_option('--proxy', dest='proxy', default=None, help='Use the specified HTTP/HTTPS proxy', metavar='URL') general.add_option('--no-check-certificate', action='store_true', dest='no_check_certificate', default=False, help='Suppress HTTPS certificate validation.') - general.add_option('--test', action='store_true', dest='test', default=False, help=optparse.SUPPRESS_HELP) + selection.add_option('--playlist-start', dest='playliststart', metavar='NUMBER', help='playlist video to start at (default is %default)', default=1) @@ -171,6 +186,8 @@ def parseOpts(overrideArguments=None): dest='password', metavar='PASSWORD', help='account password') authentication.add_option('-n', '--netrc', action='store_true', dest='usenetrc', help='use .netrc authentication data', default=False) + authentication.add_option('--video-password', + dest='videopassword', metavar='PASSWORD', help='video password (vimeo only)') video_format.add_option('-f', '--format', @@ -184,24 +201,40 @@ def parseOpts(overrideArguments=None): action='store', dest='format_limit', metavar='FORMAT', help='highest quality format to download') video_format.add_option('-F', '--list-formats', action='store_true', dest='listformats', help='list all available formats (currently youtube only)') - video_format.add_option('--write-sub', '--write-srt', + + subtitles.add_option('--write-sub', '--write-srt', action='store_true', dest='writesubtitles', help='write subtitle file (currently youtube only)', default=False) - video_format.add_option('--only-sub', + subtitles.add_option('--write-auto-sub', '--write-automatic-sub', + action='store_true', dest='writeautomaticsub', + help='write automatic subtitle file (currently youtube only)', default=False) + subtitles.add_option('--only-sub', action='store_true', dest='skip_download', help='[deprecated] alias of --skip-download', default=False) - video_format.add_option('--all-subs', + subtitles.add_option('--all-subs', action='store_true', dest='allsubtitles', - help='downloads all the available subtitles of the video (currently youtube only)', default=False) - video_format.add_option('--list-subs', + help='downloads all the available subtitles of the video', default=False) + subtitles.add_option('--list-subs', action='store_true', dest='listsubtitles', - help='lists all available subtitles for the video (currently youtube only)', default=False) - video_format.add_option('--sub-format', + help='lists all available subtitles for the video', default=False) + subtitles.add_option('--sub-format', action='store', dest='subtitlesformat', metavar='FORMAT', - help='subtitle format [srt/sbv] (default=srt) (currently youtube only)', default='srt') - video_format.add_option('--sub-lang', '--srt-lang', - action='store', dest='subtitleslang', metavar='LANG', - help='language of the subtitles to download (optional) use IETF language tags like \'en\'') + help='subtitle format (default=srt) ([sbv/vtt] youtube only)', default='srt') + subtitles.add_option('--sub-lang', '--sub-langs', '--srt-lang', + action='callback', dest='subtitleslang', metavar='LANGS', type='str', + default=[], callback=_comma_separated_values_options_callback, + help='languages of the subtitles to download (optional) separated by commas, use IETF language tags like \'en,pt\'') + + downloader.add_option('-r', '--rate-limit', + dest='ratelimit', metavar='LIMIT', help='maximum download rate (e.g. 50k or 44.6m)') + downloader.add_option('-R', '--retries', + dest='retries', metavar='RETRIES', help='number of retries (default is %default)', default=10) + downloader.add_option('--buffer-size', + dest='buffersize', metavar='SIZE', help='size of download buffer (e.g. 1024 or 16k) (default is %default)', default="1024") + downloader.add_option('--no-resize-buffer', + action='store_true', dest='noresizebuffer', + help='do not automatically adjust the buffer size. By default, the buffer size is automatically resized from an initial value of SIZE.', default=False) + downloader.add_option('--test', action='store_true', dest='test', default=False, help=optparse.SUPPRESS_HELP) verbosity.add_option('-q', '--quiet', action='store_true', dest='quiet', help='activates quiet mode', default=False) @@ -305,20 +338,24 @@ def parseOpts(overrideArguments=None): help='keeps the video file on disk after the post-processing; the video is erased by default') postproc.add_option('--no-post-overwrites', action='store_true', dest='nopostoverwrites', default=False, help='do not overwrite post-processed files; the post-processed files are overwritten by default') + postproc.add_option('--embed-subs', action='store_true', dest='embedsubtitles', default=False, + help='embed subtitles in the video (only for mp4 videos)') parser.add_option_group(general) parser.add_option_group(selection) + parser.add_option_group(downloader) parser.add_option_group(filesystem) parser.add_option_group(verbosity) parser.add_option_group(video_format) + parser.add_option_group(subtitles) parser.add_option_group(authentication) parser.add_option_group(postproc) if overrideArguments is not None: opts, args = parser.parse_args(overrideArguments) if opts.verbose: - print(u'[debug] Override config: ' + repr(overrideArguments)) + sys.stderr.write(u'[debug] Override config: ' + repr(overrideArguments) + '\n') else: xdg_config_home = os.environ.get('XDG_CONFIG_HOME') if xdg_config_home: @@ -327,13 +364,13 @@ def parseOpts(overrideArguments=None): userConfFile = os.path.join(os.path.expanduser('~'), '.config', 'youtube-dl.conf') systemConf = _readOptions('/etc/youtube-dl.conf') userConf = _readOptions(userConfFile) - commandLineConf = sys.argv[1:] + commandLineConf = sys.argv[1:] argv = systemConf + userConf + commandLineConf opts, args = parser.parse_args(argv) if opts.verbose: - print(u'[debug] System config: ' + repr(systemConf)) - print(u'[debug] User config: ' + repr(userConf)) - print(u'[debug] Command-line args: ' + repr(commandLineConf)) + sys.stderr.write(u'[debug] System config: ' + repr(_hide_login_info(systemConf)) + '\n') + sys.stderr.write(u'[debug] User config: ' + repr(_hide_login_info(userConf)) + '\n') + sys.stderr.write(u'[debug] Command-line args: ' + repr(_hide_login_info(commandLineConf)) + '\n') return parser, opts, args @@ -361,14 +398,14 @@ def _real_main(argv=None): # Set user agent if opts.user_agent is not None: std_headers['User-Agent'] = opts.user_agent - + # Set referer if opts.referer is not None: std_headers['Referer'] = opts.referer # Dump user agent if opts.dump_user_agent: - print(std_headers['User-Agent']) + compat_print(std_headers['User-Agent']) sys.exit(0) # Batch file verification @@ -382,6 +419,8 @@ def _real_main(argv=None): batchurls = batchfd.readlines() batchurls = [x.strip() for x in batchurls] batchurls = [x for x in batchurls if len(x) > 0 and not re.search(r'^[#/;]', x)] + if opts.verbose: + sys.stderr.write(u'[debug] Batch file urls: ' + repr(batchurls) + u'\n') except IOError: sys.exit(u'ERROR: batch file could not be read') all_urls = batchurls + args @@ -402,25 +441,41 @@ def _real_main(argv=None): proxy_handler = compat_urllib_request.ProxyHandler(proxies) https_handler = make_HTTPS_handler(opts) opener = compat_urllib_request.build_opener(https_handler, proxy_handler, cookie_processor, YoutubeDLHandler()) + # 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) + opener.addheaders =[] compat_urllib_request.install_opener(opener) socket.setdefaulttimeout(300) # 5 minutes should be enough (famous last words) extractors = gen_extractors() if opts.list_extractors: - for ie in extractors: - print(ie.IE_NAME + (' (CURRENTLY BROKEN)' if not ie._WORKING else '')) + for ie in sorted(extractors, key=lambda ie: ie.IE_NAME.lower()): + compat_print(ie.IE_NAME + (' (CURRENTLY BROKEN)' if not ie._WORKING else '')) matchedUrls = [url for url in all_urls if ie.suitable(url)] all_urls = [url for url in all_urls if url not in matchedUrls] for mu in matchedUrls: - print(u' ' + mu) + compat_print(u' ' + mu) + sys.exit(0) + if opts.list_extractor_descriptions: + for ie in sorted(extractors, key=lambda ie: ie.IE_NAME.lower()): + if not ie._WORKING: + continue + desc = getattr(ie, 'IE_DESC', ie.IE_NAME) + if hasattr(ie, 'SEARCH_KEY'): + _SEARCHES = (u'cute kittens', u'slithering pythons', u'falling cat', u'angry poodle', u'purple fish', u'running tortoise') + _COUNTS = (u'', u'5', u'10', u'all') + desc += u' (Example: "%s%s:%s" )' % (ie.SEARCH_KEY, random.choice(_COUNTS), random.choice(_SEARCHES)) + compat_print(desc) sys.exit(0) + # Conflicting, missing and erroneous options if opts.usenetrc and (opts.username is not None or opts.password is not None): parser.error(u'using .netrc conflicts with giving username/password') if opts.password is not None and opts.username is None: - print(u'WARNING: account username missing') + parser.error(u' account username missing\n') if opts.outtmpl is not None and (opts.usetitle or opts.autonumber or opts.useid): parser.error(u'using output template conflicts with using title, video ID or auto number') if opts.usetitle and opts.useid: @@ -492,11 +547,12 @@ def _real_main(argv=None): or (opts.autonumber and u'%(autonumber)s-%(id)s.%(ext)s') or u'%(title)s-%(id)s.%(ext)s') - # File downloader - fd = FileDownloader({ + # YoutubeDL + ydl = YoutubeDL({ 'usenetrc': opts.usenetrc, 'username': opts.username, 'password': opts.password, + 'videopassword': opts.videopassword, 'quiet': (opts.quiet or opts.geturl or opts.gettitle or opts.getid or opts.getthumbnail or opts.getdescription or opts.getfilename or opts.getformat), 'forceurl': opts.geturl, 'forcetitle': opts.gettitle, @@ -532,10 +588,11 @@ def _real_main(argv=None): 'writeinfojson': opts.writeinfojson, 'writethumbnail': opts.writethumbnail, 'writesubtitles': opts.writesubtitles, + 'writeautomaticsub': opts.writeautomaticsub, 'allsubtitles': opts.allsubtitles, 'listsubtitles': opts.listsubtitles, 'subtitlesformat': opts.subtitlesformat, - 'subtitleslang': opts.subtitleslang, + 'subtitleslangs': opts.subtitleslang, 'matchtitle': decodeOption(opts.matchtitle), 'rejecttitle': decodeOption(opts.rejecttitle), 'max_downloads': opts.max_downloads, @@ -550,31 +607,37 @@ def _real_main(argv=None): }) if opts.verbose: - fd.to_screen(u'[debug] youtube-dl version ' + __version__) + sys.stderr.write(u'[debug] youtube-dl version ' + __version__ + u'\n') try: - sp = subprocess.Popen(['git', 'rev-parse', '--short', 'HEAD'], stdout=subprocess.PIPE, stderr=subprocess.PIPE, - cwd=os.path.dirname(os.path.abspath(__file__))) + sp = subprocess.Popen( + ['git', 'rev-parse', '--short', 'HEAD'], + stdout=subprocess.PIPE, stderr=subprocess.PIPE, + cwd=os.path.dirname(os.path.abspath(__file__))) out, err = sp.communicate() out = out.decode().strip() if re.match('[0-9a-f]+', out): - fd.to_screen(u'[debug] Git HEAD: ' + out) + sys.stderr.write(u'[debug] Git HEAD: ' + out + u'\n') except: - pass - fd.to_screen(u'[debug] Python version %s - %s' %(platform.python_version(), platform.platform())) - fd.to_screen(u'[debug] Proxy map: ' + str(proxy_handler.proxies)) + try: + sys.exc_clear() + except: + pass + sys.stderr.write(u'[debug] Python version %s - %s' %(platform.python_version(), platform_name()) + u'\n') + sys.stderr.write(u'[debug] Proxy map: ' + str(proxy_handler.proxies) + u'\n') - for extractor in extractors: - fd.add_info_extractor(extractor) + ydl.add_default_info_extractors() # PostProcessors if opts.extractaudio: - fd.add_post_processor(FFmpegExtractAudioPP(preferredcodec=opts.audioformat, preferredquality=opts.audioquality, nopostoverwrites=opts.nopostoverwrites)) + ydl.add_post_processor(FFmpegExtractAudioPP(preferredcodec=opts.audioformat, preferredquality=opts.audioquality, nopostoverwrites=opts.nopostoverwrites)) if opts.recodevideo: - fd.add_post_processor(FFmpegVideoConvertor(preferedformat=opts.recodevideo)) + ydl.add_post_processor(FFmpegVideoConvertor(preferedformat=opts.recodevideo)) + if opts.embedsubtitles: + ydl.add_post_processor(FFmpegEmbedSubtitlePP(subtitlesformat=opts.subtitlesformat)) # Update version if opts.update_self: - update_self(fd.to_screen, opts.verbose, sys.argv[0]) + update_self(ydl.to_screen, opts.verbose, sys.argv[0]) # Maybe do nothing if len(all_urls) < 1: @@ -584,9 +647,9 @@ def _real_main(argv=None): sys.exit() try: - retcode = fd.download(all_urls) + retcode = ydl.download(all_urls) except MaxDownloadsReached: - fd.to_screen(u'--max-download limit reached, aborting.') + ydl.to_screen(u'--max-download limit reached, aborting.') retcode = 101 # Dump cookie jar if requested