X-Git-Url: http://git.bitcoin.ninja/index.cgi?p=youtube-dl;a=blobdiff_plain;f=youtube_dl%2F__init__.py;h=c21bf6d4afb41eeba91f2d82c556a31ccf6b0601;hp=c4f64893d491f8e6f9927fd644b967cf0d9c3542;hb=18b4e04f1c663e0ea695f6501b860f85af9d7ca1;hpb=b9fc428494b22623529d364387b8693cc3cb1503 diff --git a/youtube_dl/__init__.py b/youtube_dl/__init__.py index c4f64893d..c21bf6d4a 100644 --- a/youtube_dl/__init__.py +++ b/youtube_dl/__init__.py @@ -1,9 +1,6 @@ #!/usr/bin/env python # -*- coding: utf-8 -*- -from __future__ import with_statement -from __future__ import absolute_import - __authors__ = ( 'Ricardo Garcia Gonzalez', 'Danny Colligan', @@ -24,13 +21,22 @@ __authors__ = ( 'Jaime Marquínez Ferrándiz', 'Jeff Crouse', 'Osama Khalid', - ) + 'Michael Walter', + 'M. Yasoob Ullah Khalid', + 'Julien Fraichard', + 'Johny Mo Swag', + 'Axel Noack', + 'Albert Kim', + 'Pierre Rudloff', +) __license__ = 'Public Domain' +import codecs import getpass import optparse import os +import random import re import shlex import socket @@ -43,10 +49,11 @@ from .utils import * from .update import update_self from .version import __version__ from .FileDownloader import * -from .InfoExtractors import gen_extractors +from .extractor import gen_extractors +from .YoutubeDL import YoutubeDL from .PostProcessor import * -def parseOpts(): +def parseOpts(overrideArguments=None): def _readOptions(filename_bytes): try: optionf = open(filename_bytes) @@ -113,6 +120,8 @@ def parseOpts(): 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') @@ -122,27 +131,26 @@ def parseOpts(): 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='download rate limit (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) general.add_option('--user-agent', dest='user_agent', help='specify a custom user agent', metavar='UA') + general.add_option('--referer', + dest='referer', help='specify a custom referer, use if the video access is restricted to one domain', + metavar='REF', default=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('--test', action='store_true', dest='test', default=False, help=optparse.SUPPRESS_HELP) + 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.') + selection.add_option('--playlist-start', dest='playliststart', metavar='NUMBER', help='playlist video to start at (default is %default)', default=1) @@ -153,6 +161,9 @@ def parseOpts(): selection.add_option('--max-downloads', metavar='NUMBER', dest='max_downloads', help='Abort after downloading NUMBER files', default=None) selection.add_option('--min-filesize', metavar='SIZE', dest='min_filesize', help="Do not download any videos smaller than SIZE (e.g. 50k or 44.6m)", default=None) selection.add_option('--max-filesize', metavar='SIZE', dest='max_filesize', help="Do not download any videos larger than SIZE (e.g. 50k or 44.6m)", default=None) + selection.add_option('--date', metavar='DATE', dest='date', help='download only videos uploaded in this date', default=None) + selection.add_option('--datebefore', metavar='DATE', dest='datebefore', help='download only videos uploaded before this date', default=None) + selection.add_option('--dateafter', metavar='DATE', dest='dateafter', help='download only videos uploaded after this date', default=None) authentication.add_option('-u', '--username', @@ -161,10 +172,13 @@ def parseOpts(): 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', - action='store', dest='format', metavar='FORMAT', help='video format code') + action='store', dest='format', metavar='FORMAT', + help='video format code, specifiy the order of preference using slashes: "-f 22/17/18"') video_format.add_option('--all-formats', action='store_const', dest='format', help='download all available video formats', const='all') video_format.add_option('--prefer-free-formats', @@ -173,25 +187,37 @@ def parseOpts(): 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', - action='store_true', dest='onlysubtitles', - help='downloads only the subtitles (no video)', default=False) - video_format.add_option('--all-subs', + help='write subtitle file', default=False) + subtitles.add_option('--write-auto-sub', '--write-automatic-sub', + action='store_true', dest='writeautomaticsub', + help='write automatic subtitle file (youtube only)', default=False) + 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', - action='store', dest='subtitlesformat', metavar='LANG', - help='subtitle format [srt/sbv] (default=srt) (currently youtube only)', default='srt') - video_format.add_option('--sub-lang', '--srt-lang', + help='lists all available subtitles for the video', default=False) + subtitles.add_option('--sub-format', + action='store', dest='subtitlesformat', metavar='FORMAT', + help='subtitle format (default=srt) ([sbv/vtt] youtube only)', default='srt') + subtitles.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\'') + 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) verbosity.add_option('-s', '--simulate', @@ -202,6 +228,8 @@ def parseOpts(): action='store_true', dest='geturl', help='simulate, quiet but print URL', default=False) verbosity.add_option('-e', '--get-title', action='store_true', dest='gettitle', help='simulate, quiet but print title', default=False) + verbosity.add_option('--get-id', + action='store_true', dest='getid', help='simulate, quiet but print id', default=False) verbosity.add_option('--get-thumbnail', action='store_true', dest='getthumbnail', help='simulate, quiet but print thumbnail URL', default=False) @@ -223,18 +251,33 @@ def parseOpts(): help='display progress in console titlebar', default=False) verbosity.add_option('-v', '--verbose', action='store_true', dest='verbose', help='print various debugging information', default=False) + verbosity.add_option('--dump-intermediate-pages', + action='store_true', dest='dump_intermediate_pages', default=False, + help='print downloaded pages to debug problems(very verbose)') filesystem.add_option('-t', '--title', - action='store_true', dest='usetitle', help='use title in file name', default=False) + action='store_true', dest='usetitle', help='use title in file name (default)', default=False) filesystem.add_option('--id', - action='store_true', dest='useid', help='use video ID in file name', default=False) + action='store_true', dest='useid', help='use only video ID in file name', default=False) filesystem.add_option('-l', '--literal', action='store_true', dest='usetitle', help='[deprecated] alias of --title', default=False) filesystem.add_option('-A', '--auto-number', action='store_true', dest='autonumber', help='number downloaded files starting from 00000', default=False) filesystem.add_option('-o', '--output', - dest='outtmpl', metavar='TEMPLATE', help='output filename template. Use %(title)s to get the title, %(uploader)s for the uploader name, %(uploader_id)s for the uploader nickname if different, %(autonumber)s to get an automatically incremented number, %(ext)s for the filename extension, %(upload_date)s for the upload date (YYYYMMDD), %(extractor)s for the provider (youtube, metacafe, etc), %(id)s for the video id and %% for a literal percent. Use - to output to stdout. Can also be used to download to a different directory, for example with -o \'/my/downloads/%(uploader)s/%(title)s-%(id)s.%(ext)s\' .') + dest='outtmpl', metavar='TEMPLATE', + help=('output filename template. Use %(title)s to get the title, ' + '%(uploader)s for the uploader name, %(uploader_id)s for the uploader nickname if different, ' + '%(autonumber)s to get an automatically incremented number, ' + '%(ext)s for the filename extension, %(upload_date)s for the upload date (YYYYMMDD), ' + '%(extractor)s for the provider (youtube, metacafe, etc), ' + '%(id)s for the video id , %(playlist)s for the playlist the video is in, ' + '%(playlist_index)s for the position in the playlist and %% for a literal percent. ' + 'Use - to output to stdout. Can also be used to download to a different directory, ' + 'for example with -o \'/my/downloads/%(uploader)s/%(title)s-%(id)s.%(ext)s\' .')) + filesystem.add_option('--autonumber-size', + dest='autonumber_size', metavar='NUMBER', + help='Specifies the number of digits in %(autonumber)s when it is present in output filename template or --autonumber option is given') filesystem.add_option('--restrict-filenames', action='store_true', dest='restrictfilenames', help='Restrict filenames to only ASCII characters, and avoid "&" and spaces in filenames', default=False) @@ -260,6 +303,9 @@ def parseOpts(): filesystem.add_option('--write-info-json', action='store_true', dest='writeinfojson', help='write video metadata to a .info.json file', default=False) + filesystem.add_option('--write-thumbnail', + action='store_true', dest='writethumbnail', + help='write thumbnail image to disk', default=False) postproc.add_option('-x', '--extract-audio', action='store_true', dest='extractaudio', default=False, @@ -278,24 +324,43 @@ def parseOpts(): 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) - xdg_config_home = os.environ.get('XDG_CONFIG_HOME') - if xdg_config_home: - userConf = os.path.join(xdg_config_home, 'youtube-dl.conf') + if overrideArguments is not None: + opts, args = parser.parse_args(overrideArguments) + if opts.verbose: + sys.stderr.write(u'[debug] Override config: ' + repr(overrideArguments) + '\n') else: - userConf = os.path.join(os.path.expanduser('~'), '.config', 'youtube-dl.conf') - argv = _readOptions('/etc/youtube-dl.conf') + _readOptions(userConf) + sys.argv[1:] - opts, args = parser.parse_args(argv) + xdg_config_home = os.environ.get('XDG_CONFIG_HOME') + if xdg_config_home: + userConfFile = os.path.join(xdg_config_home, 'youtube-dl.conf') + else: + userConfFile = os.path.join(os.path.expanduser('~'), '.config', 'youtube-dl.conf') + systemConf = _readOptions('/etc/youtube-dl.conf') + userConf = _readOptions(userConfFile) + commandLineConf = sys.argv[1:] + argv = systemConf + userConf + commandLineConf + opts, args = parser.parse_args(argv) + if opts.verbose: + sys.stderr.write(u'[debug] System config: ' + repr(systemConf) + '\n') + sys.stderr.write(u'[debug] User config: ' + repr(userConf) + '\n') + sys.stderr.write(u'[debug] Command-line args: ' + repr(commandLineConf) + '\n') return parser, opts, args -def _real_main(): - parser, opts, args = parseOpts() +def _real_main(argv=None): + # Compatibility fixes for Windows + if sys.platform == 'win32': + # https://github.com/rg3/youtube-dl/issues/820 + codecs.register(lambda name: codecs.lookup('utf-8') if name == 'cp65001' else None) + + parser, opts, args = parseOpts(argv) # Open appropriate CookieJar if opts.cookiefile is None: @@ -314,9 +379,13 @@ def _real_main(): 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 @@ -330,6 +399,8 @@ def _real_main(): 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 @@ -337,27 +408,50 @@ def _real_main(): # General configuration cookie_processor = compat_urllib_request.HTTPCookieProcessor(jar) - proxy_handler = compat_urllib_request.ProxyHandler() - opener = compat_urllib_request.build_opener(proxy_handler, cookie_processor, YoutubeDLHandler()) + if opts.proxy is not None: + if opts.proxy == '': + proxies = {} + else: + 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) + if 'http' in proxies and 'https' not in proxies: + proxies['https'] = proxies['http'] + 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()) 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: - parser.error(u'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: @@ -411,6 +505,10 @@ def _real_main(): if opts.recodevideo is not None: if opts.recodevideo not in ['mp4', 'flv', 'webm', 'ogg']: parser.error(u'invalid video recode format specified') + if opts.date is not None: + date = DateRange.day(opts.date) + else: + date = DateRange(opts.dateafter, opts.datebefore) if sys.version_info < (3,): # In Python 2, sys.argv is a bytestring (also note http://bugs.python.org/issue2128 for Windows systems) @@ -423,26 +521,29 @@ def _real_main(): or (opts.usetitle and u'%(title)s-%(id)s.%(ext)s') or (opts.useid and u'%(id)s.%(ext)s') or (opts.autonumber and u'%(autonumber)s-%(id)s.%(ext)s') - or u'%(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, - 'quiet': (opts.quiet or opts.geturl or opts.gettitle or opts.getthumbnail or opts.getdescription or opts.getfilename or opts.getformat), + '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, + 'forceid': opts.getid, 'forcethumbnail': opts.getthumbnail, 'forcedescription': opts.getdescription, 'forcefilename': opts.getfilename, 'forceformat': opts.getformat, 'simulate': opts.simulate, - 'skip_download': (opts.skip_download or opts.simulate or opts.geturl or opts.gettitle or opts.getthumbnail or opts.getdescription or opts.getfilename or opts.getformat), + 'skip_download': (opts.skip_download or opts.simulate or opts.geturl or opts.gettitle or opts.getid or opts.getthumbnail or opts.getdescription or opts.getfilename or opts.getformat), 'format': opts.format, 'format_limit': opts.format_limit, 'listformats': opts.listformats, 'outtmpl': outtmpl, + 'autonumber_size': opts.autonumber_size, 'restrictfilenames': opts.restrictfilenames, 'ignoreerrors': opts.ignoreerrors, 'ratelimit': opts.ratelimit, @@ -461,8 +562,9 @@ def _real_main(): 'updatetime': opts.updatetime, 'writedescription': opts.writedescription, 'writeinfojson': opts.writeinfojson, + 'writethumbnail': opts.writethumbnail, 'writesubtitles': opts.writesubtitles, - 'onlysubtitles': opts.onlysubtitles, + 'writeautomaticsub': opts.writeautomaticsub, 'allsubtitles': opts.allsubtitles, 'listsubtitles': opts.listsubtitles, 'subtitlesformat': opts.subtitlesformat, @@ -472,38 +574,44 @@ def _real_main(): 'max_downloads': opts.max_downloads, 'prefer_free_formats': opts.prefer_free_formats, 'verbose': opts.verbose, + 'dump_intermediate_pages': opts.dump_intermediate_pages, 'test': opts.test, 'keepvideo': opts.keepvideo, 'min_filesize': opts.min_filesize, - 'max_filesize': opts.max_filesize + 'max_filesize': opts.max_filesize, + 'daterange': date, }) 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.platform()) + 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)) # 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: @@ -513,9 +621,9 @@ def _real_main(): 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 @@ -527,9 +635,9 @@ def _real_main(): sys.exit(retcode) -def main(): +def main(argv=None): try: - _real_main() + _real_main(argv) except DownloadError: sys.exit(1) except SameFileError: