X-Git-Url: https://git.rapsys.eu/youtubedl/blobdiff_plain/af478477605bdf3f5d57562035885cfee905f379..252064fa082f84e06d1251185428b44c75c59109:/youtube_dl/YoutubeDL.py?ds=sidebyside diff --git a/youtube_dl/YoutubeDL.py b/youtube_dl/YoutubeDL.py old mode 100644 new mode 100755 index 42cbcf6..3dff723 --- a/youtube_dl/YoutubeDL.py +++ b/youtube_dl/YoutubeDL.py @@ -4,9 +4,11 @@ from __future__ import absolute_import, unicode_literals import collections +import datetime import errno import io import json +import locale import os import platform import re @@ -29,6 +31,7 @@ from .utils import ( ContentTooShortError, date_from_str, DateRange, + DEFAULT_OUTTMPL, determine_ext, DownloadError, encodeFilename, @@ -93,6 +96,7 @@ class YoutubeDL(object): usenetrc: Use netrc for authentication instead. verbose: Print additional info to stdout. quiet: Do not print messages to stdout. + no_warnings: Do not print out anything for warnings. forceurl: Force printing final URL. forcetitle: Force printing title. forceid: Force printing ID. @@ -147,6 +151,8 @@ class YoutubeDL(object): again. cookiefile: File name where cookies should be read from and dumped to. nocheckcertificate:Do not verify SSL certificates + prefer_insecure: Use HTTP instead of HTTPS to retrieve information. + At the moment, this is only supported by YouTube. proxy: URL of the proxy server to use socket_timeout: Time to wait for unresponsive hosts, in seconds bidi_workaround: Work around buggy terminals without bidirectional text @@ -155,6 +161,7 @@ class YoutubeDL(object): include_ads: Download ads as well default_search: Prepend this string if an input url is not valid. 'auto' for elaborate guessing + encoding: Use this encoding instead of the system-specified. The following parameters are not used by YoutubeDL itself, they are used by the FileDownloader: @@ -280,6 +287,9 @@ class YoutubeDL(object): """Print message to stdout if not in quiet mode.""" return self.to_stdout(message, skip_eol, check_quiet=True) + def _write_string(self, s, out=None): + write_string(s, out=out, encoding=self.params.get('encoding')) + def to_stdout(self, message, skip_eol=False, check_quiet=False): """Print message to stdout if not in quiet mode.""" if self.params.get('logger'): @@ -289,7 +299,7 @@ class YoutubeDL(object): terminator = ['\n', ''][skip_eol] output = message + terminator - write_string(output, self._screen_file) + self._write_string(output, self._screen_file) def to_stderr(self, message): """Print message to stderr.""" @@ -299,7 +309,7 @@ class YoutubeDL(object): else: message = self._bidi_workaround(message) output = message + '\n' - write_string(output, self._err_file) + self._write_string(output, self._err_file) def to_console_title(self, message): if not self.params.get('consoletitle', False): @@ -309,21 +319,21 @@ class YoutubeDL(object): # already of type unicode() ctypes.windll.kernel32.SetConsoleTitleW(ctypes.c_wchar_p(message)) elif 'TERM' in os.environ: - write_string('\033]0;%s\007' % message, self._screen_file) + self._write_string('\033]0;%s\007' % message, self._screen_file) def save_console_title(self): if not self.params.get('consoletitle', False): return if 'TERM' in os.environ: # Save the title on stack - write_string('\033[22;0t', self._screen_file) + self._write_string('\033[22;0t', self._screen_file) def restore_console_title(self): if not self.params.get('consoletitle', False): return if 'TERM' in os.environ: # Restore the title from stack - write_string('\033[23;0t', self._screen_file) + self._write_string('\033[23;0t', self._screen_file) def __enter__(self): self.save_console_title() @@ -370,12 +380,17 @@ class YoutubeDL(object): Print the message to stderr, it will be prefixed with 'WARNING:' If stderr is a tty file the 'WARNING:' will be colored ''' - if self._err_file.isatty() and os.name != 'nt': - _msg_header = '\033[0;33mWARNING:\033[0m' + if self.params.get('logger') is not None: + self.params['logger'].warning(message) else: - _msg_header = 'WARNING:' - warning_message = '%s %s' % (_msg_header, message) - self.to_stderr(warning_message) + if self.params.get('no_warnings'): + return + if self._err_file.isatty() and os.name != 'nt': + _msg_header = '\033[0;33mWARNING:\033[0m' + else: + _msg_header = 'WARNING:' + warning_message = '%s %s' % (_msg_header, message) + self.to_stderr(warning_message) def report_error(self, message, tb=None): ''' @@ -409,6 +424,13 @@ class YoutubeDL(object): template_dict['autonumber'] = autonumber_templ % self._num_downloads if template_dict.get('playlist_index') is not None: template_dict['playlist_index'] = '%05d' % template_dict['playlist_index'] + if template_dict.get('resolution') is None: + if template_dict.get('width') and template_dict.get('height'): + template_dict['resolution'] = '%dx%d' % (template_dict['width'], template_dict['height']) + elif template_dict.get('height'): + template_dict['resolution'] = '%sp' % template_dict['height'] + elif template_dict.get('width'): + template_dict['resolution'] = '?x%d' % template_dict['width'] sanitize = lambda k, v: sanitize_filename( compat_str(v), @@ -419,7 +441,8 @@ class YoutubeDL(object): if v is not None) template_dict = collections.defaultdict(lambda: 'NA', template_dict) - tmpl = os.path.expanduser(self.params['outtmpl']) + outtmpl = self.params.get('outtmpl', DEFAULT_OUTTMPL) + tmpl = os.path.expanduser(outtmpl) filename = tmpl % template_dict return filename except ValueError as err: @@ -499,13 +522,7 @@ class YoutubeDL(object): '_type': 'compat_list', 'entries': ie_result, } - self.add_extra_info(ie_result, - { - 'extractor': ie.IE_NAME, - 'webpage_url': url, - 'webpage_url_basename': url_basename(url), - 'extractor_key': ie.ie_key(), - }) + self.add_default_extra_info(ie_result, ie, url) if process: return self.process_ie_result(ie_result, download, extra_info) else: @@ -522,7 +539,15 @@ class YoutubeDL(object): else: raise else: - self.report_error('no suitable InfoExtractor: %s' % url) + self.report_error('no suitable InfoExtractor for URL %s' % url) + + def add_default_extra_info(self, ie_result, ie, url): + self.add_extra_info(ie_result, { + 'extractor': ie.IE_NAME, + 'webpage_url': url, + 'webpage_url_basename': url_basename(url), + 'extractor_key': ie.ie_key(), + }) def process_ie_result(self, ie_result, download=True, extra_info={}): """ @@ -656,6 +681,18 @@ class YoutubeDL(object): if f.get('vcodec') == 'none'] if audio_formats: return audio_formats[0] + elif format_spec == 'bestvideo': + video_formats = [ + f for f in available_formats + if f.get('acodec') == 'none'] + if video_formats: + return video_formats[-1] + elif format_spec == 'worstvideo': + video_formats = [ + f for f in available_formats + if f.get('acodec') == 'none'] + if video_formats: + return video_formats[0] else: extensions = ['mp4', 'flv', 'webm', '3gp'] if format_spec in extensions: @@ -670,11 +707,35 @@ class YoutubeDL(object): def process_video_result(self, info_dict, download=True): assert info_dict.get('_type', 'video') == 'video' + if 'id' not in info_dict: + raise ExtractorError('Missing "id" field in extractor result') + if 'title' not in info_dict: + raise ExtractorError('Missing "title" field in extractor result') + if 'playlist' not in info_dict: # It isn't part of a playlist info_dict['playlist'] = None info_dict['playlist_index'] = None + thumbnails = info_dict.get('thumbnails') + if thumbnails: + thumbnails.sort(key=lambda t: ( + t.get('width'), t.get('height'), t.get('url'))) + for t in thumbnails: + if 'width' in t and 'height' in t: + t['resolution'] = '%dx%d' % (t['width'], t['height']) + + if thumbnails and 'thumbnail' not in info_dict: + info_dict['thumbnail'] = thumbnails[-1]['url'] + + if 'display_id' not in info_dict and 'id' in info_dict: + info_dict['display_id'] = info_dict['id'] + + if info_dict.get('upload_date') is None and info_dict.get('timestamp') is not None: + upload_date = datetime.datetime.utcfromtimestamp( + info_dict['timestamp']) + info_dict['upload_date'] = upload_date.strftime('%Y%m%d') + # This extractors handle format selection themselves if info_dict['extractor'] in ['Youku']: if download: @@ -688,8 +749,14 @@ class YoutubeDL(object): else: formats = info_dict['formats'] + if not formats: + raise ExtractorError('No video formats found!') + # We check that all the formats have the format and format_id fields - for (i, format) in enumerate(formats): + for i, format in enumerate(formats): + if 'url' not in format: + raise ExtractorError('Missing "url" key in result (index %d)' % i) + if format.get('format_id') is None: format['format_id'] = compat_str(i) if format.get('format') is None: @@ -700,7 +767,7 @@ class YoutubeDL(object): ) # Automatically determine file extension if missing if 'ext' not in format: - format['ext'] = determine_ext(format['url']) + format['ext'] = determine_ext(format['url']).lower() format_limit = self.params.get('format_limit', None) if format_limit: @@ -825,7 +892,7 @@ class YoutubeDL(object): try: dn = os.path.dirname(encodeFilename(filename)) - if dn != '' and not os.path.exists(dn): + if dn and not os.path.exists(dn): os.makedirs(dn) except (OSError, IOError) as err: self.report_error('unable to create directory ' + compat_str(err)) @@ -882,7 +949,7 @@ class YoutubeDL(object): with io.open(encodeFilename(sub_filename), 'w', encoding='utf-8') as subfile: subfile.write(sub) except (OSError, IOError): - self.report_error('Cannot write subtitles file ' + descfn) + self.report_error('Cannot write subtitles file ' + sub_filename) return if self.params.get('writeinfojson', False): @@ -908,7 +975,7 @@ class YoutubeDL(object): self.to_screen('[%s] %s: Downloading thumbnail ...' % (info_dict['extractor'], info_dict['id'])) try: - uf = compat_urllib_request.urlopen(info_dict['thumbnail']) + uf = self.urlopen(info_dict['thumbnail']) with open(thumb_filename, 'wb') as thumbf: shutil.copyfileobj(uf, thumbf) self.to_screen('[%s] %s: Writing thumbnail to: %s' % @@ -926,6 +993,8 @@ class YoutubeDL(object): fd = get_suitable_downloader(info)(self, self.params) for ph in self._progress_hooks: fd.add_progress_hook(ph) + if self.params.get('verbose'): + self.to_stdout('[debug] Invoking downloader on %r' % info.get('url')) return fd.download(name, info) if info_dict.get('requested_formats') is not None: downloaded = [] @@ -971,10 +1040,11 @@ 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 - '%' not in self.params['outtmpl'] + '%' not in outtmpl and self.params.get('max_downloads') != 1): - raise SameFileError(self.params['outtmpl']) + raise SameFileError(outtmpl) for url in url_list: try: @@ -1085,57 +1155,57 @@ class YoutubeDL(object): res = default return res - def list_formats(self, info_dict): - def format_note(fdict): - res = '' - if fdict.get('ext') in ['f4f', 'f4m']: - res += '(unsupported) ' - if fdict.get('format_note') is not None: - res += fdict['format_note'] + ' ' - if fdict.get('tbr') is not None: - res += '%4dk ' % fdict['tbr'] - if fdict.get('container') is not None: - if res: - res += ', ' - res += '%s container' % fdict['container'] - if (fdict.get('vcodec') is not None and - fdict.get('vcodec') != 'none'): - if res: - res += ', ' - res += fdict['vcodec'] - if fdict.get('vbr') is not None: - res += '@' - elif fdict.get('vbr') is not None and fdict.get('abr') is not None: - res += 'video@' + def _format_note(self, fdict): + res = '' + if fdict.get('ext') in ['f4f', 'f4m']: + res += '(unsupported) ' + if fdict.get('format_note') is not None: + res += fdict['format_note'] + ' ' + if fdict.get('tbr') is not None: + res += '%4dk ' % fdict['tbr'] + if fdict.get('container') is not None: + if res: + res += ', ' + res += '%s container' % fdict['container'] + if (fdict.get('vcodec') is not None and + fdict.get('vcodec') != 'none'): + if res: + res += ', ' + res += fdict['vcodec'] if fdict.get('vbr') is not None: - res += '%4dk' % fdict['vbr'] - if fdict.get('acodec') is not None: - if res: - res += ', ' - if fdict['acodec'] == 'none': - res += 'video only' - else: - res += '%-5s' % fdict['acodec'] - elif fdict.get('abr') is not None: - if res: - res += ', ' - res += 'audio' - if fdict.get('abr') is not None: - res += '@%3dk' % fdict['abr'] - if fdict.get('asr') is not None: - res += ' (%5dHz)' % fdict['asr'] - if fdict.get('filesize') is not None: - if res: - res += ', ' - res += format_bytes(fdict['filesize']) - return res + res += '@' + elif fdict.get('vbr') is not None and fdict.get('abr') is not None: + res += 'video@' + if fdict.get('vbr') is not None: + res += '%4dk' % fdict['vbr'] + if fdict.get('acodec') is not None: + if res: + res += ', ' + if fdict['acodec'] == 'none': + res += 'video only' + else: + res += '%-5s' % fdict['acodec'] + elif fdict.get('abr') is not None: + if res: + res += ', ' + res += 'audio' + if fdict.get('abr') is not None: + res += '@%3dk' % fdict['abr'] + if fdict.get('asr') is not None: + res += ' (%5dHz)' % fdict['asr'] + if fdict.get('filesize') is not None: + if res: + res += ', ' + res += format_bytes(fdict['filesize']) + return res + def list_formats(self, info_dict): def line(format, idlen=20): return (('%-' + compat_str(idlen + 1) + 's%-10s%-12s%s') % ( format['format_id'], format['ext'], self.format_resolution(format), - format_note(format), + self._format_note(format), )) formats = info_dict.get('formats', [info_dict]) @@ -1143,8 +1213,8 @@ class YoutubeDL(object): max(len(f['format_id']) for f in formats)) formats_s = [line(f, idlen) for f in formats] if len(formats) > 1: - formats_s[0] += (' ' if format_note(formats[0]) else '') + '(worst)' - formats_s[-1] += (' ' if format_note(formats[-1]) else '') + '(best)' + formats_s[0] += (' ' if self._format_note(formats[0]) else '') + '(worst)' + formats_s[-1] += (' ' if self._format_note(formats[-1]) else '') + '(best)' header_line = line({ 'format_id': 'format code', 'ext': 'extension', @@ -1154,12 +1224,22 @@ class YoutubeDL(object): def urlopen(self, req): """ Start an HTTP download """ - return self._opener.open(req) + return self._opener.open(req, timeout=self._socket_timeout) def print_debug_header(self): if not self.params.get('verbose'): return - write_string('[debug] youtube-dl version ' + __version__ + '\n') + + write_string( + '[debug] Encodings: locale %s, fs %s, out %s, pref %s\n' % ( + locale.getpreferredencoding(), + sys.getfilesystemencoding(), + sys.stdout.encoding, + self.get_encoding()), + encoding=None + ) + + self._write_string('[debug] youtube-dl version ' + __version__ + '\n') try: sp = subprocess.Popen( ['git', 'rev-parse', '--short', 'HEAD'], @@ -1168,24 +1248,24 @@ class YoutubeDL(object): out, err = sp.communicate() out = out.decode().strip() if re.match('[0-9a-f]+', out): - write_string('[debug] Git HEAD: ' + out + '\n') + self._write_string('[debug] Git HEAD: ' + out + '\n') except: try: sys.exc_clear() except: pass - write_string('[debug] Python version %s - %s' % + self._write_string('[debug] Python version %s - %s' % (platform.python_version(), platform_name()) + '\n') proxy_map = {} for handler in self._opener.handlers: if hasattr(handler, 'proxies'): proxy_map.update(handler.proxies) - write_string('[debug] Proxy map: ' + compat_str(proxy_map) + '\n') + self._write_string('[debug] Proxy map: ' + compat_str(proxy_map) + '\n') def _setup_opener(self): timeout_val = self.params.get('socket_timeout') - timeout = 600 if timeout_val is None else float(timeout_val) + self._socket_timeout = 600 if timeout_val is None else float(timeout_val) opts_cookiefile = self.params.get('cookiefile') opts_proxy = self.params.get('proxy') @@ -1224,6 +1304,18 @@ class YoutubeDL(object): opener.addheaders = [] self._opener = opener - # TODO remove this global modification - compat_urllib_request.install_opener(opener) - socket.setdefaulttimeout(timeout) + def encode(self, s): + if isinstance(s, bytes): + return s # Already encoded + + try: + return s.encode(self.get_encoding()) + except UnicodeEncodeError as err: + err.reason = err.reason + '. Check your system encoding configuration or use the --encoding option.' + raise + + def get_encoding(self): + encoding = self.params.get('encoding') + if encoding is None: + encoding = preferredencoding() + return encoding