Raphaël G. Git Repositories - youtubedl/blob - youtube_dl/utils.py

   1 #!/usr/bin/env python
   2 # -*- coding: utf-8 -*-
   3
   4 from __future__ import unicode_literals
   5
   6 import calendar
   7 import codecs
   8 import contextlib
   9 import ctypes
  10 import datetime
  11 import email.utils
  12 import errno
  13 import gzip
  14 import itertools
  15 import io
  16 import json
  17 import locale
  18 import math
  19 import os
  20 import pipes
  21 import platform
  22 import re
  23 import ssl
  24 import socket
  25 import struct
  26 import subprocess
  27 import sys
  28 import tempfile
  29 import traceback
  30 import xml.etree.ElementTree
  31 import zlib
  32
  33 from .compat import (
  34     compat_chr,
  35     compat_getenv,
  36     compat_html_entities,
  37     compat_parse_qs,
  38     compat_str,
  39     compat_urllib_error,
  40     compat_urllib_parse,
  41     compat_urllib_parse_urlparse,
  42     compat_urllib_request,
  43     compat_urlparse,
  44     shlex_quote,
  45 )
  46
  47
  48 # This is not clearly defined otherwise
  49 compiled_regex_type = type(re.compile(''))
  50
  51 std_headers = {
  52     'User-Agent': 'Mozilla/5.0 (X11; Linux x86_64; rv:10.0) Gecko/20100101 Firefox/10.0 (Chrome)',
  53     'Accept-Charset': 'ISO-8859-1,utf-8;q=0.7,*;q=0.7',
  54     'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
  55     'Accept-Encoding': 'gzip, deflate',
  56     'Accept-Language': 'en-us,en;q=0.5',
  57 }
  58
  59
  60 def preferredencoding():
  61     """Get preferred encoding.
  62
  63     Returns the best encoding scheme for the system, based on
  64     locale.getpreferredencoding() and some further tweaks.
  65     """
  66     try:
  67         pref = locale.getpreferredencoding()
  68         'TEST'.encode(pref)
  69     except:
  70         pref = 'UTF-8'
  71
  72     return pref
  73
  74
  75 def write_json_file(obj, fn):
  76     """ Encode obj as JSON and write it to fn, atomically if possible """
  77
  78     fn = encodeFilename(fn)
  79     if sys.version_info < (3, 0) and sys.platform != 'win32':
  80         encoding = get_filesystem_encoding()
  81         # os.path.basename returns a bytes object, but NamedTemporaryFile
  82         # will fail if the filename contains non ascii characters unless we
  83         # use a unicode object
  84         path_basename = lambda f: os.path.basename(fn).decode(encoding)
  85         # the same for os.path.dirname
  86         path_dirname = lambda f: os.path.dirname(fn).decode(encoding)
  87     else:
  88         path_basename = os.path.basename
  89         path_dirname = os.path.dirname
  90
  91     args = {
  92         'suffix': '.tmp',
  93         'prefix': path_basename(fn) + '.',
  94         'dir': path_dirname(fn),
  95         'delete': False,
  96     }
  97
  98     # In Python 2.x, json.dump expects a bytestream.
  99     # In Python 3.x, it writes to a character stream
 100     if sys.version_info < (3, 0):
 101         args['mode'] = 'wb'
 102     else:
 103         args.update({
 104             'mode': 'w',
 105             'encoding': 'utf-8',
 106         })
 107
 108     tf = tempfile.NamedTemporaryFile(**args)
 109
 110     try:
 111         with tf:
 112             json.dump(obj, tf)
 113         if sys.platform == 'win32':
 114             # Need to remove existing file on Windows, else os.rename raises
 115             # WindowsError or FileExistsError.
 116             try:
 117                 os.unlink(fn)
 118             except OSError:
 119                 pass
 120         os.rename(tf.name, fn)
 121     except:
 122         try:
 123             os.remove(tf.name)
 124         except OSError:
 125             pass
 126         raise
 127
 128
 129 if sys.version_info >= (2, 7):
 130     def find_xpath_attr(node, xpath, key, val):
 131         """ Find the xpath xpath[@key=val] """
 132         assert re.match(r'^[a-zA-Z-]+$', key)
 133         assert re.match(r'^[a-zA-Z0-9@\s:._-]*$', val)
 134         expr = xpath + "[@%s='%s']" % (key, val)
 135         return node.find(expr)
 136 else:
 137     def find_xpath_attr(node, xpath, key, val):
 138         # Here comes the crazy part: In 2.6, if the xpath is a unicode,
 139         # .//node does not match if a node is a direct child of . !
 140         if isinstance(xpath, unicode):
 141             xpath = xpath.encode('ascii')
 142
 143         for f in node.findall(xpath):
 144             if f.attrib.get(key) == val:
 145                 return f
 146         return None
 147
 148 # On python2.6 the xml.etree.ElementTree.Element methods don't support
 149 # the namespace parameter
 150
 151
 152 def xpath_with_ns(path, ns_map):
 153     components = [c.split(':') for c in path.split('/')]
 154     replaced = []
 155     for c in components:
 156         if len(c) == 1:
 157             replaced.append(c[0])
 158         else:
 159             ns, tag = c
 160             replaced.append('{%s}%s' % (ns_map[ns], tag))
 161     return '/'.join(replaced)
 162
 163
 164 def xpath_text(node, xpath, name=None, fatal=False):
 165     if sys.version_info < (2, 7):  # Crazy 2.6
 166         xpath = xpath.encode('ascii')
 167
 168     n = node.find(xpath)
 169     if n is None:
 170         if fatal:
 171             name = xpath if name is None else name
 172             raise ExtractorError('Could not find XML element %s' % name)
 173         else:
 174             return None
 175     return n.text
 176
 177
 178 def get_element_by_id(id, html):
 179     """Return the content of the tag with the specified ID in the passed HTML document"""
 180     return get_element_by_attribute("id", id, html)
 181
 182
 183 def get_element_by_attribute(attribute, value, html):
 184     """Return the content of the tag with the specified attribute in the passed HTML document"""
 185
 186     m = re.search(r'''(?xs)
 187         <([a-zA-Z0-9:._-]+)
 188          (?:\s+[a-zA-Z0-9:._-]+(?:=[a-zA-Z0-9:._-]+|="[^"]+"|='[^']+'))*?
 189          \s+%s=['"]?%s['"]?
 190          (?:\s+[a-zA-Z0-9:._-]+(?:=[a-zA-Z0-9:._-]+|="[^"]+"|='[^']+'))*?
 191         \s*>
 192         (?P<content>.*?)
 193         </\1>
 194     ''' % (re.escape(attribute), re.escape(value)), html)
 195
 196     if not m:
 197         return None
 198     res = m.group('content')
 199
 200     if res.startswith('"') or res.startswith("'"):
 201         res = res[1:-1]
 202
 203     return unescapeHTML(res)
 204
 205
 206 def clean_html(html):
 207     """Clean an HTML snippet into a readable string"""
 208     # Newline vs <br />
 209     html = html.replace('\n', ' ')
 210     html = re.sub(r'\s*<\s*br\s*/?\s*>\s*', '\n', html)
 211     html = re.sub(r'<\s*/\s*p\s*>\s*<\s*p[^>]*>', '\n', html)
 212     # Strip html tags
 213     html = re.sub('<.*?>', '', html)
 214     # Replace html entities
 215     html = unescapeHTML(html)
 216     return html.strip()
 217
 218
 219 def sanitize_open(filename, open_mode):
 220     """Try to open the given filename, and slightly tweak it if this fails.
 221
 222     Attempts to open the given filename. If this fails, it tries to change
 223     the filename slightly, step by step, until it's either able to open it
 224     or it fails and raises a final exception, like the standard open()
 225     function.
 226
 227     It returns the tuple (stream, definitive_file_name).
 228     """
 229     try:
 230         if filename == '-':
 231             if sys.platform == 'win32':
 232                 import msvcrt
 233                 msvcrt.setmode(sys.stdout.fileno(), os.O_BINARY)
 234             return (sys.stdout.buffer if hasattr(sys.stdout, 'buffer') else sys.stdout, filename)
 235         stream = open(encodeFilename(filename), open_mode)
 236         return (stream, filename)
 237     except (IOError, OSError) as err:
 238         if err.errno in (errno.EACCES,):
 239             raise
 240
 241         # In case of error, try to remove win32 forbidden chars
 242         alt_filename = os.path.join(
 243             re.sub('[/<>:"\\|\\\\?\\*]', '#', path_part)
 244             for path_part in os.path.split(filename)
 245         )
 246         if alt_filename == filename:
 247             raise
 248         else:
 249             # An exception here should be caught in the caller
 250             stream = open(encodeFilename(filename), open_mode)
 251             return (stream, alt_filename)
 252
 253
 254 def timeconvert(timestr):
 255     """Convert RFC 2822 defined time string into system timestamp"""
 256     timestamp = None
 257     timetuple = email.utils.parsedate_tz(timestr)
 258     if timetuple is not None:
 259         timestamp = email.utils.mktime_tz(timetuple)
 260     return timestamp
 261
 262
 263 def sanitize_filename(s, restricted=False, is_id=False):
 264     """Sanitizes a string so it could be used as part of a filename.
 265     If restricted is set, use a stricter subset of allowed characters.
 266     Set is_id if this is not an arbitrary string, but an ID that should be kept if possible
 267     """
 268     def replace_insane(char):
 269         if char == '?' or ord(char) < 32 or ord(char) == 127:
 270             return ''
 271         elif char == '"':
 272             return '' if restricted else '\''
 273         elif char == ':':
 274             return '_-' if restricted else ' -'
 275         elif char in '\\/|*<>':
 276             return '_'
 277         if restricted and (char in '!&\'()[]{}$;`^,#' or char.isspace()):
 278             return '_'
 279         if restricted and ord(char) > 127:
 280             return '_'
 281         return char
 282
 283     result = ''.join(map(replace_insane, s))
 284     if not is_id:
 285         while '__' in result:
 286             result = result.replace('__', '_')
 287         result = result.strip('_')
 288         # Common case of "Foreign band name - English song title"
 289         if restricted and result.startswith('-_'):
 290             result = result[2:]
 291         if not result:
 292             result = '_'
 293     return result
 294
 295
 296 def orderedSet(iterable):
 297     """ Remove all duplicates from the input iterable """
 298     res = []
 299     for el in iterable:
 300         if el not in res:
 301             res.append(el)
 302     return res
 303
 304
 305 def _htmlentity_transform(entity):
 306     """Transforms an HTML entity to a character."""
 307     # Known non-numeric HTML entity
 308     if entity in compat_html_entities.name2codepoint:
 309         return compat_chr(compat_html_entities.name2codepoint[entity])
 310
 311     mobj = re.match(r'#(x?[0-9]+)', entity)
 312     if mobj is not None:
 313         numstr = mobj.group(1)
 314         if numstr.startswith('x'):
 315             base = 16
 316             numstr = '0%s' % numstr
 317         else:
 318             base = 10
 319         return compat_chr(int(numstr, base))
 320
 321     # Unknown entity in name, return its literal representation
 322     return ('&%s;' % entity)
 323
 324
 325 def unescapeHTML(s):
 326     if s is None:
 327         return None
 328     assert type(s) == compat_str
 329
 330     return re.sub(
 331         r'&([^;]+);', lambda m: _htmlentity_transform(m.group(1)), s)
 332
 333
 334 def encodeFilename(s, for_subprocess=False):
 335     """
 336     @param s The name of the file
 337     """
 338
 339     assert type(s) == compat_str
 340
 341     # Python 3 has a Unicode API
 342     if sys.version_info >= (3, 0):
 343         return s
 344
 345     if sys.platform == 'win32' and sys.getwindowsversion()[0] >= 5:
 346         # Pass '' directly to use Unicode APIs on Windows 2000 and up
 347         # (Detecting Windows NT 4 is tricky because 'major >= 4' would
 348         # match Windows 9x series as well. Besides, NT 4 is obsolete.)
 349         if not for_subprocess:
 350             return s
 351         else:
 352             # For subprocess calls, encode with locale encoding
 353             # Refer to http://stackoverflow.com/a/9951851/35070
 354             encoding = preferredencoding()
 355     else:
 356         encoding = sys.getfilesystemencoding()
 357     if encoding is None:
 358         encoding = 'utf-8'
 359     return s.encode(encoding, 'ignore')
 360
 361
 362 def encodeArgument(s):
 363     if not isinstance(s, compat_str):
 364         # Legacy code that uses byte strings
 365         # Uncomment the following line after fixing all post processors
 366         #assert False, 'Internal error: %r should be of type %r, is %r' % (s, compat_str, type(s))
 367         s = s.decode('ascii')
 368     return encodeFilename(s, True)
 369
 370
 371 def decodeOption(optval):
 372     if optval is None:
 373         return optval
 374     if isinstance(optval, bytes):
 375         optval = optval.decode(preferredencoding())
 376
 377     assert isinstance(optval, compat_str)
 378     return optval
 379
 380
 381 def formatSeconds(secs):
 382     if secs > 3600:
 383         return '%d:%02d:%02d' % (secs // 3600, (secs % 3600) // 60, secs % 60)
 384     elif secs > 60:
 385         return '%d:%02d' % (secs // 60, secs % 60)
 386     else:
 387         return '%d' % secs
 388
 389
 390 def make_HTTPS_handler(opts_no_check_certificate, **kwargs):
 391     if sys.version_info < (3, 2):
 392         import httplib
 393
 394         class HTTPSConnectionV3(httplib.HTTPSConnection):
 395             def __init__(self, *args, **kwargs):
 396                 httplib.HTTPSConnection.__init__(self, *args, **kwargs)
 397
 398             def connect(self):
 399                 sock = socket.create_connection((self.host, self.port), self.timeout)
 400                 if getattr(self, '_tunnel_host', False):
 401                     self.sock = sock
 402                     self._tunnel()
 403                 try:
 404                     self.sock = ssl.wrap_socket(sock, self.key_file, self.cert_file, ssl_version=ssl.PROTOCOL_TLSv1)
 405                 except ssl.SSLError:
 406                     self.sock = ssl.wrap_socket(sock, self.key_file, self.cert_file, ssl_version=ssl.PROTOCOL_SSLv23)
 407
 408         class HTTPSHandlerV3(compat_urllib_request.HTTPSHandler):
 409             def https_open(self, req):
 410                 return self.do_open(HTTPSConnectionV3, req)
 411         return HTTPSHandlerV3(**kwargs)
 412     elif hasattr(ssl, 'create_default_context'):  # Python >= 3.4
 413         context = ssl.create_default_context(ssl.Purpose.CLIENT_AUTH)
 414         context.options &= ~ssl.OP_NO_SSLv3  # Allow older, not-as-secure SSLv3
 415         if opts_no_check_certificate:
 416             context.verify_mode = ssl.CERT_NONE
 417         return compat_urllib_request.HTTPSHandler(context=context, **kwargs)
 418     else:  # Python < 3.4
 419         context = ssl.SSLContext(ssl.PROTOCOL_SSLv23)
 420         context.verify_mode = (ssl.CERT_NONE
 421                                if opts_no_check_certificate
 422                                else ssl.CERT_REQUIRED)
 423         context.set_default_verify_paths()
 424         try:
 425             context.load_default_certs()
 426         except AttributeError:
 427             pass  # Python < 3.4
 428         return compat_urllib_request.HTTPSHandler(context=context, **kwargs)
 429
 430
 431 class ExtractorError(Exception):
 432     """Error during info extraction."""
 433
 434     def __init__(self, msg, tb=None, expected=False, cause=None, video_id=None):
 435         """ tb, if given, is the original traceback (so that it can be printed out).
 436         If expected is set, this is a normal error message and most likely not a bug in youtube-dl.
 437         """
 438
 439         if sys.exc_info()[0] in (compat_urllib_error.URLError, socket.timeout, UnavailableVideoError):
 440             expected = True
 441         if video_id is not None:
 442             msg = video_id + ': ' + msg
 443         if cause:
 444             msg += ' (caused by %r)' % cause
 445         if not expected:
 446             if ytdl_is_updateable():
 447                 update_cmd = 'type  youtube-dl -U  to update'
 448             else:
 449                 update_cmd = 'see  https://yt-dl.org/update  on how to update'
 450             msg += '; please report this issue on https://yt-dl.org/bug .'
 451             msg += ' Make sure you are using the latest version; %s.' % update_cmd
 452             msg += ' Be sure to call youtube-dl with the --verbose flag and include its complete output.'
 453         super(ExtractorError, self).__init__(msg)
 454
 455         self.traceback = tb
 456         self.exc_info = sys.exc_info()  # preserve original exception
 457         self.cause = cause
 458         self.video_id = video_id
 459
 460     def format_traceback(self):
 461         if self.traceback is None:
 462             return None
 463         return ''.join(traceback.format_tb(self.traceback))
 464
 465
 466 class RegexNotFoundError(ExtractorError):
 467     """Error when a regex didn't match"""
 468     pass
 469
 470
 471 class DownloadError(Exception):
 472     """Download Error exception.
 473
 474     This exception may be thrown by FileDownloader objects if they are not
 475     configured to continue on errors. They will contain the appropriate
 476     error message.
 477     """
 478
 479     def __init__(self, msg, exc_info=None):
 480         """ exc_info, if given, is the original exception that caused the trouble (as returned by sys.exc_info()). """
 481         super(DownloadError, self).__init__(msg)
 482         self.exc_info = exc_info
 483
 484
 485 class SameFileError(Exception):
 486     """Same File exception.
 487
 488     This exception will be thrown by FileDownloader objects if they detect
 489     multiple files would have to be downloaded to the same file on disk.
 490     """
 491     pass
 492
 493
 494 class PostProcessingError(Exception):
 495     """Post Processing exception.
 496
 497     This exception may be raised by PostProcessor's .run() method to
 498     indicate an error in the postprocessing task.
 499     """
 500
 501     def __init__(self, msg):
 502         self.msg = msg
 503
 504
 505 class MaxDownloadsReached(Exception):
 506     """ --max-downloads limit has been reached. """
 507     pass
 508
 509
 510 class UnavailableVideoError(Exception):
 511     """Unavailable Format exception.
 512
 513     This exception will be thrown when a video is requested
 514     in a format that is not available for that video.
 515     """
 516     pass
 517
 518
 519 class ContentTooShortError(Exception):
 520     """Content Too Short exception.
 521
 522     This exception may be raised by FileDownloader objects when a file they
 523     download is too small for what the server announced first, indicating
 524     the connection was probably interrupted.
 525     """
 526     # Both in bytes
 527     downloaded = None
 528     expected = None
 529
 530     def __init__(self, downloaded, expected):
 531         self.downloaded = downloaded
 532         self.expected = expected
 533
 534
 535 class YoutubeDLHandler(compat_urllib_request.HTTPHandler):
 536     """Handler for HTTP requests and responses.
 537
 538     This class, when installed with an OpenerDirector, automatically adds
 539     the standard headers to every HTTP request and handles gzipped and
 540     deflated responses from web servers. If compression is to be avoided in
 541     a particular request, the original request in the program code only has
 542     to include the HTTP header "Youtubedl-No-Compression", which will be
 543     removed before making the real request.
 544
 545     Part of this code was copied from:
 546
 547     http://techknack.net/python-urllib2-handlers/
 548
 549     Andrew Rowls, the author of that code, agreed to release it to the
 550     public domain.
 551     """
 552
 553     @staticmethod
 554     def deflate(data):
 555         try:
 556             return zlib.decompress(data, -zlib.MAX_WBITS)
 557         except zlib.error:
 558             return zlib.decompress(data)
 559
 560     @staticmethod
 561     def addinfourl_wrapper(stream, headers, url, code):
 562         if hasattr(compat_urllib_request.addinfourl, 'getcode'):
 563             return compat_urllib_request.addinfourl(stream, headers, url, code)
 564         ret = compat_urllib_request.addinfourl(stream, headers, url)
 565         ret.code = code
 566         return ret
 567
 568     def http_request(self, req):
 569         for h, v in std_headers.items():
 570             if h not in req.headers:
 571                 req.add_header(h, v)
 572         if 'Youtubedl-no-compression' in req.headers:
 573             if 'Accept-encoding' in req.headers:
 574                 del req.headers['Accept-encoding']
 575             del req.headers['Youtubedl-no-compression']
 576         if 'Youtubedl-user-agent' in req.headers:
 577             if 'User-agent' in req.headers:
 578                 del req.headers['User-agent']
 579             req.headers['User-agent'] = req.headers['Youtubedl-user-agent']
 580             del req.headers['Youtubedl-user-agent']
 581
 582         if sys.version_info < (2, 7) and '#' in req.get_full_url():
 583             # Python 2.6 is brain-dead when it comes to fragments
 584             req._Request__original = req._Request__original.partition('#')[0]
 585             req._Request__r_type = req._Request__r_type.partition('#')[0]
 586
 587         return req
 588
 589     def http_response(self, req, resp):
 590         old_resp = resp
 591         # gzip
 592         if resp.headers.get('Content-encoding', '') == 'gzip':
 593             content = resp.read()
 594             gz = gzip.GzipFile(fileobj=io.BytesIO(content), mode='rb')
 595             try:
 596                 uncompressed = io.BytesIO(gz.read())
 597             except IOError as original_ioerror:
 598                 # There may be junk add the end of the file
 599                 # See http://stackoverflow.com/q/4928560/35070 for details
 600                 for i in range(1, 1024):
 601                     try:
 602                         gz = gzip.GzipFile(fileobj=io.BytesIO(content[:-i]), mode='rb')
 603                         uncompressed = io.BytesIO(gz.read())
 604                     except IOError:
 605                         continue
 606                     break
 607                 else:
 608                     raise original_ioerror
 609             resp = self.addinfourl_wrapper(uncompressed, old_resp.headers, old_resp.url, old_resp.code)
 610             resp.msg = old_resp.msg
 611         # deflate
 612         if resp.headers.get('Content-encoding', '') == 'deflate':
 613             gz = io.BytesIO(self.deflate(resp.read()))
 614             resp = self.addinfourl_wrapper(gz, old_resp.headers, old_resp.url, old_resp.code)
 615             resp.msg = old_resp.msg
 616         return resp
 617
 618     https_request = http_request
 619     https_response = http_response
 620
 621
 622 def parse_iso8601(date_str, delimiter='T'):
 623     """ Return a UNIX timestamp from the given date """
 624
 625     if date_str is None:
 626         return None
 627
 628     m = re.search(
 629         r'(\.[0-9]+)?(?:Z$| ?(?P<sign>\+|-)(?P<hours>[0-9]{2}):?(?P<minutes>[0-9]{2})$)',
 630         date_str)
 631     if not m:
 632         timezone = datetime.timedelta()
 633     else:
 634         date_str = date_str[:-len(m.group(0))]
 635         if not m.group('sign'):
 636             timezone = datetime.timedelta()
 637         else:
 638             sign = 1 if m.group('sign') == '+' else -1
 639             timezone = datetime.timedelta(
 640                 hours=sign * int(m.group('hours')),
 641                 minutes=sign * int(m.group('minutes')))
 642     date_format = '%Y-%m-%d{0}%H:%M:%S'.format(delimiter)
 643     dt = datetime.datetime.strptime(date_str, date_format) - timezone
 644     return calendar.timegm(dt.timetuple())
 645
 646
 647 def unified_strdate(date_str):
 648     """Return a string with the date in the format YYYYMMDD"""
 649
 650     if date_str is None:
 651         return None
 652
 653     upload_date = None
 654     # Replace commas
 655     date_str = date_str.replace(',', ' ')
 656     # %z (UTC offset) is only supported in python>=3.2
 657     date_str = re.sub(r' ?(\+|-)[0-9]{2}:?[0-9]{2}$', '', date_str)
 658     format_expressions = [
 659         '%d %B %Y',
 660         '%d %b %Y',
 661         '%B %d %Y',
 662         '%b %d %Y',
 663         '%b %dst %Y %I:%M%p',
 664         '%b %dnd %Y %I:%M%p',
 665         '%b %dth %Y %I:%M%p',
 666         '%Y-%m-%d',
 667         '%Y/%m/%d',
 668         '%d.%m.%Y',
 669         '%d/%m/%Y',
 670         '%d/%m/%y',
 671         '%Y/%m/%d %H:%M:%S',
 672         '%d/%m/%Y %H:%M:%S',
 673         '%Y-%m-%d %H:%M:%S',
 674         '%Y-%m-%d %H:%M:%S.%f',
 675         '%d.%m.%Y %H:%M',
 676         '%d.%m.%Y %H.%M',
 677         '%Y-%m-%dT%H:%M:%SZ',
 678         '%Y-%m-%dT%H:%M:%S.%fZ',
 679         '%Y-%m-%dT%H:%M:%S.%f0Z',
 680         '%Y-%m-%dT%H:%M:%S',
 681         '%Y-%m-%dT%H:%M:%S.%f',
 682         '%Y-%m-%dT%H:%M',
 683     ]
 684     for expression in format_expressions:
 685         try:
 686             upload_date = datetime.datetime.strptime(date_str, expression).strftime('%Y%m%d')
 687         except ValueError:
 688             pass
 689     if upload_date is None:
 690         timetuple = email.utils.parsedate_tz(date_str)
 691         if timetuple:
 692             upload_date = datetime.datetime(*timetuple[:6]).strftime('%Y%m%d')
 693     return upload_date
 694
 695
 696 def determine_ext(url, default_ext='unknown_video'):
 697     if url is None:
 698         return default_ext
 699     guess = url.partition('?')[0].rpartition('.')[2]
 700     if re.match(r'^[A-Za-z0-9]+$', guess):
 701         return guess
 702     else:
 703         return default_ext
 704
 705
 706 def subtitles_filename(filename, sub_lang, sub_format):
 707     return filename.rsplit('.', 1)[0] + '.' + sub_lang + '.' + sub_format
 708
 709
 710 def date_from_str(date_str):
 711     """
 712     Return a datetime object from a string in the format YYYYMMDD or
 713     (now|today)[+-][0-9](day|week|month|year)(s)?"""
 714     today = datetime.date.today()
 715     if date_str == 'now'or date_str == 'today':
 716         return today
 717     match = re.match('(now|today)(?P<sign>[+-])(?P<time>\d+)(?P<unit>day|week|month|year)(s)?', date_str)
 718     if match is not None:
 719         sign = match.group('sign')
 720         time = int(match.group('time'))
 721         if sign == '-':
 722             time = -time
 723         unit = match.group('unit')
 724         # A bad aproximation?
 725         if unit == 'month':
 726             unit = 'day'
 727             time *= 30
 728         elif unit == 'year':
 729             unit = 'day'
 730             time *= 365
 731         unit += 's'
 732         delta = datetime.timedelta(**{unit: time})
 733         return today + delta
 734     return datetime.datetime.strptime(date_str, "%Y%m%d").date()
 735
 736
 737 def hyphenate_date(date_str):
 738     """
 739     Convert a date in 'YYYYMMDD' format to 'YYYY-MM-DD' format"""
 740     match = re.match(r'^(\d\d\d\d)(\d\d)(\d\d)$', date_str)
 741     if match is not None:
 742         return '-'.join(match.groups())
 743     else:
 744         return date_str
 745
 746
 747 class DateRange(object):
 748     """Represents a time interval between two dates"""
 749
 750     def __init__(self, start=None, end=None):
 751         """start and end must be strings in the format accepted by date"""
 752         if start is not None:
 753             self.start = date_from_str(start)
 754         else:
 755             self.start = datetime.datetime.min.date()
 756         if end is not None:
 757             self.end = date_from_str(end)
 758         else:
 759             self.end = datetime.datetime.max.date()
 760         if self.start > self.end:
 761             raise ValueError('Date range: "%s" , the start date must be before the end date' % self)
 762
 763     @classmethod
 764     def day(cls, day):
 765         """Returns a range that only contains the given day"""
 766         return cls(day, day)
 767
 768     def __contains__(self, date):
 769         """Check if the date is in the range"""
 770         if not isinstance(date, datetime.date):
 771             date = date_from_str(date)
 772         return self.start <= date <= self.end
 773
 774     def __str__(self):
 775         return '%s - %s' % (self.start.isoformat(), self.end.isoformat())
 776
 777
 778 def platform_name():
 779     """ Returns the platform name as a compat_str """
 780     res = platform.platform()
 781     if isinstance(res, bytes):
 782         res = res.decode(preferredencoding())
 783
 784     assert isinstance(res, compat_str)
 785     return res
 786
 787
 788 def _windows_write_string(s, out):
 789     """ Returns True if the string was written using special methods,
 790     False if it has yet to be written out."""
 791     # Adapted from http://stackoverflow.com/a/3259271/35070
 792
 793     import ctypes
 794     import ctypes.wintypes
 795
 796     WIN_OUTPUT_IDS = {
 797         1: -11,
 798         2: -12,
 799     }
 800
 801     try:
 802         fileno = out.fileno()
 803     except AttributeError:
 804         # If the output stream doesn't have a fileno, it's virtual
 805         return False
 806     if fileno not in WIN_OUTPUT_IDS:
 807         return False
 808
 809     GetStdHandle = ctypes.WINFUNCTYPE(
 810         ctypes.wintypes.HANDLE, ctypes.wintypes.DWORD)(
 811         ("GetStdHandle", ctypes.windll.kernel32))
 812     h = GetStdHandle(WIN_OUTPUT_IDS[fileno])
 813
 814     WriteConsoleW = ctypes.WINFUNCTYPE(
 815         ctypes.wintypes.BOOL, ctypes.wintypes.HANDLE, ctypes.wintypes.LPWSTR,
 816         ctypes.wintypes.DWORD, ctypes.POINTER(ctypes.wintypes.DWORD),
 817         ctypes.wintypes.LPVOID)(("WriteConsoleW", ctypes.windll.kernel32))
 818     written = ctypes.wintypes.DWORD(0)
 819
 820     GetFileType = ctypes.WINFUNCTYPE(ctypes.wintypes.DWORD, ctypes.wintypes.DWORD)(("GetFileType", ctypes.windll.kernel32))
 821     FILE_TYPE_CHAR = 0x0002
 822     FILE_TYPE_REMOTE = 0x8000
 823     GetConsoleMode = ctypes.WINFUNCTYPE(
 824         ctypes.wintypes.BOOL, ctypes.wintypes.HANDLE,
 825         ctypes.POINTER(ctypes.wintypes.DWORD))(
 826         ("GetConsoleMode", ctypes.windll.kernel32))
 827     INVALID_HANDLE_VALUE = ctypes.wintypes.DWORD(-1).value
 828
 829     def not_a_console(handle):
 830         if handle == INVALID_HANDLE_VALUE or handle is None:
 831             return True
 832         return ((GetFileType(handle) & ~FILE_TYPE_REMOTE) != FILE_TYPE_CHAR
 833                 or GetConsoleMode(handle, ctypes.byref(ctypes.wintypes.DWORD())) == 0)
 834
 835     if not_a_console(h):
 836         return False
 837
 838     def next_nonbmp_pos(s):
 839         try:
 840             return next(i for i, c in enumerate(s) if ord(c) > 0xffff)
 841         except StopIteration:
 842             return len(s)
 843
 844     while s:
 845         count = min(next_nonbmp_pos(s), 1024)
 846
 847         ret = WriteConsoleW(
 848             h, s, count if count else 2, ctypes.byref(written), None)
 849         if ret == 0:
 850             raise OSError('Failed to write string')
 851         if not count:  # We just wrote a non-BMP character
 852             assert written.value == 2
 853             s = s[1:]
 854         else:
 855             assert written.value > 0
 856             s = s[written.value:]
 857     return True
 858
 859
 860 def write_string(s, out=None, encoding=None):
 861     if out is None:
 862         out = sys.stderr
 863     assert type(s) == compat_str
 864
 865     if sys.platform == 'win32' and encoding is None and hasattr(out, 'fileno'):
 866         if _windows_write_string(s, out):
 867             return
 868
 869     if ('b' in getattr(out, 'mode', '') or
 870             sys.version_info[0] < 3):  # Python 2 lies about mode of sys.stderr
 871         byt = s.encode(encoding or preferredencoding(), 'ignore')
 872         out.write(byt)
 873     elif hasattr(out, 'buffer'):
 874         enc = encoding or getattr(out, 'encoding', None) or preferredencoding()
 875         byt = s.encode(enc, 'ignore')
 876         out.buffer.write(byt)
 877     else:
 878         out.write(s)
 879     out.flush()
 880
 881
 882 def bytes_to_intlist(bs):
 883     if not bs:
 884         return []
 885     if isinstance(bs[0], int):  # Python 3
 886         return list(bs)
 887     else:
 888         return [ord(c) for c in bs]
 889
 890
 891 def intlist_to_bytes(xs):
 892     if not xs:
 893         return b''
 894     return struct_pack('%dB' % len(xs), *xs)
 895
 896
 897 # Cross-platform file locking
 898 if sys.platform == 'win32':
 899     import ctypes.wintypes
 900     import msvcrt
 901
 902     class OVERLAPPED(ctypes.Structure):
 903         _fields_ = [
 904             ('Internal', ctypes.wintypes.LPVOID),
 905             ('InternalHigh', ctypes.wintypes.LPVOID),
 906             ('Offset', ctypes.wintypes.DWORD),
 907             ('OffsetHigh', ctypes.wintypes.DWORD),
 908             ('hEvent', ctypes.wintypes.HANDLE),
 909         ]
 910
 911     kernel32 = ctypes.windll.kernel32
 912     LockFileEx = kernel32.LockFileEx
 913     LockFileEx.argtypes = [
 914         ctypes.wintypes.HANDLE,     # hFile
 915         ctypes.wintypes.DWORD,      # dwFlags
 916         ctypes.wintypes.DWORD,      # dwReserved
 917         ctypes.wintypes.DWORD,      # nNumberOfBytesToLockLow
 918         ctypes.wintypes.DWORD,      # nNumberOfBytesToLockHigh
 919         ctypes.POINTER(OVERLAPPED)  # Overlapped
 920     ]
 921     LockFileEx.restype = ctypes.wintypes.BOOL
 922     UnlockFileEx = kernel32.UnlockFileEx
 923     UnlockFileEx.argtypes = [
 924         ctypes.wintypes.HANDLE,     # hFile
 925         ctypes.wintypes.DWORD,      # dwReserved
 926         ctypes.wintypes.DWORD,      # nNumberOfBytesToLockLow
 927         ctypes.wintypes.DWORD,      # nNumberOfBytesToLockHigh
 928         ctypes.POINTER(OVERLAPPED)  # Overlapped
 929     ]
 930     UnlockFileEx.restype = ctypes.wintypes.BOOL
 931     whole_low = 0xffffffff
 932     whole_high = 0x7fffffff
 933
 934     def _lock_file(f, exclusive):
 935         overlapped = OVERLAPPED()
 936         overlapped.Offset = 0
 937         overlapped.OffsetHigh = 0
 938         overlapped.hEvent = 0
 939         f._lock_file_overlapped_p = ctypes.pointer(overlapped)
 940         handle = msvcrt.get_osfhandle(f.fileno())
 941         if not LockFileEx(handle, 0x2 if exclusive else 0x0, 0,
 942                           whole_low, whole_high, f._lock_file_overlapped_p):
 943             raise OSError('Locking file failed: %r' % ctypes.FormatError())
 944
 945     def _unlock_file(f):
 946         assert f._lock_file_overlapped_p
 947         handle = msvcrt.get_osfhandle(f.fileno())
 948         if not UnlockFileEx(handle, 0,
 949                             whole_low, whole_high, f._lock_file_overlapped_p):
 950             raise OSError('Unlocking file failed: %r' % ctypes.FormatError())
 951
 952 else:
 953     import fcntl
 954
 955     def _lock_file(f, exclusive):
 956         fcntl.flock(f, fcntl.LOCK_EX if exclusive else fcntl.LOCK_SH)
 957
 958     def _unlock_file(f):
 959         fcntl.flock(f, fcntl.LOCK_UN)
 960
 961
 962 class locked_file(object):
 963     def __init__(self, filename, mode, encoding=None):
 964         assert mode in ['r', 'a', 'w']
 965         self.f = io.open(filename, mode, encoding=encoding)
 966         self.mode = mode
 967
 968     def __enter__(self):
 969         exclusive = self.mode != 'r'
 970         try:
 971             _lock_file(self.f, exclusive)
 972         except IOError:
 973             self.f.close()
 974             raise
 975         return self
 976
 977     def __exit__(self, etype, value, traceback):
 978         try:
 979             _unlock_file(self.f)
 980         finally:
 981             self.f.close()
 982
 983     def __iter__(self):
 984         return iter(self.f)
 985
 986     def write(self, *args):
 987         return self.f.write(*args)
 988
 989     def read(self, *args):
 990         return self.f.read(*args)
 991
 992
 993 def get_filesystem_encoding():
 994     encoding = sys.getfilesystemencoding()
 995     return encoding if encoding is not None else 'utf-8'
 996
 997
 998 def shell_quote(args):
 999     quoted_args = []
1000     encoding = get_filesystem_encoding()
1001     for a in args:
1002         if isinstance(a, bytes):
1003             # We may get a filename encoded with 'encodeFilename'
1004             a = a.decode(encoding)
1005         quoted_args.append(pipes.quote(a))
1006     return ' '.join(quoted_args)
1007
1008
1009 def takewhile_inclusive(pred, seq):
1010     """ Like itertools.takewhile, but include the latest evaluated element
1011         (the first element so that Not pred(e)) """
1012     for e in seq:
1013         yield e
1014         if not pred(e):
1015             return
1016
1017
1018 def smuggle_url(url, data):
1019     """ Pass additional data in a URL for internal use. """
1020
1021     sdata = compat_urllib_parse.urlencode(
1022         {'__youtubedl_smuggle': json.dumps(data)})
1023     return url + '#' + sdata
1024
1025
1026 def unsmuggle_url(smug_url, default=None):
1027     if not '#__youtubedl_smuggle' in smug_url:
1028         return smug_url, default
1029     url, _, sdata = smug_url.rpartition('#')
1030     jsond = compat_parse_qs(sdata)['__youtubedl_smuggle'][0]
1031     data = json.loads(jsond)
1032     return url, data
1033
1034
1035 def format_bytes(bytes):
1036     if bytes is None:
1037         return 'N/A'
1038     if type(bytes) is str:
1039         bytes = float(bytes)
1040     if bytes == 0.0:
1041         exponent = 0
1042     else:
1043         exponent = int(math.log(bytes, 1024.0))
1044     suffix = ['B', 'KiB', 'MiB', 'GiB', 'TiB', 'PiB', 'EiB', 'ZiB', 'YiB'][exponent]
1045     converted = float(bytes) / float(1024 ** exponent)
1046     return '%.2f%s' % (converted, suffix)
1047
1048
1049 def parse_filesize(s):
1050     if s is None:
1051         return None
1052
1053     # The lower-case forms are of course incorrect and inofficial,
1054     # but we support those too
1055     _UNIT_TABLE = {
1056         'B': 1,
1057         'b': 1,
1058         'KiB': 1024,
1059         'KB': 1000,
1060         'kB': 1024,
1061         'Kb': 1000,
1062         'MiB': 1024 ** 2,
1063         'MB': 1000 ** 2,
1064         'mB': 1024 ** 2,
1065         'Mb': 1000 ** 2,
1066         'GiB': 1024 ** 3,
1067         'GB': 1000 ** 3,
1068         'gB': 1024 ** 3,
1069         'Gb': 1000 ** 3,
1070         'TiB': 1024 ** 4,
1071         'TB': 1000 ** 4,
1072         'tB': 1024 ** 4,
1073         'Tb': 1000 ** 4,
1074         'PiB': 1024 ** 5,
1075         'PB': 1000 ** 5,
1076         'pB': 1024 ** 5,
1077         'Pb': 1000 ** 5,
1078         'EiB': 1024 ** 6,
1079         'EB': 1000 ** 6,
1080         'eB': 1024 ** 6,
1081         'Eb': 1000 ** 6,
1082         'ZiB': 1024 ** 7,
1083         'ZB': 1000 ** 7,
1084         'zB': 1024 ** 7,
1085         'Zb': 1000 ** 7,
1086         'YiB': 1024 ** 8,
1087         'YB': 1000 ** 8,
1088         'yB': 1024 ** 8,
1089         'Yb': 1000 ** 8,
1090     }
1091
1092     units_re = '|'.join(re.escape(u) for u in _UNIT_TABLE)
1093     m = re.match(r'(?P<num>[0-9]+(?:\.[0-9]*)?)\s*(?P<unit>%s)' % units_re, s)
1094     if not m:
1095         return None
1096
1097     return int(float(m.group('num')) * _UNIT_TABLE[m.group('unit')])
1098
1099
1100 def get_term_width():
1101     columns = compat_getenv('COLUMNS', None)
1102     if columns:
1103         return int(columns)
1104
1105     try:
1106         sp = subprocess.Popen(
1107             ['stty', 'size'],
1108             stdout=subprocess.PIPE, stderr=subprocess.PIPE)
1109         out, err = sp.communicate()
1110         return int(out.split()[1])
1111     except:
1112         pass
1113     return None
1114
1115
1116 def month_by_name(name):
1117     """ Return the number of a month by (locale-independently) English name """
1118
1119     ENGLISH_NAMES = [
1120         'January', 'February', 'March', 'April', 'May', 'June',
1121         'July', 'August', 'September', 'October', 'November', 'December']
1122     try:
1123         return ENGLISH_NAMES.index(name) + 1
1124     except ValueError:
1125         return None
1126
1127
1128 def fix_xml_ampersands(xml_str):
1129     """Replace all the '&' by '&amp;' in XML"""
1130     return re.sub(
1131         r'&(?!amp;|lt;|gt;|apos;|quot;|#x[0-9a-fA-F]{,4};|#[0-9]{,4};)',
1132         '&amp;',
1133         xml_str)
1134
1135
1136 def setproctitle(title):
1137     assert isinstance(title, compat_str)
1138     try:
1139         libc = ctypes.cdll.LoadLibrary("libc.so.6")
1140     except OSError:
1141         return
1142     title_bytes = title.encode('utf-8')
1143     buf = ctypes.create_string_buffer(len(title_bytes))
1144     buf.value = title_bytes
1145     try:
1146         libc.prctl(15, buf, 0, 0, 0)
1147     except AttributeError:
1148         return  # Strange libc, just skip this
1149
1150
1151 def remove_start(s, start):
1152     if s.startswith(start):
1153         return s[len(start):]
1154     return s
1155
1156
1157 def remove_end(s, end):
1158     if s.endswith(end):
1159         return s[:-len(end)]
1160     return s
1161
1162
1163 def url_basename(url):
1164     path = compat_urlparse.urlparse(url).path
1165     return path.strip('/').split('/')[-1]
1166
1167
1168 class HEADRequest(compat_urllib_request.Request):
1169     def get_method(self):
1170         return "HEAD"
1171
1172
1173 def int_or_none(v, scale=1, default=None, get_attr=None, invscale=1):
1174     if get_attr:
1175         if v is not None:
1176             v = getattr(v, get_attr, None)
1177     if v == '':
1178         v = None
1179     return default if v is None else (int(v) * invscale // scale)
1180
1181
1182 def str_or_none(v, default=None):
1183     return default if v is None else compat_str(v)
1184
1185
1186 def str_to_int(int_str):
1187     """ A more relaxed version of int_or_none """
1188     if int_str is None:
1189         return None
1190     int_str = re.sub(r'[,\.\+]', '', int_str)
1191     return int(int_str)
1192
1193
1194 def float_or_none(v, scale=1, invscale=1, default=None):
1195     return default if v is None else (float(v) * invscale / scale)
1196
1197
1198 def parse_duration(s):
1199     if s is None:
1200         return None
1201
1202     s = s.strip()
1203
1204     m = re.match(
1205         r'''(?ix)T?
1206             (?:
1207                 (?:(?P<hours>[0-9]+)\s*(?:[:h]|hours?)\s*)?
1208                 (?P<mins>[0-9]+)\s*(?:[:m]|mins?|minutes?)\s*
1209             )?
1210             (?P<secs>[0-9]+)(?P<ms>\.[0-9]+)?\s*(?:s|secs?|seconds?)?$''', s)
1211     if not m:
1212         return None
1213     res = int(m.group('secs'))
1214     if m.group('mins'):
1215         res += int(m.group('mins')) * 60
1216         if m.group('hours'):
1217             res += int(m.group('hours')) * 60 * 60
1218     if m.group('ms'):
1219         res += float(m.group('ms'))
1220     return res
1221
1222
1223 def prepend_extension(filename, ext):
1224     name, real_ext = os.path.splitext(filename)
1225     return '{0}.{1}{2}'.format(name, ext, real_ext)
1226
1227
1228 def check_executable(exe, args=[]):
1229     """ Checks if the given binary is installed somewhere in PATH, and returns its name.
1230     args can be a list of arguments for a short output (like -version) """
1231     try:
1232         subprocess.Popen([exe] + args, stdout=subprocess.PIPE, stderr=subprocess.PIPE).communicate()
1233     except OSError:
1234         return False
1235     return exe
1236
1237
1238 def get_exe_version(exe, args=['--version'],
1239                     version_re=r'version\s+([0-9._-a-zA-Z]+)',
1240                     unrecognized='present'):
1241     """ Returns the version of the specified executable,
1242     or False if the executable is not present """
1243     try:
1244         out, err = subprocess.Popen(
1245             [exe] + args,
1246             stdout=subprocess.PIPE, stderr=subprocess.STDOUT).communicate()
1247     except OSError:
1248         return False
1249     firstline = out.partition(b'\n')[0].decode('ascii', 'ignore')
1250     m = re.search(version_re, firstline)
1251     if m:
1252         return m.group(1)
1253     else:
1254         return unrecognized
1255
1256
1257 class PagedList(object):
1258     def __len__(self):
1259         # This is only useful for tests
1260         return len(self.getslice())
1261
1262
1263 class OnDemandPagedList(PagedList):
1264     def __init__(self, pagefunc, pagesize):
1265         self._pagefunc = pagefunc
1266         self._pagesize = pagesize
1267
1268     def getslice(self, start=0, end=None):
1269         res = []
1270         for pagenum in itertools.count(start // self._pagesize):
1271             firstid = pagenum * self._pagesize
1272             nextfirstid = pagenum * self._pagesize + self._pagesize
1273             if start >= nextfirstid:
1274                 continue
1275
1276             page_results = list(self._pagefunc(pagenum))
1277
1278             startv = (
1279                 start % self._pagesize
1280                 if firstid <= start < nextfirstid
1281                 else 0)
1282
1283             endv = (
1284                 ((end - 1) % self._pagesize) + 1
1285                 if (end is not None and firstid <= end <= nextfirstid)
1286                 else None)
1287
1288             if startv != 0 or endv is not None:
1289                 page_results = page_results[startv:endv]
1290             res.extend(page_results)
1291
1292             # A little optimization - if current page is not "full", ie. does
1293             # not contain page_size videos then we can assume that this page
1294             # is the last one - there are no more ids on further pages -
1295             # i.e. no need to query again.
1296             if len(page_results) + startv < self._pagesize:
1297                 break
1298
1299             # If we got the whole page, but the next page is not interesting,
1300             # break out early as well
1301             if end == nextfirstid:
1302                 break
1303         return res
1304
1305
1306 class InAdvancePagedList(PagedList):
1307     def __init__(self, pagefunc, pagecount, pagesize):
1308         self._pagefunc = pagefunc
1309         self._pagecount = pagecount
1310         self._pagesize = pagesize
1311
1312     def getslice(self, start=0, end=None):
1313         res = []
1314         start_page = start // self._pagesize
1315         end_page = (
1316             self._pagecount if end is None else (end // self._pagesize + 1))
1317         skip_elems = start - start_page * self._pagesize
1318         only_more = None if end is None else end - start
1319         for pagenum in range(start_page, end_page):
1320             page = list(self._pagefunc(pagenum))
1321             if skip_elems:
1322                 page = page[skip_elems:]
1323                 skip_elems = None
1324             if only_more is not None:
1325                 if len(page) < only_more:
1326                     only_more -= len(page)
1327                 else:
1328                     page = page[:only_more]
1329                     res.extend(page)
1330                     break
1331             res.extend(page)
1332         return res
1333
1334
1335 def uppercase_escape(s):
1336     unicode_escape = codecs.getdecoder('unicode_escape')
1337     return re.sub(
1338         r'\\U[0-9a-fA-F]{8}',
1339         lambda m: unicode_escape(m.group(0))[0],
1340         s)
1341
1342
1343 def escape_rfc3986(s):
1344     """Escape non-ASCII characters as suggested by RFC 3986"""
1345     if sys.version_info < (3, 0) and isinstance(s, unicode):
1346         s = s.encode('utf-8')
1347     return compat_urllib_parse.quote(s, b"%/;:@&=+$,!~*'()?#[]")
1348
1349
1350 def escape_url(url):
1351     """Escape URL as suggested by RFC 3986"""
1352     url_parsed = compat_urllib_parse_urlparse(url)
1353     return url_parsed._replace(
1354         path=escape_rfc3986(url_parsed.path),
1355         params=escape_rfc3986(url_parsed.params),
1356         query=escape_rfc3986(url_parsed.query),
1357         fragment=escape_rfc3986(url_parsed.fragment)
1358     ).geturl()
1359
1360 try:
1361     struct.pack('!I', 0)
1362 except TypeError:
1363     # In Python 2.6 (and some 2.7 versions), struct requires a bytes argument
1364     def struct_pack(spec, *args):
1365         if isinstance(spec, compat_str):
1366             spec = spec.encode('ascii')
1367         return struct.pack(spec, *args)
1368
1369     def struct_unpack(spec, *args):
1370         if isinstance(spec, compat_str):
1371             spec = spec.encode('ascii')
1372         return struct.unpack(spec, *args)
1373 else:
1374     struct_pack = struct.pack
1375     struct_unpack = struct.unpack
1376
1377
1378 def read_batch_urls(batch_fd):
1379     def fixup(url):
1380         if not isinstance(url, compat_str):
1381             url = url.decode('utf-8', 'replace')
1382         BOM_UTF8 = '\xef\xbb\xbf'
1383         if url.startswith(BOM_UTF8):
1384             url = url[len(BOM_UTF8):]
1385         url = url.strip()
1386         if url.startswith(('#', ';', ']')):
1387             return False
1388         return url
1389
1390     with contextlib.closing(batch_fd) as fd:
1391         return [url for url in map(fixup, fd) if url]
1392
1393
1394 def urlencode_postdata(*args, **kargs):
1395     return compat_urllib_parse.urlencode(*args, **kargs).encode('ascii')
1396
1397
1398 try:
1399     etree_iter = xml.etree.ElementTree.Element.iter
1400 except AttributeError:  # Python <=2.6
1401     etree_iter = lambda n: n.findall('.//*')
1402
1403
1404 def parse_xml(s):
1405     class TreeBuilder(xml.etree.ElementTree.TreeBuilder):
1406         def doctype(self, name, pubid, system):
1407             pass  # Ignore doctypes
1408
1409     parser = xml.etree.ElementTree.XMLParser(target=TreeBuilder())
1410     kwargs = {'parser': parser} if sys.version_info >= (2, 7) else {}
1411     tree = xml.etree.ElementTree.XML(s.encode('utf-8'), **kwargs)
1412     # Fix up XML parser in Python 2.x
1413     if sys.version_info < (3, 0):
1414         for n in etree_iter(tree):
1415             if n.text is not None:
1416                 if not isinstance(n.text, compat_str):
1417                     n.text = n.text.decode('utf-8')
1418     return tree
1419
1420
1421 US_RATINGS = {
1422     'G': 0,
1423     'PG': 10,
1424     'PG-13': 13,
1425     'R': 16,
1426     'NC': 18,
1427 }
1428
1429
1430 def parse_age_limit(s):
1431     if s is None:
1432         return None
1433     m = re.match(r'^(?P<age>\d{1,2})\+?$', s)
1434     return int(m.group('age')) if m else US_RATINGS.get(s, None)
1435
1436
1437 def strip_jsonp(code):
1438     return re.sub(
1439         r'(?s)^[a-zA-Z0-9_]+\s*\(\s*(.*)\);?\s*?(?://[^\n]*)*$', r'\1', code)
1440
1441
1442 def js_to_json(code):
1443     def fix_kv(m):
1444         v = m.group(0)
1445         if v in ('true', 'false', 'null'):
1446             return v
1447         if v.startswith('"'):
1448             return v
1449         if v.startswith("'"):
1450             v = v[1:-1]
1451             v = re.sub(r"\\\\|\\'|\"", lambda m: {
1452                 '\\\\': '\\\\',
1453                 "\\'": "'",
1454                 '"': '\\"',
1455             }[m.group(0)], v)
1456         return '"%s"' % v
1457
1458     res = re.sub(r'''(?x)
1459         "(?:[^"\\]*(?:\\\\|\\")?)*"|
1460         '(?:[^'\\]*(?:\\\\|\\')?)*'|
1461         [a-zA-Z_][a-zA-Z_0-9]*
1462         ''', fix_kv, code)
1463     res = re.sub(r',(\s*\])', lambda m: m.group(1), res)
1464     return res
1465
1466
1467 def qualities(quality_ids):
1468     """ Get a numeric quality value out of a list of possible values """
1469     def q(qid):
1470         try:
1471             return quality_ids.index(qid)
1472         except ValueError:
1473             return -1
1474     return q
1475
1476
1477 DEFAULT_OUTTMPL = '%(title)s-%(id)s.%(ext)s'
1478
1479
1480 def limit_length(s, length):
1481     """ Add ellipses to overly long strings """
1482     if s is None:
1483         return None
1484     ELLIPSES = '...'
1485     if len(s) > length:
1486         return s[:length - len(ELLIPSES)] + ELLIPSES
1487     return s
1488
1489
1490 def version_tuple(v):
1491     return [int(e) for e in v.split('.')]
1492
1493
1494 def is_outdated_version(version, limit, assume_new=True):
1495     if not version:
1496         return not assume_new
1497     try:
1498         return version_tuple(version) < version_tuple(limit)
1499     except ValueError:
1500         return not assume_new
1501
1502
1503 def ytdl_is_updateable():
1504     """ Returns if youtube-dl can be updated with -U """
1505     from zipimport import zipimporter
1506
1507     return isinstance(globals().get('__loader__'), zipimporter) or hasattr(sys, 'frozen')
1508
1509
1510 def args_to_str(args):
1511     # Get a short string representation for a subprocess command
1512     return ' '.join(shlex_quote(a) for a in args)