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 base64
   7 import binascii
   8 import calendar
   9 import codecs
  10 import contextlib
  11 import ctypes
  12 import datetime
  13 import email.utils
  14 import errno
  15 import functools
  16 import gzip
  17 import itertools
  18 import io
  19 import json
  20 import locale
  21 import math
  22 import operator
  23 import os
  24 import pipes
  25 import platform
  26 import re
  27 import ssl
  28 import socket
  29 import struct
  30 import subprocess
  31 import sys
  32 import tempfile
  33 import traceback
  34 import xml.etree.ElementTree
  35 import zlib
  36
  37 from .compat import (
  38     compat_basestring,
  39     compat_chr,
  40     compat_etree_fromstring,
  41     compat_html_entities,
  42     compat_http_client,
  43     compat_kwargs,
  44     compat_parse_qs,
  45     compat_socket_create_connection,
  46     compat_str,
  47     compat_urllib_error,
  48     compat_urllib_parse,
  49     compat_urllib_parse_urlparse,
  50     compat_urllib_request,
  51     compat_urlparse,
  52     shlex_quote,
  53 )
  54
  55
  56 # This is not clearly defined otherwise
  57 compiled_regex_type = type(re.compile(''))
  58
  59 std_headers = {
  60     'User-Agent': 'Mozilla/5.0 (X11; Linux x86_64; rv:10.0) Gecko/20150101 Firefox/44.0 (Chrome)',
  61     'Accept-Charset': 'ISO-8859-1,utf-8;q=0.7,*;q=0.7',
  62     'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
  63     'Accept-Encoding': 'gzip, deflate',
  64     'Accept-Language': 'en-us,en;q=0.5',
  65 }
  66
  67
  68 NO_DEFAULT = object()
  69
  70 ENGLISH_MONTH_NAMES = [
  71     'January', 'February', 'March', 'April', 'May', 'June',
  72     'July', 'August', 'September', 'October', 'November', 'December']
  73
  74 KNOWN_EXTENSIONS = (
  75     'mp4', 'm4a', 'm4p', 'm4b', 'm4r', 'm4v', 'aac',
  76     'flv', 'f4v', 'f4a', 'f4b',
  77     'webm', 'ogg', 'ogv', 'oga', 'ogx', 'spx', 'opus',
  78     'mkv', 'mka', 'mk3d',
  79     'avi', 'divx',
  80     'mov',
  81     'asf', 'wmv', 'wma',
  82     '3gp', '3g2',
  83     'mp3',
  84     'flac',
  85     'ape',
  86     'wav',
  87     'f4f', 'f4m', 'm3u8', 'smil')
  88
  89
  90 def preferredencoding():
  91     """Get preferred encoding.
  92
  93     Returns the best encoding scheme for the system, based on
  94     locale.getpreferredencoding() and some further tweaks.
  95     """
  96     try:
  97         pref = locale.getpreferredencoding()
  98         'TEST'.encode(pref)
  99     except Exception:
 100         pref = 'UTF-8'
 101
 102     return pref
 103
 104
 105 def write_json_file(obj, fn):
 106     """ Encode obj as JSON and write it to fn, atomically if possible """
 107
 108     fn = encodeFilename(fn)
 109     if sys.version_info < (3, 0) and sys.platform != 'win32':
 110         encoding = get_filesystem_encoding()
 111         # os.path.basename returns a bytes object, but NamedTemporaryFile
 112         # will fail if the filename contains non ascii characters unless we
 113         # use a unicode object
 114         path_basename = lambda f: os.path.basename(fn).decode(encoding)
 115         # the same for os.path.dirname
 116         path_dirname = lambda f: os.path.dirname(fn).decode(encoding)
 117     else:
 118         path_basename = os.path.basename
 119         path_dirname = os.path.dirname
 120
 121     args = {
 122         'suffix': '.tmp',
 123         'prefix': path_basename(fn) + '.',
 124         'dir': path_dirname(fn),
 125         'delete': False,
 126     }
 127
 128     # In Python 2.x, json.dump expects a bytestream.
 129     # In Python 3.x, it writes to a character stream
 130     if sys.version_info < (3, 0):
 131         args['mode'] = 'wb'
 132     else:
 133         args.update({
 134             'mode': 'w',
 135             'encoding': 'utf-8',
 136         })
 137
 138     tf = tempfile.NamedTemporaryFile(**compat_kwargs(args))
 139
 140     try:
 141         with tf:
 142             json.dump(obj, tf)
 143         if sys.platform == 'win32':
 144             # Need to remove existing file on Windows, else os.rename raises
 145             # WindowsError or FileExistsError.
 146             try:
 147                 os.unlink(fn)
 148             except OSError:
 149                 pass
 150         os.rename(tf.name, fn)
 151     except Exception:
 152         try:
 153             os.remove(tf.name)
 154         except OSError:
 155             pass
 156         raise
 157
 158
 159 if sys.version_info >= (2, 7):
 160     def find_xpath_attr(node, xpath, key, val=None):
 161         """ Find the xpath xpath[@key=val] """
 162         assert re.match(r'^[a-zA-Z_-]+$', key)
 163         if val:
 164             assert re.match(r'^[a-zA-Z0-9@\s:._-]*$', val)
 165         expr = xpath + ('[@%s]' % key if val is None else "[@%s='%s']" % (key, val))
 166         return node.find(expr)
 167 else:
 168     def find_xpath_attr(node, xpath, key, val=None):
 169         # Here comes the crazy part: In 2.6, if the xpath is a unicode,
 170         # .//node does not match if a node is a direct child of . !
 171         if isinstance(xpath, compat_str):
 172             xpath = xpath.encode('ascii')
 173
 174         for f in node.findall(xpath):
 175             if key not in f.attrib:
 176                 continue
 177             if val is None or f.attrib.get(key) == val:
 178                 return f
 179         return None
 180
 181 # On python2.6 the xml.etree.ElementTree.Element methods don't support
 182 # the namespace parameter
 183
 184
 185 def xpath_with_ns(path, ns_map):
 186     components = [c.split(':') for c in path.split('/')]
 187     replaced = []
 188     for c in components:
 189         if len(c) == 1:
 190             replaced.append(c[0])
 191         else:
 192             ns, tag = c
 193             replaced.append('{%s}%s' % (ns_map[ns], tag))
 194     return '/'.join(replaced)
 195
 196
 197 def xpath_element(node, xpath, name=None, fatal=False, default=NO_DEFAULT):
 198     def _find_xpath(xpath):
 199         if sys.version_info < (2, 7):  # Crazy 2.6
 200             xpath = xpath.encode('ascii')
 201         return node.find(xpath)
 202
 203     if isinstance(xpath, (str, compat_str)):
 204         n = _find_xpath(xpath)
 205     else:
 206         for xp in xpath:
 207             n = _find_xpath(xp)
 208             if n is not None:
 209                 break
 210
 211     if n is None:
 212         if default is not NO_DEFAULT:
 213             return default
 214         elif fatal:
 215             name = xpath if name is None else name
 216             raise ExtractorError('Could not find XML element %s' % name)
 217         else:
 218             return None
 219     return n
 220
 221
 222 def xpath_text(node, xpath, name=None, fatal=False, default=NO_DEFAULT):
 223     n = xpath_element(node, xpath, name, fatal=fatal, default=default)
 224     if n is None or n == default:
 225         return n
 226     if n.text is None:
 227         if default is not NO_DEFAULT:
 228             return default
 229         elif fatal:
 230             name = xpath if name is None else name
 231             raise ExtractorError('Could not find XML element\'s text %s' % name)
 232         else:
 233             return None
 234     return n.text
 235
 236
 237 def xpath_attr(node, xpath, key, name=None, fatal=False, default=NO_DEFAULT):
 238     n = find_xpath_attr(node, xpath, key)
 239     if n is None:
 240         if default is not NO_DEFAULT:
 241             return default
 242         elif fatal:
 243             name = '%s[@%s]' % (xpath, key) if name is None else name
 244             raise ExtractorError('Could not find XML attribute %s' % name)
 245         else:
 246             return None
 247     return n.attrib[key]
 248
 249
 250 def get_element_by_id(id, html):
 251     """Return the content of the tag with the specified ID in the passed HTML document"""
 252     return get_element_by_attribute('id', id, html)
 253
 254
 255 def get_element_by_attribute(attribute, value, html):
 256     """Return the content of the tag with the specified attribute in the passed HTML document"""
 257
 258     m = re.search(r'''(?xs)
 259         <([a-zA-Z0-9:._-]+)
 260          (?:\s+[a-zA-Z0-9:._-]+(?:=[a-zA-Z0-9:._-]+|="[^"]+"|='[^']+'))*?
 261          \s+%s=['"]?%s['"]?
 262          (?:\s+[a-zA-Z0-9:._-]+(?:=[a-zA-Z0-9:._-]+|="[^"]+"|='[^']+'))*?
 263         \s*>
 264         (?P<content>.*?)
 265         </\1>
 266     ''' % (re.escape(attribute), re.escape(value)), html)
 267
 268     if not m:
 269         return None
 270     res = m.group('content')
 271
 272     if res.startswith('"') or res.startswith("'"):
 273         res = res[1:-1]
 274
 275     return unescapeHTML(res)
 276
 277
 278 def clean_html(html):
 279     """Clean an HTML snippet into a readable string"""
 280
 281     if html is None:  # Convenience for sanitizing descriptions etc.
 282         return html
 283
 284     # Newline vs <br />
 285     html = html.replace('\n', ' ')
 286     html = re.sub(r'\s*<\s*br\s*/?\s*>\s*', '\n', html)
 287     html = re.sub(r'<\s*/\s*p\s*>\s*<\s*p[^>]*>', '\n', html)
 288     # Strip html tags
 289     html = re.sub('<.*?>', '', html)
 290     # Replace html entities
 291     html = unescapeHTML(html)
 292     return html.strip()
 293
 294
 295 def sanitize_open(filename, open_mode):
 296     """Try to open the given filename, and slightly tweak it if this fails.
 297
 298     Attempts to open the given filename. If this fails, it tries to change
 299     the filename slightly, step by step, until it's either able to open it
 300     or it fails and raises a final exception, like the standard open()
 301     function.
 302
 303     It returns the tuple (stream, definitive_file_name).
 304     """
 305     try:
 306         if filename == '-':
 307             if sys.platform == 'win32':
 308                 import msvcrt
 309                 msvcrt.setmode(sys.stdout.fileno(), os.O_BINARY)
 310             return (sys.stdout.buffer if hasattr(sys.stdout, 'buffer') else sys.stdout, filename)
 311         stream = open(encodeFilename(filename), open_mode)
 312         return (stream, filename)
 313     except (IOError, OSError) as err:
 314         if err.errno in (errno.EACCES,):
 315             raise
 316
 317         # In case of error, try to remove win32 forbidden chars
 318         alt_filename = sanitize_path(filename)
 319         if alt_filename == filename:
 320             raise
 321         else:
 322             # An exception here should be caught in the caller
 323             stream = open(encodeFilename(alt_filename), open_mode)
 324             return (stream, alt_filename)
 325
 326
 327 def timeconvert(timestr):
 328     """Convert RFC 2822 defined time string into system timestamp"""
 329     timestamp = None
 330     timetuple = email.utils.parsedate_tz(timestr)
 331     if timetuple is not None:
 332         timestamp = email.utils.mktime_tz(timetuple)
 333     return timestamp
 334
 335
 336 def sanitize_filename(s, restricted=False, is_id=False):
 337     """Sanitizes a string so it could be used as part of a filename.
 338     If restricted is set, use a stricter subset of allowed characters.
 339     Set is_id if this is not an arbitrary string, but an ID that should be kept if possible
 340     """
 341     def replace_insane(char):
 342         if char == '?' or ord(char) < 32 or ord(char) == 127:
 343             return ''
 344         elif char == '"':
 345             return '' if restricted else '\''
 346         elif char == ':':
 347             return '_-' if restricted else ' -'
 348         elif char in '\\/|*<>':
 349             return '_'
 350         if restricted and (char in '!&\'()[]{}$;`^,#' or char.isspace()):
 351             return '_'
 352         if restricted and ord(char) > 127:
 353             return '_'
 354         return char
 355
 356     # Handle timestamps
 357     s = re.sub(r'[0-9]+(?::[0-9]+)+', lambda m: m.group(0).replace(':', '_'), s)
 358     result = ''.join(map(replace_insane, s))
 359     if not is_id:
 360         while '__' in result:
 361             result = result.replace('__', '_')
 362         result = result.strip('_')
 363         # Common case of "Foreign band name - English song title"
 364         if restricted and result.startswith('-_'):
 365             result = result[2:]
 366         if result.startswith('-'):
 367             result = '_' + result[len('-'):]
 368         result = result.lstrip('.')
 369         if not result:
 370             result = '_'
 371     return result
 372
 373
 374 def sanitize_path(s):
 375     """Sanitizes and normalizes path on Windows"""
 376     if sys.platform != 'win32':
 377         return s
 378     drive_or_unc, _ = os.path.splitdrive(s)
 379     if sys.version_info < (2, 7) and not drive_or_unc:
 380         drive_or_unc, _ = os.path.splitunc(s)
 381     norm_path = os.path.normpath(remove_start(s, drive_or_unc)).split(os.path.sep)
 382     if drive_or_unc:
 383         norm_path.pop(0)
 384     sanitized_path = [
 385         path_part if path_part in ['.', '..'] else re.sub('(?:[/<>:"\\|\\\\?\\*]|[\s.]$)', '#', path_part)
 386         for path_part in norm_path]
 387     if drive_or_unc:
 388         sanitized_path.insert(0, drive_or_unc + os.path.sep)
 389     return os.path.join(*sanitized_path)
 390
 391
 392 # Prepend protocol-less URLs with `http:` scheme in order to mitigate the number of
 393 # unwanted failures due to missing protocol
 394 def sanitized_Request(url, *args, **kwargs):
 395     return compat_urllib_request.Request(
 396         'http:%s' % url if url.startswith('//') else url, *args, **kwargs)
 397
 398
 399 def orderedSet(iterable):
 400     """ Remove all duplicates from the input iterable """
 401     res = []
 402     for el in iterable:
 403         if el not in res:
 404             res.append(el)
 405     return res
 406
 407
 408 def _htmlentity_transform(entity):
 409     """Transforms an HTML entity to a character."""
 410     # Known non-numeric HTML entity
 411     if entity in compat_html_entities.name2codepoint:
 412         return compat_chr(compat_html_entities.name2codepoint[entity])
 413
 414     mobj = re.match(r'#(x[0-9a-fA-F]+|[0-9]+)', entity)
 415     if mobj is not None:
 416         numstr = mobj.group(1)
 417         if numstr.startswith('x'):
 418             base = 16
 419             numstr = '0%s' % numstr
 420         else:
 421             base = 10
 422         # See https://github.com/rg3/youtube-dl/issues/7518
 423         try:
 424             return compat_chr(int(numstr, base))
 425         except ValueError:
 426             pass
 427
 428     # Unknown entity in name, return its literal representation
 429     return '&%s;' % entity
 430
 431
 432 def unescapeHTML(s):
 433     if s is None:
 434         return None
 435     assert type(s) == compat_str
 436
 437     return re.sub(
 438         r'&([^;]+);', lambda m: _htmlentity_transform(m.group(1)), s)
 439
 440
 441 def get_subprocess_encoding():
 442     if sys.platform == 'win32' and sys.getwindowsversion()[0] >= 5:
 443         # For subprocess calls, encode with locale encoding
 444         # Refer to http://stackoverflow.com/a/9951851/35070
 445         encoding = preferredencoding()
 446     else:
 447         encoding = sys.getfilesystemencoding()
 448     if encoding is None:
 449         encoding = 'utf-8'
 450     return encoding
 451
 452
 453 def encodeFilename(s, for_subprocess=False):
 454     """
 455     @param s The name of the file
 456     """
 457
 458     assert type(s) == compat_str
 459
 460     # Python 3 has a Unicode API
 461     if sys.version_info >= (3, 0):
 462         return s
 463
 464     # Pass '' directly to use Unicode APIs on Windows 2000 and up
 465     # (Detecting Windows NT 4 is tricky because 'major >= 4' would
 466     # match Windows 9x series as well. Besides, NT 4 is obsolete.)
 467     if not for_subprocess and sys.platform == 'win32' and sys.getwindowsversion()[0] >= 5:
 468         return s
 469
 470     return s.encode(get_subprocess_encoding(), 'ignore')
 471
 472
 473 def decodeFilename(b, for_subprocess=False):
 474
 475     if sys.version_info >= (3, 0):
 476         return b
 477
 478     if not isinstance(b, bytes):
 479         return b
 480
 481     return b.decode(get_subprocess_encoding(), 'ignore')
 482
 483
 484 def encodeArgument(s):
 485     if not isinstance(s, compat_str):
 486         # Legacy code that uses byte strings
 487         # Uncomment the following line after fixing all post processors
 488         # assert False, 'Internal error: %r should be of type %r, is %r' % (s, compat_str, type(s))
 489         s = s.decode('ascii')
 490     return encodeFilename(s, True)
 491
 492
 493 def decodeArgument(b):
 494     return decodeFilename(b, True)
 495
 496
 497 def decodeOption(optval):
 498     if optval is None:
 499         return optval
 500     if isinstance(optval, bytes):
 501         optval = optval.decode(preferredencoding())
 502
 503     assert isinstance(optval, compat_str)
 504     return optval
 505
 506
 507 def formatSeconds(secs):
 508     if secs > 3600:
 509         return '%d:%02d:%02d' % (secs // 3600, (secs % 3600) // 60, secs % 60)
 510     elif secs > 60:
 511         return '%d:%02d' % (secs // 60, secs % 60)
 512     else:
 513         return '%d' % secs
 514
 515
 516 def make_HTTPS_handler(params, **kwargs):
 517     opts_no_check_certificate = params.get('nocheckcertificate', False)
 518     if hasattr(ssl, 'create_default_context'):  # Python >= 3.4 or 2.7.9
 519         context = ssl.create_default_context(ssl.Purpose.SERVER_AUTH)
 520         if opts_no_check_certificate:
 521             context.check_hostname = False
 522             context.verify_mode = ssl.CERT_NONE
 523         try:
 524             return YoutubeDLHTTPSHandler(params, context=context, **kwargs)
 525         except TypeError:
 526             # Python 2.7.8
 527             # (create_default_context present but HTTPSHandler has no context=)
 528             pass
 529
 530     if sys.version_info < (3, 2):
 531         return YoutubeDLHTTPSHandler(params, **kwargs)
 532     else:  # Python < 3.4
 533         context = ssl.SSLContext(ssl.PROTOCOL_TLSv1)
 534         context.verify_mode = (ssl.CERT_NONE
 535                                if opts_no_check_certificate
 536                                else ssl.CERT_REQUIRED)
 537         context.set_default_verify_paths()
 538         return YoutubeDLHTTPSHandler(params, context=context, **kwargs)
 539
 540
 541 def bug_reports_message():
 542     if ytdl_is_updateable():
 543         update_cmd = 'type  youtube-dl -U  to update'
 544     else:
 545         update_cmd = 'see  https://yt-dl.org/update  on how to update'
 546     msg = '; please report this issue on https://yt-dl.org/bug .'
 547     msg += ' Make sure you are using the latest version; %s.' % update_cmd
 548     msg += ' Be sure to call youtube-dl with the --verbose flag and include its complete output.'
 549     return msg
 550
 551
 552 class ExtractorError(Exception):
 553     """Error during info extraction."""
 554
 555     def __init__(self, msg, tb=None, expected=False, cause=None, video_id=None):
 556         """ tb, if given, is the original traceback (so that it can be printed out).
 557         If expected is set, this is a normal error message and most likely not a bug in youtube-dl.
 558         """
 559
 560         if sys.exc_info()[0] in (compat_urllib_error.URLError, socket.timeout, UnavailableVideoError):
 561             expected = True
 562         if video_id is not None:
 563             msg = video_id + ': ' + msg
 564         if cause:
 565             msg += ' (caused by %r)' % cause
 566         if not expected:
 567             msg += bug_reports_message()
 568         super(ExtractorError, self).__init__(msg)
 569
 570         self.traceback = tb
 571         self.exc_info = sys.exc_info()  # preserve original exception
 572         self.cause = cause
 573         self.video_id = video_id
 574
 575     def format_traceback(self):
 576         if self.traceback is None:
 577             return None
 578         return ''.join(traceback.format_tb(self.traceback))
 579
 580
 581 class UnsupportedError(ExtractorError):
 582     def __init__(self, url):
 583         super(UnsupportedError, self).__init__(
 584             'Unsupported URL: %s' % url, expected=True)
 585         self.url = url
 586
 587
 588 class RegexNotFoundError(ExtractorError):
 589     """Error when a regex didn't match"""
 590     pass
 591
 592
 593 class DownloadError(Exception):
 594     """Download Error exception.
 595
 596     This exception may be thrown by FileDownloader objects if they are not
 597     configured to continue on errors. They will contain the appropriate
 598     error message.
 599     """
 600
 601     def __init__(self, msg, exc_info=None):
 602         """ exc_info, if given, is the original exception that caused the trouble (as returned by sys.exc_info()). """
 603         super(DownloadError, self).__init__(msg)
 604         self.exc_info = exc_info
 605
 606
 607 class SameFileError(Exception):
 608     """Same File exception.
 609
 610     This exception will be thrown by FileDownloader objects if they detect
 611     multiple files would have to be downloaded to the same file on disk.
 612     """
 613     pass
 614
 615
 616 class PostProcessingError(Exception):
 617     """Post Processing exception.
 618
 619     This exception may be raised by PostProcessor's .run() method to
 620     indicate an error in the postprocessing task.
 621     """
 622
 623     def __init__(self, msg):
 624         self.msg = msg
 625
 626
 627 class MaxDownloadsReached(Exception):
 628     """ --max-downloads limit has been reached. """
 629     pass
 630
 631
 632 class UnavailableVideoError(Exception):
 633     """Unavailable Format exception.
 634
 635     This exception will be thrown when a video is requested
 636     in a format that is not available for that video.
 637     """
 638     pass
 639
 640
 641 class ContentTooShortError(Exception):
 642     """Content Too Short exception.
 643
 644     This exception may be raised by FileDownloader objects when a file they
 645     download is too small for what the server announced first, indicating
 646     the connection was probably interrupted.
 647     """
 648
 649     def __init__(self, downloaded, expected):
 650         # Both in bytes
 651         self.downloaded = downloaded
 652         self.expected = expected
 653
 654
 655 def _create_http_connection(ydl_handler, http_class, is_https, *args, **kwargs):
 656     # Working around python 2 bug (see http://bugs.python.org/issue17849) by limiting
 657     # expected HTTP responses to meet HTTP/1.0 or later (see also
 658     # https://github.com/rg3/youtube-dl/issues/6727)
 659     if sys.version_info < (3, 0):
 660         kwargs[b'strict'] = True
 661     hc = http_class(*args, **kwargs)
 662     source_address = ydl_handler._params.get('source_address')
 663     if source_address is not None:
 664         sa = (source_address, 0)
 665         if hasattr(hc, 'source_address'):  # Python 2.7+
 666             hc.source_address = sa
 667         else:  # Python 2.6
 668             def _hc_connect(self, *args, **kwargs):
 669                 sock = compat_socket_create_connection(
 670                     (self.host, self.port), self.timeout, sa)
 671                 if is_https:
 672                     self.sock = ssl.wrap_socket(
 673                         sock, self.key_file, self.cert_file,
 674                         ssl_version=ssl.PROTOCOL_TLSv1)
 675                 else:
 676                     self.sock = sock
 677             hc.connect = functools.partial(_hc_connect, hc)
 678
 679     return hc
 680
 681
 682 def handle_youtubedl_headers(headers):
 683     filtered_headers = headers
 684
 685     if 'Youtubedl-no-compression' in filtered_headers:
 686         filtered_headers = dict((k, v) for k, v in filtered_headers.items() if k.lower() != 'accept-encoding')
 687         del filtered_headers['Youtubedl-no-compression']
 688
 689     return filtered_headers
 690
 691
 692 class YoutubeDLHandler(compat_urllib_request.HTTPHandler):
 693     """Handler for HTTP requests and responses.
 694
 695     This class, when installed with an OpenerDirector, automatically adds
 696     the standard headers to every HTTP request and handles gzipped and
 697     deflated responses from web servers. If compression is to be avoided in
 698     a particular request, the original request in the program code only has
 699     to include the HTTP header "Youtubedl-no-compression", which will be
 700     removed before making the real request.
 701
 702     Part of this code was copied from:
 703
 704     http://techknack.net/python-urllib2-handlers/
 705
 706     Andrew Rowls, the author of that code, agreed to release it to the
 707     public domain.
 708     """
 709
 710     def __init__(self, params, *args, **kwargs):
 711         compat_urllib_request.HTTPHandler.__init__(self, *args, **kwargs)
 712         self._params = params
 713
 714     def http_open(self, req):
 715         return self.do_open(functools.partial(
 716             _create_http_connection, self, compat_http_client.HTTPConnection, False),
 717             req)
 718
 719     @staticmethod
 720     def deflate(data):
 721         try:
 722             return zlib.decompress(data, -zlib.MAX_WBITS)
 723         except zlib.error:
 724             return zlib.decompress(data)
 725
 726     @staticmethod
 727     def addinfourl_wrapper(stream, headers, url, code):
 728         if hasattr(compat_urllib_request.addinfourl, 'getcode'):
 729             return compat_urllib_request.addinfourl(stream, headers, url, code)
 730         ret = compat_urllib_request.addinfourl(stream, headers, url)
 731         ret.code = code
 732         return ret
 733
 734     def http_request(self, req):
 735         # According to RFC 3986, URLs can not contain non-ASCII characters, however this is not
 736         # always respected by websites, some tend to give out URLs with non percent-encoded
 737         # non-ASCII characters (see telemb.py, ard.py [#3412])
 738         # urllib chokes on URLs with non-ASCII characters (see http://bugs.python.org/issue3991)
 739         # To work around aforementioned issue we will replace request's original URL with
 740         # percent-encoded one
 741         # Since redirects are also affected (e.g. http://www.southpark.de/alle-episoden/s18e09)
 742         # the code of this workaround has been moved here from YoutubeDL.urlopen()
 743         url = req.get_full_url()
 744         url_escaped = escape_url(url)
 745
 746         # Substitute URL if any change after escaping
 747         if url != url_escaped:
 748             req_type = HEADRequest if req.get_method() == 'HEAD' else compat_urllib_request.Request
 749             new_req = req_type(
 750                 url_escaped, data=req.data, headers=req.headers,
 751                 origin_req_host=req.origin_req_host, unverifiable=req.unverifiable)
 752             new_req.timeout = req.timeout
 753             req = new_req
 754
 755         for h, v in std_headers.items():
 756             # Capitalize is needed because of Python bug 2275: http://bugs.python.org/issue2275
 757             # The dict keys are capitalized because of this bug by urllib
 758             if h.capitalize() not in req.headers:
 759                 req.add_header(h, v)
 760
 761         req.headers = handle_youtubedl_headers(req.headers)
 762
 763         if sys.version_info < (2, 7) and '#' in req.get_full_url():
 764             # Python 2.6 is brain-dead when it comes to fragments
 765             req._Request__original = req._Request__original.partition('#')[0]
 766             req._Request__r_type = req._Request__r_type.partition('#')[0]
 767
 768         return req
 769
 770     def http_response(self, req, resp):
 771         old_resp = resp
 772         # gzip
 773         if resp.headers.get('Content-encoding', '') == 'gzip':
 774             content = resp.read()
 775             gz = gzip.GzipFile(fileobj=io.BytesIO(content), mode='rb')
 776             try:
 777                 uncompressed = io.BytesIO(gz.read())
 778             except IOError as original_ioerror:
 779                 # There may be junk add the end of the file
 780                 # See http://stackoverflow.com/q/4928560/35070 for details
 781                 for i in range(1, 1024):
 782                     try:
 783                         gz = gzip.GzipFile(fileobj=io.BytesIO(content[:-i]), mode='rb')
 784                         uncompressed = io.BytesIO(gz.read())
 785                     except IOError:
 786                         continue
 787                     break
 788                 else:
 789                     raise original_ioerror
 790             resp = self.addinfourl_wrapper(uncompressed, old_resp.headers, old_resp.url, old_resp.code)
 791             resp.msg = old_resp.msg
 792             del resp.headers['Content-encoding']
 793         # deflate
 794         if resp.headers.get('Content-encoding', '') == 'deflate':
 795             gz = io.BytesIO(self.deflate(resp.read()))
 796             resp = self.addinfourl_wrapper(gz, old_resp.headers, old_resp.url, old_resp.code)
 797             resp.msg = old_resp.msg
 798             del resp.headers['Content-encoding']
 799         # Percent-encode redirect URL of Location HTTP header to satisfy RFC 3986 (see
 800         # https://github.com/rg3/youtube-dl/issues/6457).
 801         if 300 <= resp.code < 400:
 802             location = resp.headers.get('Location')
 803             if location:
 804                 # As of RFC 2616 default charset is iso-8859-1 that is respected by python 3
 805                 if sys.version_info >= (3, 0):
 806                     location = location.encode('iso-8859-1').decode('utf-8')
 807                 location_escaped = escape_url(location)
 808                 if location != location_escaped:
 809                     del resp.headers['Location']
 810                     resp.headers['Location'] = location_escaped
 811         return resp
 812
 813     https_request = http_request
 814     https_response = http_response
 815
 816
 817 class YoutubeDLHTTPSHandler(compat_urllib_request.HTTPSHandler):
 818     def __init__(self, params, https_conn_class=None, *args, **kwargs):
 819         compat_urllib_request.HTTPSHandler.__init__(self, *args, **kwargs)
 820         self._https_conn_class = https_conn_class or compat_http_client.HTTPSConnection
 821         self._params = params
 822
 823     def https_open(self, req):
 824         kwargs = {}
 825         if hasattr(self, '_context'):  # python > 2.6
 826             kwargs['context'] = self._context
 827         if hasattr(self, '_check_hostname'):  # python 3.x
 828             kwargs['check_hostname'] = self._check_hostname
 829         return self.do_open(functools.partial(
 830             _create_http_connection, self, self._https_conn_class, True),
 831             req, **kwargs)
 832
 833
 834 class YoutubeDLCookieProcessor(compat_urllib_request.HTTPCookieProcessor):
 835     def __init__(self, cookiejar=None):
 836         compat_urllib_request.HTTPCookieProcessor.__init__(self, cookiejar)
 837
 838     def http_response(self, request, response):
 839         # Python 2 will choke on next HTTP request in row if there are non-ASCII
 840         # characters in Set-Cookie HTTP header of last response (see
 841         # https://github.com/rg3/youtube-dl/issues/6769).
 842         # In order to at least prevent crashing we will percent encode Set-Cookie
 843         # header before HTTPCookieProcessor starts processing it.
 844         # if sys.version_info < (3, 0) and response.headers:
 845         #     for set_cookie_header in ('Set-Cookie', 'Set-Cookie2'):
 846         #         set_cookie = response.headers.get(set_cookie_header)
 847         #         if set_cookie:
 848         #             set_cookie_escaped = compat_urllib_parse.quote(set_cookie, b"%/;:@&=+$,!~*'()?#[] ")
 849         #             if set_cookie != set_cookie_escaped:
 850         #                 del response.headers[set_cookie_header]
 851         #                 response.headers[set_cookie_header] = set_cookie_escaped
 852         return compat_urllib_request.HTTPCookieProcessor.http_response(self, request, response)
 853
 854     https_request = compat_urllib_request.HTTPCookieProcessor.http_request
 855     https_response = http_response
 856
 857
 858 def parse_iso8601(date_str, delimiter='T', timezone=None):
 859     """ Return a UNIX timestamp from the given date """
 860
 861     if date_str is None:
 862         return None
 863
 864     date_str = re.sub(r'\.[0-9]+', '', date_str)
 865
 866     if timezone is None:
 867         m = re.search(
 868             r'(?:Z$| ?(?P<sign>\+|-)(?P<hours>[0-9]{2}):?(?P<minutes>[0-9]{2})$)',
 869             date_str)
 870         if not m:
 871             timezone = datetime.timedelta()
 872         else:
 873             date_str = date_str[:-len(m.group(0))]
 874             if not m.group('sign'):
 875                 timezone = datetime.timedelta()
 876             else:
 877                 sign = 1 if m.group('sign') == '+' else -1
 878                 timezone = datetime.timedelta(
 879                     hours=sign * int(m.group('hours')),
 880                     minutes=sign * int(m.group('minutes')))
 881     try:
 882         date_format = '%Y-%m-%d{0}%H:%M:%S'.format(delimiter)
 883         dt = datetime.datetime.strptime(date_str, date_format) - timezone
 884         return calendar.timegm(dt.timetuple())
 885     except ValueError:
 886         pass
 887
 888
 889 def unified_strdate(date_str, day_first=True):
 890     """Return a string with the date in the format YYYYMMDD"""
 891
 892     if date_str is None:
 893         return None
 894     upload_date = None
 895     # Replace commas
 896     date_str = date_str.replace(',', ' ')
 897     # %z (UTC offset) is only supported in python>=3.2
 898     if not re.match(r'^[0-9]{1,2}-[0-9]{1,2}-[0-9]{4}$', date_str):
 899         date_str = re.sub(r' ?(\+|-)[0-9]{2}:?[0-9]{2}$', '', date_str)
 900     # Remove AM/PM + timezone
 901     date_str = re.sub(r'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str)
 902
 903     format_expressions = [
 904         '%d %B %Y',
 905         '%d %b %Y',
 906         '%B %d %Y',
 907         '%b %d %Y',
 908         '%b %dst %Y %I:%M%p',
 909         '%b %dnd %Y %I:%M%p',
 910         '%b %dth %Y %I:%M%p',
 911         '%Y %m %d',
 912         '%Y-%m-%d',
 913         '%Y/%m/%d',
 914         '%Y/%m/%d %H:%M:%S',
 915         '%Y-%m-%d %H:%M:%S',
 916         '%Y-%m-%d %H:%M:%S.%f',
 917         '%d.%m.%Y %H:%M',
 918         '%d.%m.%Y %H.%M',
 919         '%Y-%m-%dT%H:%M:%SZ',
 920         '%Y-%m-%dT%H:%M:%S.%fZ',
 921         '%Y-%m-%dT%H:%M:%S.%f0Z',
 922         '%Y-%m-%dT%H:%M:%S',
 923         '%Y-%m-%dT%H:%M:%S.%f',
 924         '%Y-%m-%dT%H:%M',
 925     ]
 926     if day_first:
 927         format_expressions.extend([
 928             '%d-%m-%Y',
 929             '%d.%m.%Y',
 930             '%d/%m/%Y',
 931             '%d/%m/%y',
 932             '%d/%m/%Y %H:%M:%S',
 933         ])
 934     else:
 935         format_expressions.extend([
 936             '%m-%d-%Y',
 937             '%m.%d.%Y',
 938             '%m/%d/%Y',
 939             '%m/%d/%y',
 940             '%m/%d/%Y %H:%M:%S',
 941         ])
 942     for expression in format_expressions:
 943         try:
 944             upload_date = datetime.datetime.strptime(date_str, expression).strftime('%Y%m%d')
 945         except ValueError:
 946             pass
 947     if upload_date is None:
 948         timetuple = email.utils.parsedate_tz(date_str)
 949         if timetuple:
 950             upload_date = datetime.datetime(*timetuple[:6]).strftime('%Y%m%d')
 951     if upload_date is not None:
 952         return compat_str(upload_date)
 953
 954
 955 def determine_ext(url, default_ext='unknown_video'):
 956     if url is None:
 957         return default_ext
 958     guess = url.partition('?')[0].rpartition('.')[2]
 959     if re.match(r'^[A-Za-z0-9]+$', guess):
 960         return guess
 961     # Try extract ext from URLs like http://example.com/foo/bar.mp4/?download
 962     elif guess.rstrip('/') in KNOWN_EXTENSIONS:
 963         return guess.rstrip('/')
 964     else:
 965         return default_ext
 966
 967
 968 def subtitles_filename(filename, sub_lang, sub_format):
 969     return filename.rsplit('.', 1)[0] + '.' + sub_lang + '.' + sub_format
 970
 971
 972 def date_from_str(date_str):
 973     """
 974     Return a datetime object from a string in the format YYYYMMDD or
 975     (now|today)[+-][0-9](day|week|month|year)(s)?"""
 976     today = datetime.date.today()
 977     if date_str in ('now', 'today'):
 978         return today
 979     if date_str == 'yesterday':
 980         return today - datetime.timedelta(days=1)
 981     match = re.match('(now|today)(?P<sign>[+-])(?P<time>\d+)(?P<unit>day|week|month|year)(s)?', date_str)
 982     if match is not None:
 983         sign = match.group('sign')
 984         time = int(match.group('time'))
 985         if sign == '-':
 986             time = -time
 987         unit = match.group('unit')
 988         # A bad approximation?
 989         if unit == 'month':
 990             unit = 'day'
 991             time *= 30
 992         elif unit == 'year':
 993             unit = 'day'
 994             time *= 365
 995         unit += 's'
 996         delta = datetime.timedelta(**{unit: time})
 997         return today + delta
 998     return datetime.datetime.strptime(date_str, '%Y%m%d').date()
 999
1000
1001 def hyphenate_date(date_str):
1002     """
1003     Convert a date in 'YYYYMMDD' format to 'YYYY-MM-DD' format"""
1004     match = re.match(r'^(\d\d\d\d)(\d\d)(\d\d)$', date_str)
1005     if match is not None:
1006         return '-'.join(match.groups())
1007     else:
1008         return date_str
1009
1010
1011 class DateRange(object):
1012     """Represents a time interval between two dates"""
1013
1014     def __init__(self, start=None, end=None):
1015         """start and end must be strings in the format accepted by date"""
1016         if start is not None:
1017             self.start = date_from_str(start)
1018         else:
1019             self.start = datetime.datetime.min.date()
1020         if end is not None:
1021             self.end = date_from_str(end)
1022         else:
1023             self.end = datetime.datetime.max.date()
1024         if self.start > self.end:
1025             raise ValueError('Date range: "%s" , the start date must be before the end date' % self)
1026
1027     @classmethod
1028     def day(cls, day):
1029         """Returns a range that only contains the given day"""
1030         return cls(day, day)
1031
1032     def __contains__(self, date):
1033         """Check if the date is in the range"""
1034         if not isinstance(date, datetime.date):
1035             date = date_from_str(date)
1036         return self.start <= date <= self.end
1037
1038     def __str__(self):
1039         return '%s - %s' % (self.start.isoformat(), self.end.isoformat())
1040
1041
1042 def platform_name():
1043     """ Returns the platform name as a compat_str """
1044     res = platform.platform()
1045     if isinstance(res, bytes):
1046         res = res.decode(preferredencoding())
1047
1048     assert isinstance(res, compat_str)
1049     return res
1050
1051
1052 def _windows_write_string(s, out):
1053     """ Returns True if the string was written using special methods,
1054     False if it has yet to be written out."""
1055     # Adapted from http://stackoverflow.com/a/3259271/35070
1056
1057     import ctypes
1058     import ctypes.wintypes
1059
1060     WIN_OUTPUT_IDS = {
1061         1: -11,
1062         2: -12,
1063     }
1064
1065     try:
1066         fileno = out.fileno()
1067     except AttributeError:
1068         # If the output stream doesn't have a fileno, it's virtual
1069         return False
1070     except io.UnsupportedOperation:
1071         # Some strange Windows pseudo files?
1072         return False
1073     if fileno not in WIN_OUTPUT_IDS:
1074         return False
1075
1076     GetStdHandle = ctypes.WINFUNCTYPE(
1077         ctypes.wintypes.HANDLE, ctypes.wintypes.DWORD)(
1078         (b'GetStdHandle', ctypes.windll.kernel32))
1079     h = GetStdHandle(WIN_OUTPUT_IDS[fileno])
1080
1081     WriteConsoleW = ctypes.WINFUNCTYPE(
1082         ctypes.wintypes.BOOL, ctypes.wintypes.HANDLE, ctypes.wintypes.LPWSTR,
1083         ctypes.wintypes.DWORD, ctypes.POINTER(ctypes.wintypes.DWORD),
1084         ctypes.wintypes.LPVOID)((b'WriteConsoleW', ctypes.windll.kernel32))
1085     written = ctypes.wintypes.DWORD(0)
1086
1087     GetFileType = ctypes.WINFUNCTYPE(ctypes.wintypes.DWORD, ctypes.wintypes.DWORD)((b'GetFileType', ctypes.windll.kernel32))
1088     FILE_TYPE_CHAR = 0x0002
1089     FILE_TYPE_REMOTE = 0x8000
1090     GetConsoleMode = ctypes.WINFUNCTYPE(
1091         ctypes.wintypes.BOOL, ctypes.wintypes.HANDLE,
1092         ctypes.POINTER(ctypes.wintypes.DWORD))(
1093         (b'GetConsoleMode', ctypes.windll.kernel32))
1094     INVALID_HANDLE_VALUE = ctypes.wintypes.DWORD(-1).value
1095
1096     def not_a_console(handle):
1097         if handle == INVALID_HANDLE_VALUE or handle is None:
1098             return True
1099         return ((GetFileType(handle) & ~FILE_TYPE_REMOTE) != FILE_TYPE_CHAR or
1100                 GetConsoleMode(handle, ctypes.byref(ctypes.wintypes.DWORD())) == 0)
1101
1102     if not_a_console(h):
1103         return False
1104
1105     def next_nonbmp_pos(s):
1106         try:
1107             return next(i for i, c in enumerate(s) if ord(c) > 0xffff)
1108         except StopIteration:
1109             return len(s)
1110
1111     while s:
1112         count = min(next_nonbmp_pos(s), 1024)
1113
1114         ret = WriteConsoleW(
1115             h, s, count if count else 2, ctypes.byref(written), None)
1116         if ret == 0:
1117             raise OSError('Failed to write string')
1118         if not count:  # We just wrote a non-BMP character
1119             assert written.value == 2
1120             s = s[1:]
1121         else:
1122             assert written.value > 0
1123             s = s[written.value:]
1124     return True
1125
1126
1127 def write_string(s, out=None, encoding=None):
1128     if out is None:
1129         out = sys.stderr
1130     assert type(s) == compat_str
1131
1132     if sys.platform == 'win32' and encoding is None and hasattr(out, 'fileno'):
1133         if _windows_write_string(s, out):
1134             return
1135
1136     if ('b' in getattr(out, 'mode', '') or
1137             sys.version_info[0] < 3):  # Python 2 lies about mode of sys.stderr
1138         byt = s.encode(encoding or preferredencoding(), 'ignore')
1139         out.write(byt)
1140     elif hasattr(out, 'buffer'):
1141         enc = encoding or getattr(out, 'encoding', None) or preferredencoding()
1142         byt = s.encode(enc, 'ignore')
1143         out.buffer.write(byt)
1144     else:
1145         out.write(s)
1146     out.flush()
1147
1148
1149 def bytes_to_intlist(bs):
1150     if not bs:
1151         return []
1152     if isinstance(bs[0], int):  # Python 3
1153         return list(bs)
1154     else:
1155         return [ord(c) for c in bs]
1156
1157
1158 def intlist_to_bytes(xs):
1159     if not xs:
1160         return b''
1161     return struct_pack('%dB' % len(xs), *xs)
1162
1163
1164 # Cross-platform file locking
1165 if sys.platform == 'win32':
1166     import ctypes.wintypes
1167     import msvcrt
1168
1169     class OVERLAPPED(ctypes.Structure):
1170         _fields_ = [
1171             ('Internal', ctypes.wintypes.LPVOID),
1172             ('InternalHigh', ctypes.wintypes.LPVOID),
1173             ('Offset', ctypes.wintypes.DWORD),
1174             ('OffsetHigh', ctypes.wintypes.DWORD),
1175             ('hEvent', ctypes.wintypes.HANDLE),
1176         ]
1177
1178     kernel32 = ctypes.windll.kernel32
1179     LockFileEx = kernel32.LockFileEx
1180     LockFileEx.argtypes = [
1181         ctypes.wintypes.HANDLE,     # hFile
1182         ctypes.wintypes.DWORD,      # dwFlags
1183         ctypes.wintypes.DWORD,      # dwReserved
1184         ctypes.wintypes.DWORD,      # nNumberOfBytesToLockLow
1185         ctypes.wintypes.DWORD,      # nNumberOfBytesToLockHigh
1186         ctypes.POINTER(OVERLAPPED)  # Overlapped
1187     ]
1188     LockFileEx.restype = ctypes.wintypes.BOOL
1189     UnlockFileEx = kernel32.UnlockFileEx
1190     UnlockFileEx.argtypes = [
1191         ctypes.wintypes.HANDLE,     # hFile
1192         ctypes.wintypes.DWORD,      # dwReserved
1193         ctypes.wintypes.DWORD,      # nNumberOfBytesToLockLow
1194         ctypes.wintypes.DWORD,      # nNumberOfBytesToLockHigh
1195         ctypes.POINTER(OVERLAPPED)  # Overlapped
1196     ]
1197     UnlockFileEx.restype = ctypes.wintypes.BOOL
1198     whole_low = 0xffffffff
1199     whole_high = 0x7fffffff
1200
1201     def _lock_file(f, exclusive):
1202         overlapped = OVERLAPPED()
1203         overlapped.Offset = 0
1204         overlapped.OffsetHigh = 0
1205         overlapped.hEvent = 0
1206         f._lock_file_overlapped_p = ctypes.pointer(overlapped)
1207         handle = msvcrt.get_osfhandle(f.fileno())
1208         if not LockFileEx(handle, 0x2 if exclusive else 0x0, 0,
1209                           whole_low, whole_high, f._lock_file_overlapped_p):
1210             raise OSError('Locking file failed: %r' % ctypes.FormatError())
1211
1212     def _unlock_file(f):
1213         assert f._lock_file_overlapped_p
1214         handle = msvcrt.get_osfhandle(f.fileno())
1215         if not UnlockFileEx(handle, 0,
1216                             whole_low, whole_high, f._lock_file_overlapped_p):
1217             raise OSError('Unlocking file failed: %r' % ctypes.FormatError())
1218
1219 else:
1220     import fcntl
1221
1222     def _lock_file(f, exclusive):
1223         fcntl.flock(f, fcntl.LOCK_EX if exclusive else fcntl.LOCK_SH)
1224
1225     def _unlock_file(f):
1226         fcntl.flock(f, fcntl.LOCK_UN)
1227
1228
1229 class locked_file(object):
1230     def __init__(self, filename, mode, encoding=None):
1231         assert mode in ['r', 'a', 'w']
1232         self.f = io.open(filename, mode, encoding=encoding)
1233         self.mode = mode
1234
1235     def __enter__(self):
1236         exclusive = self.mode != 'r'
1237         try:
1238             _lock_file(self.f, exclusive)
1239         except IOError:
1240             self.f.close()
1241             raise
1242         return self
1243
1244     def __exit__(self, etype, value, traceback):
1245         try:
1246             _unlock_file(self.f)
1247         finally:
1248             self.f.close()
1249
1250     def __iter__(self):
1251         return iter(self.f)
1252
1253     def write(self, *args):
1254         return self.f.write(*args)
1255
1256     def read(self, *args):
1257         return self.f.read(*args)
1258
1259
1260 def get_filesystem_encoding():
1261     encoding = sys.getfilesystemencoding()
1262     return encoding if encoding is not None else 'utf-8'
1263
1264
1265 def shell_quote(args):
1266     quoted_args = []
1267     encoding = get_filesystem_encoding()
1268     for a in args:
1269         if isinstance(a, bytes):
1270             # We may get a filename encoded with 'encodeFilename'
1271             a = a.decode(encoding)
1272         quoted_args.append(pipes.quote(a))
1273     return ' '.join(quoted_args)
1274
1275
1276 def smuggle_url(url, data):
1277     """ Pass additional data in a URL for internal use. """
1278
1279     sdata = compat_urllib_parse.urlencode(
1280         {'__youtubedl_smuggle': json.dumps(data)})
1281     return url + '#' + sdata
1282
1283
1284 def unsmuggle_url(smug_url, default=None):
1285     if '#__youtubedl_smuggle' not in smug_url:
1286         return smug_url, default
1287     url, _, sdata = smug_url.rpartition('#')
1288     jsond = compat_parse_qs(sdata)['__youtubedl_smuggle'][0]
1289     data = json.loads(jsond)
1290     return url, data
1291
1292
1293 def format_bytes(bytes):
1294     if bytes is None:
1295         return 'N/A'
1296     if type(bytes) is str:
1297         bytes = float(bytes)
1298     if bytes == 0.0:
1299         exponent = 0
1300     else:
1301         exponent = int(math.log(bytes, 1024.0))
1302     suffix = ['B', 'KiB', 'MiB', 'GiB', 'TiB', 'PiB', 'EiB', 'ZiB', 'YiB'][exponent]
1303     converted = float(bytes) / float(1024 ** exponent)
1304     return '%.2f%s' % (converted, suffix)
1305
1306
1307 def parse_filesize(s):
1308     if s is None:
1309         return None
1310
1311     # The lower-case forms are of course incorrect and unofficial,
1312     # but we support those too
1313     _UNIT_TABLE = {
1314         'B': 1,
1315         'b': 1,
1316         'KiB': 1024,
1317         'KB': 1000,
1318         'kB': 1024,
1319         'Kb': 1000,
1320         'MiB': 1024 ** 2,
1321         'MB': 1000 ** 2,
1322         'mB': 1024 ** 2,
1323         'Mb': 1000 ** 2,
1324         'GiB': 1024 ** 3,
1325         'GB': 1000 ** 3,
1326         'gB': 1024 ** 3,
1327         'Gb': 1000 ** 3,
1328         'TiB': 1024 ** 4,
1329         'TB': 1000 ** 4,
1330         'tB': 1024 ** 4,
1331         'Tb': 1000 ** 4,
1332         'PiB': 1024 ** 5,
1333         'PB': 1000 ** 5,
1334         'pB': 1024 ** 5,
1335         'Pb': 1000 ** 5,
1336         'EiB': 1024 ** 6,
1337         'EB': 1000 ** 6,
1338         'eB': 1024 ** 6,
1339         'Eb': 1000 ** 6,
1340         'ZiB': 1024 ** 7,
1341         'ZB': 1000 ** 7,
1342         'zB': 1024 ** 7,
1343         'Zb': 1000 ** 7,
1344         'YiB': 1024 ** 8,
1345         'YB': 1000 ** 8,
1346         'yB': 1024 ** 8,
1347         'Yb': 1000 ** 8,
1348     }
1349
1350     units_re = '|'.join(re.escape(u) for u in _UNIT_TABLE)
1351     m = re.match(
1352         r'(?P<num>[0-9]+(?:[,.][0-9]*)?)\s*(?P<unit>%s)' % units_re, s)
1353     if not m:
1354         return None
1355
1356     num_str = m.group('num').replace(',', '.')
1357     mult = _UNIT_TABLE[m.group('unit')]
1358     return int(float(num_str) * mult)
1359
1360
1361 def month_by_name(name):
1362     """ Return the number of a month by (locale-independently) English name """
1363
1364     try:
1365         return ENGLISH_MONTH_NAMES.index(name) + 1
1366     except ValueError:
1367         return None
1368
1369
1370 def month_by_abbreviation(abbrev):
1371     """ Return the number of a month by (locale-independently) English
1372         abbreviations """
1373
1374     try:
1375         return [s[:3] for s in ENGLISH_MONTH_NAMES].index(abbrev) + 1
1376     except ValueError:
1377         return None
1378
1379
1380 def fix_xml_ampersands(xml_str):
1381     """Replace all the '&' by '&amp;' in XML"""
1382     return re.sub(
1383         r'&(?!amp;|lt;|gt;|apos;|quot;|#x[0-9a-fA-F]{,4};|#[0-9]{,4};)',
1384         '&amp;',
1385         xml_str)
1386
1387
1388 def setproctitle(title):
1389     assert isinstance(title, compat_str)
1390     try:
1391         libc = ctypes.cdll.LoadLibrary('libc.so.6')
1392     except OSError:
1393         return
1394     title_bytes = title.encode('utf-8')
1395     buf = ctypes.create_string_buffer(len(title_bytes))
1396     buf.value = title_bytes
1397     try:
1398         libc.prctl(15, buf, 0, 0, 0)
1399     except AttributeError:
1400         return  # Strange libc, just skip this
1401
1402
1403 def remove_start(s, start):
1404     if s.startswith(start):
1405         return s[len(start):]
1406     return s
1407
1408
1409 def remove_end(s, end):
1410     if s.endswith(end):
1411         return s[:-len(end)]
1412     return s
1413
1414
1415 def remove_quotes(s):
1416     if s is None or len(s) < 2:
1417         return s
1418     for quote in ('"', "'", ):
1419         if s[0] == quote and s[-1] == quote:
1420             return s[1:-1]
1421     return s
1422
1423
1424 def url_basename(url):
1425     path = compat_urlparse.urlparse(url).path
1426     return path.strip('/').split('/')[-1]
1427
1428
1429 class HEADRequest(compat_urllib_request.Request):
1430     def get_method(self):
1431         return 'HEAD'
1432
1433
1434 def int_or_none(v, scale=1, default=None, get_attr=None, invscale=1):
1435     if get_attr:
1436         if v is not None:
1437             v = getattr(v, get_attr, None)
1438     if v == '':
1439         v = None
1440     if v is None:
1441         return default
1442     try:
1443         return int(v) * invscale // scale
1444     except ValueError:
1445         return default
1446
1447
1448 def str_or_none(v, default=None):
1449     return default if v is None else compat_str(v)
1450
1451
1452 def str_to_int(int_str):
1453     """ A more relaxed version of int_or_none """
1454     if int_str is None:
1455         return None
1456     int_str = re.sub(r'[,\.\+]', '', int_str)
1457     return int(int_str)
1458
1459
1460 def float_or_none(v, scale=1, invscale=1, default=None):
1461     if v is None:
1462         return default
1463     try:
1464         return float(v) * invscale / scale
1465     except ValueError:
1466         return default
1467
1468
1469 def parse_duration(s):
1470     if not isinstance(s, compat_basestring):
1471         return None
1472
1473     s = s.strip()
1474
1475     m = re.match(
1476         r'''(?ix)(?:P?T)?
1477         (?:
1478             (?P<only_mins>[0-9.]+)\s*(?:mins?\.?|minutes?)\s*|
1479             (?P<only_hours>[0-9.]+)\s*(?:hours?)|
1480
1481             \s*(?P<hours_reversed>[0-9]+)\s*(?:[:h]|hours?)\s*(?P<mins_reversed>[0-9]+)\s*(?:[:m]|mins?\.?|minutes?)\s*|
1482             (?:
1483                 (?:
1484                     (?:(?P<days>[0-9]+)\s*(?:[:d]|days?)\s*)?
1485                     (?P<hours>[0-9]+)\s*(?:[:h]|hours?)\s*
1486                 )?
1487                 (?P<mins>[0-9]+)\s*(?:[:m]|mins?|minutes?)\s*
1488             )?
1489             (?P<secs>[0-9]+)(?P<ms>\.[0-9]+)?\s*(?:s|secs?|seconds?)?
1490         )$''', s)
1491     if not m:
1492         return None
1493     res = 0
1494     if m.group('only_mins'):
1495         return float_or_none(m.group('only_mins'), invscale=60)
1496     if m.group('only_hours'):
1497         return float_or_none(m.group('only_hours'), invscale=60 * 60)
1498     if m.group('secs'):
1499         res += int(m.group('secs'))
1500     if m.group('mins_reversed'):
1501         res += int(m.group('mins_reversed')) * 60
1502     if m.group('mins'):
1503         res += int(m.group('mins')) * 60
1504     if m.group('hours'):
1505         res += int(m.group('hours')) * 60 * 60
1506     if m.group('hours_reversed'):
1507         res += int(m.group('hours_reversed')) * 60 * 60
1508     if m.group('days'):
1509         res += int(m.group('days')) * 24 * 60 * 60
1510     if m.group('ms'):
1511         res += float(m.group('ms'))
1512     return res
1513
1514
1515 def prepend_extension(filename, ext, expected_real_ext=None):
1516     name, real_ext = os.path.splitext(filename)
1517     return (
1518         '{0}.{1}{2}'.format(name, ext, real_ext)
1519         if not expected_real_ext or real_ext[1:] == expected_real_ext
1520         else '{0}.{1}'.format(filename, ext))
1521
1522
1523 def replace_extension(filename, ext, expected_real_ext=None):
1524     name, real_ext = os.path.splitext(filename)
1525     return '{0}.{1}'.format(
1526         name if not expected_real_ext or real_ext[1:] == expected_real_ext else filename,
1527         ext)
1528
1529
1530 def check_executable(exe, args=[]):
1531     """ Checks if the given binary is installed somewhere in PATH, and returns its name.
1532     args can be a list of arguments for a short output (like -version) """
1533     try:
1534         subprocess.Popen([exe] + args, stdout=subprocess.PIPE, stderr=subprocess.PIPE).communicate()
1535     except OSError:
1536         return False
1537     return exe
1538
1539
1540 def get_exe_version(exe, args=['--version'],
1541                     version_re=None, unrecognized='present'):
1542     """ Returns the version of the specified executable,
1543     or False if the executable is not present """
1544     try:
1545         out, _ = subprocess.Popen(
1546             [encodeArgument(exe)] + args,
1547             stdout=subprocess.PIPE, stderr=subprocess.STDOUT).communicate()
1548     except OSError:
1549         return False
1550     if isinstance(out, bytes):  # Python 2.x
1551         out = out.decode('ascii', 'ignore')
1552     return detect_exe_version(out, version_re, unrecognized)
1553
1554
1555 def detect_exe_version(output, version_re=None, unrecognized='present'):
1556     assert isinstance(output, compat_str)
1557     if version_re is None:
1558         version_re = r'version\s+([-0-9._a-zA-Z]+)'
1559     m = re.search(version_re, output)
1560     if m:
1561         return m.group(1)
1562     else:
1563         return unrecognized
1564
1565
1566 class PagedList(object):
1567     def __len__(self):
1568         # This is only useful for tests
1569         return len(self.getslice())
1570
1571
1572 class OnDemandPagedList(PagedList):
1573     def __init__(self, pagefunc, pagesize):
1574         self._pagefunc = pagefunc
1575         self._pagesize = pagesize
1576
1577     def getslice(self, start=0, end=None):
1578         res = []
1579         for pagenum in itertools.count(start // self._pagesize):
1580             firstid = pagenum * self._pagesize
1581             nextfirstid = pagenum * self._pagesize + self._pagesize
1582             if start >= nextfirstid:
1583                 continue
1584
1585             page_results = list(self._pagefunc(pagenum))
1586
1587             startv = (
1588                 start % self._pagesize
1589                 if firstid <= start < nextfirstid
1590                 else 0)
1591
1592             endv = (
1593                 ((end - 1) % self._pagesize) + 1
1594                 if (end is not None and firstid <= end <= nextfirstid)
1595                 else None)
1596
1597             if startv != 0 or endv is not None:
1598                 page_results = page_results[startv:endv]
1599             res.extend(page_results)
1600
1601             # A little optimization - if current page is not "full", ie. does
1602             # not contain page_size videos then we can assume that this page
1603             # is the last one - there are no more ids on further pages -
1604             # i.e. no need to query again.
1605             if len(page_results) + startv < self._pagesize:
1606                 break
1607
1608             # If we got the whole page, but the next page is not interesting,
1609             # break out early as well
1610             if end == nextfirstid:
1611                 break
1612         return res
1613
1614
1615 class InAdvancePagedList(PagedList):
1616     def __init__(self, pagefunc, pagecount, pagesize):
1617         self._pagefunc = pagefunc
1618         self._pagecount = pagecount
1619         self._pagesize = pagesize
1620
1621     def getslice(self, start=0, end=None):
1622         res = []
1623         start_page = start // self._pagesize
1624         end_page = (
1625             self._pagecount if end is None else (end // self._pagesize + 1))
1626         skip_elems = start - start_page * self._pagesize
1627         only_more = None if end is None else end - start
1628         for pagenum in range(start_page, end_page):
1629             page = list(self._pagefunc(pagenum))
1630             if skip_elems:
1631                 page = page[skip_elems:]
1632                 skip_elems = None
1633             if only_more is not None:
1634                 if len(page) < only_more:
1635                     only_more -= len(page)
1636                 else:
1637                     page = page[:only_more]
1638                     res.extend(page)
1639                     break
1640             res.extend(page)
1641         return res
1642
1643
1644 def uppercase_escape(s):
1645     unicode_escape = codecs.getdecoder('unicode_escape')
1646     return re.sub(
1647         r'\\U[0-9a-fA-F]{8}',
1648         lambda m: unicode_escape(m.group(0))[0],
1649         s)
1650
1651
1652 def lowercase_escape(s):
1653     unicode_escape = codecs.getdecoder('unicode_escape')
1654     return re.sub(
1655         r'\\u[0-9a-fA-F]{4}',
1656         lambda m: unicode_escape(m.group(0))[0],
1657         s)
1658
1659
1660 def escape_rfc3986(s):
1661     """Escape non-ASCII characters as suggested by RFC 3986"""
1662     if sys.version_info < (3, 0) and isinstance(s, compat_str):
1663         s = s.encode('utf-8')
1664     return compat_urllib_parse.quote(s, b"%/;:@&=+$,!~*'()?#[]")
1665
1666
1667 def escape_url(url):
1668     """Escape URL as suggested by RFC 3986"""
1669     url_parsed = compat_urllib_parse_urlparse(url)
1670     return url_parsed._replace(
1671         path=escape_rfc3986(url_parsed.path),
1672         params=escape_rfc3986(url_parsed.params),
1673         query=escape_rfc3986(url_parsed.query),
1674         fragment=escape_rfc3986(url_parsed.fragment)
1675     ).geturl()
1676
1677 try:
1678     struct.pack('!I', 0)
1679 except TypeError:
1680     # In Python 2.6 (and some 2.7 versions), struct requires a bytes argument
1681     def struct_pack(spec, *args):
1682         if isinstance(spec, compat_str):
1683             spec = spec.encode('ascii')
1684         return struct.pack(spec, *args)
1685
1686     def struct_unpack(spec, *args):
1687         if isinstance(spec, compat_str):
1688             spec = spec.encode('ascii')
1689         return struct.unpack(spec, *args)
1690 else:
1691     struct_pack = struct.pack
1692     struct_unpack = struct.unpack
1693
1694
1695 def read_batch_urls(batch_fd):
1696     def fixup(url):
1697         if not isinstance(url, compat_str):
1698             url = url.decode('utf-8', 'replace')
1699         BOM_UTF8 = '\xef\xbb\xbf'
1700         if url.startswith(BOM_UTF8):
1701             url = url[len(BOM_UTF8):]
1702         url = url.strip()
1703         if url.startswith(('#', ';', ']')):
1704             return False
1705         return url
1706
1707     with contextlib.closing(batch_fd) as fd:
1708         return [url for url in map(fixup, fd) if url]
1709
1710
1711 def urlencode_postdata(*args, **kargs):
1712     return compat_urllib_parse.urlencode(*args, **kargs).encode('ascii')
1713
1714
1715 def encode_dict(d, encoding='utf-8'):
1716     def encode(v):
1717         return v.encode(encoding) if isinstance(v, compat_basestring) else v
1718     return dict((encode(k), encode(v)) for k, v in d.items())
1719
1720
1721 def dict_get(d, key_or_keys, default=None, skip_false_values=True):
1722     if isinstance(key_or_keys, (list, tuple)):
1723         for key in key_or_keys:
1724             if key not in d or d[key] is None or skip_false_values and not d[key]:
1725                 continue
1726             return d[key]
1727         return default
1728     return d.get(key_or_keys, default)
1729
1730
1731 def encode_compat_str(string, encoding=preferredencoding(), errors='strict'):
1732     return string if isinstance(string, compat_str) else compat_str(string, encoding, errors)
1733
1734
1735 US_RATINGS = {
1736     'G': 0,
1737     'PG': 10,
1738     'PG-13': 13,
1739     'R': 16,
1740     'NC': 18,
1741 }
1742
1743
1744 def parse_age_limit(s):
1745     if s is None:
1746         return None
1747     m = re.match(r'^(?P<age>\d{1,2})\+?$', s)
1748     return int(m.group('age')) if m else US_RATINGS.get(s)
1749
1750
1751 def strip_jsonp(code):
1752     return re.sub(
1753         r'(?s)^[a-zA-Z0-9_.]+\s*\(\s*(.*)\);?\s*?(?://[^\n]*)*$', r'\1', code)
1754
1755
1756 def js_to_json(code):
1757     def fix_kv(m):
1758         v = m.group(0)
1759         if v in ('true', 'false', 'null'):
1760             return v
1761         if v.startswith('"'):
1762             v = re.sub(r"\\'", "'", v[1:-1])
1763         elif v.startswith("'"):
1764             v = v[1:-1]
1765             v = re.sub(r"\\\\|\\'|\"", lambda m: {
1766                 '\\\\': '\\\\',
1767                 "\\'": "'",
1768                 '"': '\\"',
1769             }[m.group(0)], v)
1770         return '"%s"' % v
1771
1772     res = re.sub(r'''(?x)
1773         "(?:[^"\\]*(?:\\\\|\\['"nu]))*[^"\\]*"|
1774         '(?:[^'\\]*(?:\\\\|\\['"nu]))*[^'\\]*'|
1775         [a-zA-Z_][.a-zA-Z_0-9]*
1776         ''', fix_kv, code)
1777     res = re.sub(r',(\s*[\]}])', lambda m: m.group(1), res)
1778     return res
1779
1780
1781 def qualities(quality_ids):
1782     """ Get a numeric quality value out of a list of possible values """
1783     def q(qid):
1784         try:
1785             return quality_ids.index(qid)
1786         except ValueError:
1787             return -1
1788     return q
1789
1790
1791 DEFAULT_OUTTMPL = '%(title)s-%(id)s.%(ext)s'
1792
1793
1794 def limit_length(s, length):
1795     """ Add ellipses to overly long strings """
1796     if s is None:
1797         return None
1798     ELLIPSES = '...'
1799     if len(s) > length:
1800         return s[:length - len(ELLIPSES)] + ELLIPSES
1801     return s
1802
1803
1804 def version_tuple(v):
1805     return tuple(int(e) for e in re.split(r'[-.]', v))
1806
1807
1808 def is_outdated_version(version, limit, assume_new=True):
1809     if not version:
1810         return not assume_new
1811     try:
1812         return version_tuple(version) < version_tuple(limit)
1813     except ValueError:
1814         return not assume_new
1815
1816
1817 def ytdl_is_updateable():
1818     """ Returns if youtube-dl can be updated with -U """
1819     from zipimport import zipimporter
1820
1821     return isinstance(globals().get('__loader__'), zipimporter) or hasattr(sys, 'frozen')
1822
1823
1824 def args_to_str(args):
1825     # Get a short string representation for a subprocess command
1826     return ' '.join(shlex_quote(a) for a in args)
1827
1828
1829 def error_to_compat_str(err):
1830     err_str = str(err)
1831     # On python 2 error byte string must be decoded with proper
1832     # encoding rather than ascii
1833     if sys.version_info[0] < 3:
1834         err_str = err_str.decode(preferredencoding())
1835     return err_str
1836
1837
1838 def mimetype2ext(mt):
1839     ext = {
1840         'audio/mp4': 'm4a',
1841     }.get(mt)
1842     if ext is not None:
1843         return ext
1844
1845     _, _, res = mt.rpartition('/')
1846
1847     return {
1848         '3gpp': '3gp',
1849         'smptett+xml': 'tt',
1850         'srt': 'srt',
1851         'ttaf+xml': 'dfxp',
1852         'ttml+xml': 'ttml',
1853         'vtt': 'vtt',
1854         'x-flv': 'flv',
1855         'x-mp4-fragmented': 'mp4',
1856         'x-ms-wmv': 'wmv',
1857     }.get(res, res)
1858
1859
1860 def urlhandle_detect_ext(url_handle):
1861     try:
1862         url_handle.headers
1863         getheader = lambda h: url_handle.headers[h]
1864     except AttributeError:  # Python < 3
1865         getheader = url_handle.info().getheader
1866
1867     cd = getheader('Content-Disposition')
1868     if cd:
1869         m = re.match(r'attachment;\s*filename="(?P<filename>[^"]+)"', cd)
1870         if m:
1871             e = determine_ext(m.group('filename'), default_ext=None)
1872             if e:
1873                 return e
1874
1875     return mimetype2ext(getheader('Content-Type'))
1876
1877
1878 def encode_data_uri(data, mime_type):
1879     return 'data:%s;base64,%s' % (mime_type, base64.b64encode(data).decode('ascii'))
1880
1881
1882 def age_restricted(content_limit, age_limit):
1883     """ Returns True iff the content should be blocked """
1884
1885     if age_limit is None:  # No limit set
1886         return False
1887     if content_limit is None:
1888         return False  # Content available for everyone
1889     return age_limit < content_limit
1890
1891
1892 def is_html(first_bytes):
1893     """ Detect whether a file contains HTML by examining its first bytes. """
1894
1895     BOMS = [
1896         (b'\xef\xbb\xbf', 'utf-8'),
1897         (b'\x00\x00\xfe\xff', 'utf-32-be'),
1898         (b'\xff\xfe\x00\x00', 'utf-32-le'),
1899         (b'\xff\xfe', 'utf-16-le'),
1900         (b'\xfe\xff', 'utf-16-be'),
1901     ]
1902     for bom, enc in BOMS:
1903         if first_bytes.startswith(bom):
1904             s = first_bytes[len(bom):].decode(enc, 'replace')
1905             break
1906     else:
1907         s = first_bytes.decode('utf-8', 'replace')
1908
1909     return re.match(r'^\s*<', s)
1910
1911
1912 def determine_protocol(info_dict):
1913     protocol = info_dict.get('protocol')
1914     if protocol is not None:
1915         return protocol
1916
1917     url = info_dict['url']
1918     if url.startswith('rtmp'):
1919         return 'rtmp'
1920     elif url.startswith('mms'):
1921         return 'mms'
1922     elif url.startswith('rtsp'):
1923         return 'rtsp'
1924
1925     ext = determine_ext(url)
1926     if ext == 'm3u8':
1927         return 'm3u8'
1928     elif ext == 'f4m':
1929         return 'f4m'
1930
1931     return compat_urllib_parse_urlparse(url).scheme
1932
1933
1934 def render_table(header_row, data):
1935     """ Render a list of rows, each as a list of values """
1936     table = [header_row] + data
1937     max_lens = [max(len(compat_str(v)) for v in col) for col in zip(*table)]
1938     format_str = ' '.join('%-' + compat_str(ml + 1) + 's' for ml in max_lens[:-1]) + '%s'
1939     return '\n'.join(format_str % tuple(row) for row in table)
1940
1941
1942 def _match_one(filter_part, dct):
1943     COMPARISON_OPERATORS = {
1944         '<': operator.lt,
1945         '<=': operator.le,
1946         '>': operator.gt,
1947         '>=': operator.ge,
1948         '=': operator.eq,
1949         '!=': operator.ne,
1950     }
1951     operator_rex = re.compile(r'''(?x)\s*
1952         (?P<key>[a-z_]+)
1953         \s*(?P<op>%s)(?P<none_inclusive>\s*\?)?\s*
1954         (?:
1955             (?P<intval>[0-9.]+(?:[kKmMgGtTpPeEzZyY]i?[Bb]?)?)|
1956             (?P<strval>(?![0-9.])[a-z0-9A-Z]*)
1957         )
1958         \s*$
1959         ''' % '|'.join(map(re.escape, COMPARISON_OPERATORS.keys())))
1960     m = operator_rex.search(filter_part)
1961     if m:
1962         op = COMPARISON_OPERATORS[m.group('op')]
1963         if m.group('strval') is not None:
1964             if m.group('op') not in ('=', '!='):
1965                 raise ValueError(
1966                     'Operator %s does not support string values!' % m.group('op'))
1967             comparison_value = m.group('strval')
1968         else:
1969             try:
1970                 comparison_value = int(m.group('intval'))
1971             except ValueError:
1972                 comparison_value = parse_filesize(m.group('intval'))
1973                 if comparison_value is None:
1974                     comparison_value = parse_filesize(m.group('intval') + 'B')
1975                 if comparison_value is None:
1976                     raise ValueError(
1977                         'Invalid integer value %r in filter part %r' % (
1978                             m.group('intval'), filter_part))
1979         actual_value = dct.get(m.group('key'))
1980         if actual_value is None:
1981             return m.group('none_inclusive')
1982         return op(actual_value, comparison_value)
1983
1984     UNARY_OPERATORS = {
1985         '': lambda v: v is not None,
1986         '!': lambda v: v is None,
1987     }
1988     operator_rex = re.compile(r'''(?x)\s*
1989         (?P<op>%s)\s*(?P<key>[a-z_]+)
1990         \s*$
1991         ''' % '|'.join(map(re.escape, UNARY_OPERATORS.keys())))
1992     m = operator_rex.search(filter_part)
1993     if m:
1994         op = UNARY_OPERATORS[m.group('op')]
1995         actual_value = dct.get(m.group('key'))
1996         return op(actual_value)
1997
1998     raise ValueError('Invalid filter part %r' % filter_part)
1999
2000
2001 def match_str(filter_str, dct):
2002     """ Filter a dictionary with a simple string syntax. Returns True (=passes filter) or false """
2003
2004     return all(
2005         _match_one(filter_part, dct) for filter_part in filter_str.split('&'))
2006
2007
2008 def match_filter_func(filter_str):
2009     def _match_func(info_dict):
2010         if match_str(filter_str, info_dict):
2011             return None
2012         else:
2013             video_title = info_dict.get('title', info_dict.get('id', 'video'))
2014             return '%s does not pass filter %s, skipping ..' % (video_title, filter_str)
2015     return _match_func
2016
2017
2018 def parse_dfxp_time_expr(time_expr):
2019     if not time_expr:
2020         return
2021
2022     mobj = re.match(r'^(?P<time_offset>\d+(?:\.\d+)?)s?$', time_expr)
2023     if mobj:
2024         return float(mobj.group('time_offset'))
2025
2026     mobj = re.match(r'^(\d+):(\d\d):(\d\d(?:(?:\.|:)\d+)?)$', time_expr)
2027     if mobj:
2028         return 3600 * int(mobj.group(1)) + 60 * int(mobj.group(2)) + float(mobj.group(3).replace(':', '.'))
2029
2030
2031 def srt_subtitles_timecode(seconds):
2032     return '%02d:%02d:%02d,%03d' % (seconds / 3600, (seconds % 3600) / 60, seconds % 60, (seconds % 1) * 1000)
2033
2034
2035 def dfxp2srt(dfxp_data):
2036     _x = functools.partial(xpath_with_ns, ns_map={
2037         'ttml': 'http://www.w3.org/ns/ttml',
2038         'ttaf1': 'http://www.w3.org/2006/10/ttaf1',
2039     })
2040
2041     class TTMLPElementParser(object):
2042         out = ''
2043
2044         def start(self, tag, attrib):
2045             if tag in (_x('ttml:br'), _x('ttaf1:br'), 'br'):
2046                 self.out += '\n'
2047
2048         def end(self, tag):
2049             pass
2050
2051         def data(self, data):
2052             self.out += data
2053
2054         def close(self):
2055             return self.out.strip()
2056
2057     def parse_node(node):
2058         target = TTMLPElementParser()
2059         parser = xml.etree.ElementTree.XMLParser(target=target)
2060         parser.feed(xml.etree.ElementTree.tostring(node))
2061         return parser.close()
2062
2063     dfxp = compat_etree_fromstring(dfxp_data.encode('utf-8'))
2064     out = []
2065     paras = dfxp.findall(_x('.//ttml:p')) or dfxp.findall(_x('.//ttaf1:p')) or dfxp.findall('.//p')
2066
2067     if not paras:
2068         raise ValueError('Invalid dfxp/TTML subtitle')
2069
2070     for para, index in zip(paras, itertools.count(1)):
2071         begin_time = parse_dfxp_time_expr(para.attrib.get('begin'))
2072         end_time = parse_dfxp_time_expr(para.attrib.get('end'))
2073         dur = parse_dfxp_time_expr(para.attrib.get('dur'))
2074         if begin_time is None:
2075             continue
2076         if not end_time:
2077             if not dur:
2078                 continue
2079             end_time = begin_time + dur
2080         out.append('%d\n%s --> %s\n%s\n\n' % (
2081             index,
2082             srt_subtitles_timecode(begin_time),
2083             srt_subtitles_timecode(end_time),
2084             parse_node(para)))
2085
2086     return ''.join(out)
2087
2088
2089 def cli_option(params, command_option, param):
2090     param = params.get(param)
2091     return [command_option, param] if param is not None else []
2092
2093
2094 def cli_bool_option(params, command_option, param, true_value='true', false_value='false', separator=None):
2095     param = params.get(param)
2096     assert isinstance(param, bool)
2097     if separator:
2098         return [command_option + separator + (true_value if param else false_value)]
2099     return [command_option, true_value if param else false_value]
2100
2101
2102 def cli_valueless_option(params, command_option, param, expected_value=True):
2103     param = params.get(param)
2104     return [command_option] if param == expected_value else []
2105
2106
2107 def cli_configuration_args(params, param, default=[]):
2108     ex_args = params.get(param)
2109     if ex_args is None:
2110         return default
2111     assert isinstance(ex_args, list)
2112     return ex_args
2113
2114
2115 class ISO639Utils(object):
2116     # See http://www.loc.gov/standards/iso639-2/ISO-639-2_utf-8.txt
2117     _lang_map = {
2118         'aa': 'aar',
2119         'ab': 'abk',
2120         'ae': 'ave',
2121         'af': 'afr',
2122         'ak': 'aka',
2123         'am': 'amh',
2124         'an': 'arg',
2125         'ar': 'ara',
2126         'as': 'asm',
2127         'av': 'ava',
2128         'ay': 'aym',
2129         'az': 'aze',
2130         'ba': 'bak',
2131         'be': 'bel',
2132         'bg': 'bul',
2133         'bh': 'bih',
2134         'bi': 'bis',
2135         'bm': 'bam',
2136         'bn': 'ben',
2137         'bo': 'bod',
2138         'br': 'bre',
2139         'bs': 'bos',
2140         'ca': 'cat',
2141         'ce': 'che',
2142         'ch': 'cha',
2143         'co': 'cos',
2144         'cr': 'cre',
2145         'cs': 'ces',
2146         'cu': 'chu',
2147         'cv': 'chv',
2148         'cy': 'cym',
2149         'da': 'dan',
2150         'de': 'deu',
2151         'dv': 'div',
2152         'dz': 'dzo',
2153         'ee': 'ewe',
2154         'el': 'ell',
2155         'en': 'eng',
2156         'eo': 'epo',
2157         'es': 'spa',
2158         'et': 'est',
2159         'eu': 'eus',
2160         'fa': 'fas',
2161         'ff': 'ful',
2162         'fi': 'fin',
2163         'fj': 'fij',
2164         'fo': 'fao',
2165         'fr': 'fra',
2166         'fy': 'fry',
2167         'ga': 'gle',
2168         'gd': 'gla',
2169         'gl': 'glg',
2170         'gn': 'grn',
2171         'gu': 'guj',
2172         'gv': 'glv',
2173         'ha': 'hau',
2174         'he': 'heb',
2175         'hi': 'hin',
2176         'ho': 'hmo',
2177         'hr': 'hrv',
2178         'ht': 'hat',
2179         'hu': 'hun',
2180         'hy': 'hye',
2181         'hz': 'her',
2182         'ia': 'ina',
2183         'id': 'ind',
2184         'ie': 'ile',
2185         'ig': 'ibo',
2186         'ii': 'iii',
2187         'ik': 'ipk',
2188         'io': 'ido',
2189         'is': 'isl',
2190         'it': 'ita',
2191         'iu': 'iku',
2192         'ja': 'jpn',
2193         'jv': 'jav',
2194         'ka': 'kat',
2195         'kg': 'kon',
2196         'ki': 'kik',
2197         'kj': 'kua',
2198         'kk': 'kaz',
2199         'kl': 'kal',
2200         'km': 'khm',
2201         'kn': 'kan',
2202         'ko': 'kor',
2203         'kr': 'kau',
2204         'ks': 'kas',
2205         'ku': 'kur',
2206         'kv': 'kom',
2207         'kw': 'cor',
2208         'ky': 'kir',
2209         'la': 'lat',
2210         'lb': 'ltz',
2211         'lg': 'lug',
2212         'li': 'lim',
2213         'ln': 'lin',
2214         'lo': 'lao',
2215         'lt': 'lit',
2216         'lu': 'lub',
2217         'lv': 'lav',
2218         'mg': 'mlg',
2219         'mh': 'mah',
2220         'mi': 'mri',
2221         'mk': 'mkd',
2222         'ml': 'mal',
2223         'mn': 'mon',
2224         'mr': 'mar',
2225         'ms': 'msa',
2226         'mt': 'mlt',
2227         'my': 'mya',
2228         'na': 'nau',
2229         'nb': 'nob',
2230         'nd': 'nde',
2231         'ne': 'nep',
2232         'ng': 'ndo',
2233         'nl': 'nld',
2234         'nn': 'nno',
2235         'no': 'nor',
2236         'nr': 'nbl',
2237         'nv': 'nav',
2238         'ny': 'nya',
2239         'oc': 'oci',
2240         'oj': 'oji',
2241         'om': 'orm',
2242         'or': 'ori',
2243         'os': 'oss',
2244         'pa': 'pan',
2245         'pi': 'pli',
2246         'pl': 'pol',
2247         'ps': 'pus',
2248         'pt': 'por',
2249         'qu': 'que',
2250         'rm': 'roh',
2251         'rn': 'run',
2252         'ro': 'ron',
2253         'ru': 'rus',
2254         'rw': 'kin',
2255         'sa': 'san',
2256         'sc': 'srd',
2257         'sd': 'snd',
2258         'se': 'sme',
2259         'sg': 'sag',
2260         'si': 'sin',
2261         'sk': 'slk',
2262         'sl': 'slv',
2263         'sm': 'smo',
2264         'sn': 'sna',
2265         'so': 'som',
2266         'sq': 'sqi',
2267         'sr': 'srp',
2268         'ss': 'ssw',
2269         'st': 'sot',
2270         'su': 'sun',
2271         'sv': 'swe',
2272         'sw': 'swa',
2273         'ta': 'tam',
2274         'te': 'tel',
2275         'tg': 'tgk',
2276         'th': 'tha',
2277         'ti': 'tir',
2278         'tk': 'tuk',
2279         'tl': 'tgl',
2280         'tn': 'tsn',
2281         'to': 'ton',
2282         'tr': 'tur',
2283         'ts': 'tso',
2284         'tt': 'tat',
2285         'tw': 'twi',
2286         'ty': 'tah',
2287         'ug': 'uig',
2288         'uk': 'ukr',
2289         'ur': 'urd',
2290         'uz': 'uzb',
2291         've': 'ven',
2292         'vi': 'vie',
2293         'vo': 'vol',
2294         'wa': 'wln',
2295         'wo': 'wol',
2296         'xh': 'xho',
2297         'yi': 'yid',
2298         'yo': 'yor',
2299         'za': 'zha',
2300         'zh': 'zho',
2301         'zu': 'zul',
2302     }
2303
2304     @classmethod
2305     def short2long(cls, code):
2306         """Convert language code from ISO 639-1 to ISO 639-2/T"""
2307         return cls._lang_map.get(code[:2])
2308
2309     @classmethod
2310     def long2short(cls, code):
2311         """Convert language code from ISO 639-2/T to ISO 639-1"""
2312         for short_name, long_name in cls._lang_map.items():
2313             if long_name == code:
2314                 return short_name
2315
2316
2317 class ISO3166Utils(object):
2318     # From http://data.okfn.org/data/core/country-list
2319     _country_map = {
2320         'AF': 'Afghanistan',
2321         'AX': 'Åland Islands',
2322         'AL': 'Albania',
2323         'DZ': 'Algeria',
2324         'AS': 'American Samoa',
2325         'AD': 'Andorra',
2326         'AO': 'Angola',
2327         'AI': 'Anguilla',
2328         'AQ': 'Antarctica',
2329         'AG': 'Antigua and Barbuda',
2330         'AR': 'Argentina',
2331         'AM': 'Armenia',
2332         'AW': 'Aruba',
2333         'AU': 'Australia',
2334         'AT': 'Austria',
2335         'AZ': 'Azerbaijan',
2336         'BS': 'Bahamas',
2337         'BH': 'Bahrain',
2338         'BD': 'Bangladesh',
2339         'BB': 'Barbados',
2340         'BY': 'Belarus',
2341         'BE': 'Belgium',
2342         'BZ': 'Belize',
2343         'BJ': 'Benin',
2344         'BM': 'Bermuda',
2345         'BT': 'Bhutan',
2346         'BO': 'Bolivia, Plurinational State of',
2347         'BQ': 'Bonaire, Sint Eustatius and Saba',
2348         'BA': 'Bosnia and Herzegovina',
2349         'BW': 'Botswana',
2350         'BV': 'Bouvet Island',
2351         'BR': 'Brazil',
2352         'IO': 'British Indian Ocean Territory',
2353         'BN': 'Brunei Darussalam',
2354         'BG': 'Bulgaria',
2355         'BF': 'Burkina Faso',
2356         'BI': 'Burundi',
2357         'KH': 'Cambodia',
2358         'CM': 'Cameroon',
2359         'CA': 'Canada',
2360         'CV': 'Cape Verde',
2361         'KY': 'Cayman Islands',
2362         'CF': 'Central African Republic',
2363         'TD': 'Chad',
2364         'CL': 'Chile',
2365         'CN': 'China',
2366         'CX': 'Christmas Island',
2367         'CC': 'Cocos (Keeling) Islands',
2368         'CO': 'Colombia',
2369         'KM': 'Comoros',
2370         'CG': 'Congo',
2371         'CD': 'Congo, the Democratic Republic of the',
2372         'CK': 'Cook Islands',
2373         'CR': 'Costa Rica',
2374         'CI': 'Côte d\'Ivoire',
2375         'HR': 'Croatia',
2376         'CU': 'Cuba',
2377         'CW': 'Curaçao',
2378         'CY': 'Cyprus',
2379         'CZ': 'Czech Republic',
2380         'DK': 'Denmark',
2381         'DJ': 'Djibouti',
2382         'DM': 'Dominica',
2383         'DO': 'Dominican Republic',
2384         'EC': 'Ecuador',
2385         'EG': 'Egypt',
2386         'SV': 'El Salvador',
2387         'GQ': 'Equatorial Guinea',
2388         'ER': 'Eritrea',
2389         'EE': 'Estonia',
2390         'ET': 'Ethiopia',
2391         'FK': 'Falkland Islands (Malvinas)',
2392         'FO': 'Faroe Islands',
2393         'FJ': 'Fiji',
2394         'FI': 'Finland',
2395         'FR': 'France',
2396         'GF': 'French Guiana',
2397         'PF': 'French Polynesia',
2398         'TF': 'French Southern Territories',
2399         'GA': 'Gabon',
2400         'GM': 'Gambia',
2401         'GE': 'Georgia',
2402         'DE': 'Germany',
2403         'GH': 'Ghana',
2404         'GI': 'Gibraltar',
2405         'GR': 'Greece',
2406         'GL': 'Greenland',
2407         'GD': 'Grenada',
2408         'GP': 'Guadeloupe',
2409         'GU': 'Guam',
2410         'GT': 'Guatemala',
2411         'GG': 'Guernsey',
2412         'GN': 'Guinea',
2413         'GW': 'Guinea-Bissau',
2414         'GY': 'Guyana',
2415         'HT': 'Haiti',
2416         'HM': 'Heard Island and McDonald Islands',
2417         'VA': 'Holy See (Vatican City State)',
2418         'HN': 'Honduras',
2419         'HK': 'Hong Kong',
2420         'HU': 'Hungary',
2421         'IS': 'Iceland',
2422         'IN': 'India',
2423         'ID': 'Indonesia',
2424         'IR': 'Iran, Islamic Republic of',
2425         'IQ': 'Iraq',
2426         'IE': 'Ireland',
2427         'IM': 'Isle of Man',
2428         'IL': 'Israel',
2429         'IT': 'Italy',
2430         'JM': 'Jamaica',
2431         'JP': 'Japan',
2432         'JE': 'Jersey',
2433         'JO': 'Jordan',
2434         'KZ': 'Kazakhstan',
2435         'KE': 'Kenya',
2436         'KI': 'Kiribati',
2437         'KP': 'Korea, Democratic People\'s Republic of',
2438         'KR': 'Korea, Republic of',
2439         'KW': 'Kuwait',
2440         'KG': 'Kyrgyzstan',
2441         'LA': 'Lao People\'s Democratic Republic',
2442         'LV': 'Latvia',
2443         'LB': 'Lebanon',
2444         'LS': 'Lesotho',
2445         'LR': 'Liberia',
2446         'LY': 'Libya',
2447         'LI': 'Liechtenstein',
2448         'LT': 'Lithuania',
2449         'LU': 'Luxembourg',
2450         'MO': 'Macao',
2451         'MK': 'Macedonia, the Former Yugoslav Republic of',
2452         'MG': 'Madagascar',
2453         'MW': 'Malawi',
2454         'MY': 'Malaysia',
2455         'MV': 'Maldives',
2456         'ML': 'Mali',
2457         'MT': 'Malta',
2458         'MH': 'Marshall Islands',
2459         'MQ': 'Martinique',
2460         'MR': 'Mauritania',
2461         'MU': 'Mauritius',
2462         'YT': 'Mayotte',
2463         'MX': 'Mexico',
2464         'FM': 'Micronesia, Federated States of',
2465         'MD': 'Moldova, Republic of',
2466         'MC': 'Monaco',
2467         'MN': 'Mongolia',
2468         'ME': 'Montenegro',
2469         'MS': 'Montserrat',
2470         'MA': 'Morocco',
2471         'MZ': 'Mozambique',
2472         'MM': 'Myanmar',
2473         'NA': 'Namibia',
2474         'NR': 'Nauru',
2475         'NP': 'Nepal',
2476         'NL': 'Netherlands',
2477         'NC': 'New Caledonia',
2478         'NZ': 'New Zealand',
2479         'NI': 'Nicaragua',
2480         'NE': 'Niger',
2481         'NG': 'Nigeria',
2482         'NU': 'Niue',
2483         'NF': 'Norfolk Island',
2484         'MP': 'Northern Mariana Islands',
2485         'NO': 'Norway',
2486         'OM': 'Oman',
2487         'PK': 'Pakistan',
2488         'PW': 'Palau',
2489         'PS': 'Palestine, State of',
2490         'PA': 'Panama',
2491         'PG': 'Papua New Guinea',
2492         'PY': 'Paraguay',
2493         'PE': 'Peru',
2494         'PH': 'Philippines',
2495         'PN': 'Pitcairn',
2496         'PL': 'Poland',
2497         'PT': 'Portugal',
2498         'PR': 'Puerto Rico',
2499         'QA': 'Qatar',
2500         'RE': 'Réunion',
2501         'RO': 'Romania',
2502         'RU': 'Russian Federation',
2503         'RW': 'Rwanda',
2504         'BL': 'Saint Barthélemy',
2505         'SH': 'Saint Helena, Ascension and Tristan da Cunha',
2506         'KN': 'Saint Kitts and Nevis',
2507         'LC': 'Saint Lucia',
2508         'MF': 'Saint Martin (French part)',
2509         'PM': 'Saint Pierre and Miquelon',
2510         'VC': 'Saint Vincent and the Grenadines',
2511         'WS': 'Samoa',
2512         'SM': 'San Marino',
2513         'ST': 'Sao Tome and Principe',
2514         'SA': 'Saudi Arabia',
2515         'SN': 'Senegal',
2516         'RS': 'Serbia',
2517         'SC': 'Seychelles',
2518         'SL': 'Sierra Leone',
2519         'SG': 'Singapore',
2520         'SX': 'Sint Maarten (Dutch part)',
2521         'SK': 'Slovakia',
2522         'SI': 'Slovenia',
2523         'SB': 'Solomon Islands',
2524         'SO': 'Somalia',
2525         'ZA': 'South Africa',
2526         'GS': 'South Georgia and the South Sandwich Islands',
2527         'SS': 'South Sudan',
2528         'ES': 'Spain',
2529         'LK': 'Sri Lanka',
2530         'SD': 'Sudan',
2531         'SR': 'Suriname',
2532         'SJ': 'Svalbard and Jan Mayen',
2533         'SZ': 'Swaziland',
2534         'SE': 'Sweden',
2535         'CH': 'Switzerland',
2536         'SY': 'Syrian Arab Republic',
2537         'TW': 'Taiwan, Province of China',
2538         'TJ': 'Tajikistan',
2539         'TZ': 'Tanzania, United Republic of',
2540         'TH': 'Thailand',
2541         'TL': 'Timor-Leste',
2542         'TG': 'Togo',
2543         'TK': 'Tokelau',
2544         'TO': 'Tonga',
2545         'TT': 'Trinidad and Tobago',
2546         'TN': 'Tunisia',
2547         'TR': 'Turkey',
2548         'TM': 'Turkmenistan',
2549         'TC': 'Turks and Caicos Islands',
2550         'TV': 'Tuvalu',
2551         'UG': 'Uganda',
2552         'UA': 'Ukraine',
2553         'AE': 'United Arab Emirates',
2554         'GB': 'United Kingdom',
2555         'US': 'United States',
2556         'UM': 'United States Minor Outlying Islands',
2557         'UY': 'Uruguay',
2558         'UZ': 'Uzbekistan',
2559         'VU': 'Vanuatu',
2560         'VE': 'Venezuela, Bolivarian Republic of',
2561         'VN': 'Viet Nam',
2562         'VG': 'Virgin Islands, British',
2563         'VI': 'Virgin Islands, U.S.',
2564         'WF': 'Wallis and Futuna',
2565         'EH': 'Western Sahara',
2566         'YE': 'Yemen',
2567         'ZM': 'Zambia',
2568         'ZW': 'Zimbabwe',
2569     }
2570
2571     @classmethod
2572     def short2full(cls, code):
2573         """Convert an ISO 3166-2 country code to the corresponding full name"""
2574         return cls._country_map.get(code.upper())
2575
2576
2577 class PerRequestProxyHandler(compat_urllib_request.ProxyHandler):
2578     def __init__(self, proxies=None):
2579         # Set default handlers
2580         for type in ('http', 'https'):
2581             setattr(self, '%s_open' % type,
2582                     lambda r, proxy='__noproxy__', type=type, meth=self.proxy_open:
2583                         meth(r, proxy, type))
2584         return compat_urllib_request.ProxyHandler.__init__(self, proxies)
2585
2586     def proxy_open(self, req, proxy, type):
2587         req_proxy = req.headers.get('Ytdl-request-proxy')
2588         if req_proxy is not None:
2589             proxy = req_proxy
2590             del req.headers['Ytdl-request-proxy']
2591
2592         if proxy == '__noproxy__':
2593             return None  # No Proxy
2594         return compat_urllib_request.ProxyHandler.proxy_open(
2595             self, req, proxy, type)
2596
2597
2598 def ohdave_rsa_encrypt(data, exponent, modulus):
2599     '''
2600     Implement OHDave's RSA algorithm. See http://www.ohdave.com/rsa/
2601
2602     Input:
2603         data: data to encrypt, bytes-like object
2604         exponent, modulus: parameter e and N of RSA algorithm, both integer
2605     Output: hex string of encrypted data
2606
2607     Limitation: supports one block encryption only
2608     '''
2609
2610     payload = int(binascii.hexlify(data[::-1]), 16)
2611     encrypted = pow(payload, exponent, modulus)
2612     return '%x' % encrypted