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

   1 #!/usr/bin/env python
   2 # -*- coding: utf-8 -*-
   3
   4 import calendar
   5 import codecs
   6 import contextlib
   7 import ctypes
   8 import datetime
   9 import email.utils
  10 import errno
  11 import getpass
  12 import gzip
  13 import itertools
  14 import io
  15 import json
  16 import locale
  17 import math
  18 import os
  19 import pipes
  20 import platform
  21 import re
  22 import ssl
  23 import socket
  24 import struct
  25 import subprocess
  26 import sys
  27 import traceback
  28 import xml.etree.ElementTree
  29 import zlib
  30
  31 try:
  32     import urllib.request as compat_urllib_request
  33 except ImportError: # Python 2
  34     import urllib2 as compat_urllib_request
  35
  36 try:
  37     import urllib.error as compat_urllib_error
  38 except ImportError: # Python 2
  39     import urllib2 as compat_urllib_error
  40
  41 try:
  42     import urllib.parse as compat_urllib_parse
  43 except ImportError: # Python 2
  44     import urllib as compat_urllib_parse
  45
  46 try:
  47     from urllib.parse import urlparse as compat_urllib_parse_urlparse
  48 except ImportError: # Python 2
  49     from urlparse import urlparse as compat_urllib_parse_urlparse
  50
  51 try:
  52     import urllib.parse as compat_urlparse
  53 except ImportError: # Python 2
  54     import urlparse as compat_urlparse
  55
  56 try:
  57     import http.cookiejar as compat_cookiejar
  58 except ImportError: # Python 2
  59     import cookielib as compat_cookiejar
  60
  61 try:
  62     import html.entities as compat_html_entities
  63 except ImportError: # Python 2
  64     import htmlentitydefs as compat_html_entities
  65
  66 try:
  67     import html.parser as compat_html_parser
  68 except ImportError: # Python 2
  69     import HTMLParser as compat_html_parser
  70
  71 try:
  72     import http.client as compat_http_client
  73 except ImportError: # Python 2
  74     import httplib as compat_http_client
  75
  76 try:
  77     from urllib.error import HTTPError as compat_HTTPError
  78 except ImportError:  # Python 2
  79     from urllib2 import HTTPError as compat_HTTPError
  80
  81 try:
  82     from urllib.request import urlretrieve as compat_urlretrieve
  83 except ImportError:  # Python 2
  84     from urllib import urlretrieve as compat_urlretrieve
  85
  86
  87 try:
  88     from subprocess import DEVNULL
  89     compat_subprocess_get_DEVNULL = lambda: DEVNULL
  90 except ImportError:
  91     compat_subprocess_get_DEVNULL = lambda: open(os.path.devnull, 'w')
  92
  93 try:
  94     from urllib.parse import parse_qs as compat_parse_qs
  95 except ImportError: # Python 2
  96     # HACK: The following is the correct parse_qs implementation from cpython 3's stdlib.
  97     # Python 2's version is apparently totally broken
  98     def _unquote(string, encoding='utf-8', errors='replace'):
  99         if string == '':
 100             return string
 101         res = string.split('%')
 102         if len(res) == 1:
 103             return string
 104         if encoding is None:
 105             encoding = 'utf-8'
 106         if errors is None:
 107             errors = 'replace'
 108         # pct_sequence: contiguous sequence of percent-encoded bytes, decoded
 109         pct_sequence = b''
 110         string = res[0]
 111         for item in res[1:]:
 112             try:
 113                 if not item:
 114                     raise ValueError
 115                 pct_sequence += item[:2].decode('hex')
 116                 rest = item[2:]
 117                 if not rest:
 118                     # This segment was just a single percent-encoded character.
 119                     # May be part of a sequence of code units, so delay decoding.
 120                     # (Stored in pct_sequence).
 121                     continue
 122             except ValueError:
 123                 rest = '%' + item
 124             # Encountered non-percent-encoded characters. Flush the current
 125             # pct_sequence.
 126             string += pct_sequence.decode(encoding, errors) + rest
 127             pct_sequence = b''
 128         if pct_sequence:
 129             # Flush the final pct_sequence
 130             string += pct_sequence.decode(encoding, errors)
 131         return string
 132
 133     def _parse_qsl(qs, keep_blank_values=False, strict_parsing=False,
 134                 encoding='utf-8', errors='replace'):
 135         qs, _coerce_result = qs, unicode
 136         pairs = [s2 for s1 in qs.split('&') for s2 in s1.split(';')]
 137         r = []
 138         for name_value in pairs:
 139             if not name_value and not strict_parsing:
 140                 continue
 141             nv = name_value.split('=', 1)
 142             if len(nv) != 2:
 143                 if strict_parsing:
 144                     raise ValueError("bad query field: %r" % (name_value,))
 145                 # Handle case of a control-name with no equal sign
 146                 if keep_blank_values:
 147                     nv.append('')
 148                 else:
 149                     continue
 150             if len(nv[1]) or keep_blank_values:
 151                 name = nv[0].replace('+', ' ')
 152                 name = _unquote(name, encoding=encoding, errors=errors)
 153                 name = _coerce_result(name)
 154                 value = nv[1].replace('+', ' ')
 155                 value = _unquote(value, encoding=encoding, errors=errors)
 156                 value = _coerce_result(value)
 157                 r.append((name, value))
 158         return r
 159
 160     def compat_parse_qs(qs, keep_blank_values=False, strict_parsing=False,
 161                 encoding='utf-8', errors='replace'):
 162         parsed_result = {}
 163         pairs = _parse_qsl(qs, keep_blank_values, strict_parsing,
 164                         encoding=encoding, errors=errors)
 165         for name, value in pairs:
 166             if name in parsed_result:
 167                 parsed_result[name].append(value)
 168             else:
 169                 parsed_result[name] = [value]
 170         return parsed_result
 171
 172 try:
 173     compat_str = unicode # Python 2
 174 except NameError:
 175     compat_str = str
 176
 177 try:
 178     compat_chr = unichr # Python 2
 179 except NameError:
 180     compat_chr = chr
 181
 182 try:
 183     from xml.etree.ElementTree import ParseError as compat_xml_parse_error
 184 except ImportError:  # Python 2.6
 185     from xml.parsers.expat import ExpatError as compat_xml_parse_error
 186
 187 def compat_ord(c):
 188     if type(c) is int: return c
 189     else: return ord(c)
 190
 191 # This is not clearly defined otherwise
 192 compiled_regex_type = type(re.compile(''))
 193
 194 std_headers = {
 195     'User-Agent': 'Mozilla/5.0 (X11; Linux x86_64; rv:10.0) Gecko/20100101 Firefox/10.0 (Chrome)',
 196     'Accept-Charset': 'ISO-8859-1,utf-8;q=0.7,*;q=0.7',
 197     'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
 198     'Accept-Encoding': 'gzip, deflate',
 199     'Accept-Language': 'en-us,en;q=0.5',
 200 }
 201
 202 def preferredencoding():
 203     """Get preferred encoding.
 204
 205     Returns the best encoding scheme for the system, based on
 206     locale.getpreferredencoding() and some further tweaks.
 207     """
 208     try:
 209         pref = locale.getpreferredencoding()
 210         u'TEST'.encode(pref)
 211     except:
 212         pref = 'UTF-8'
 213
 214     return pref
 215
 216 if sys.version_info < (3,0):
 217     def compat_print(s):
 218         print(s.encode(preferredencoding(), 'xmlcharrefreplace'))
 219 else:
 220     def compat_print(s):
 221         assert type(s) == type(u'')
 222         print(s)
 223
 224 # In Python 2.x, json.dump expects a bytestream.
 225 # In Python 3.x, it writes to a character stream
 226 if sys.version_info < (3,0):
 227     def write_json_file(obj, fn):
 228         with open(fn, 'wb') as f:
 229             json.dump(obj, f)
 230 else:
 231     def write_json_file(obj, fn):
 232         with open(fn, 'w', encoding='utf-8') as f:
 233             json.dump(obj, f)
 234
 235 if sys.version_info >= (2,7):
 236     def find_xpath_attr(node, xpath, key, val):
 237         """ Find the xpath xpath[@key=val] """
 238         assert re.match(r'^[a-zA-Z]+$', key)
 239         assert re.match(r'^[a-zA-Z0-9@\s:._]*$', val)
 240         expr = xpath + u"[@%s='%s']" % (key, val)
 241         return node.find(expr)
 242 else:
 243     def find_xpath_attr(node, xpath, key, val):
 244         for f in node.findall(xpath):
 245             if f.attrib.get(key) == val:
 246                 return f
 247         return None
 248
 249 # On python2.6 the xml.etree.ElementTree.Element methods don't support
 250 # the namespace parameter
 251 def xpath_with_ns(path, ns_map):
 252     components = [c.split(':') for c in path.split('/')]
 253     replaced = []
 254     for c in components:
 255         if len(c) == 1:
 256             replaced.append(c[0])
 257         else:
 258             ns, tag = c
 259             replaced.append('{%s}%s' % (ns_map[ns], tag))
 260     return '/'.join(replaced)
 261
 262 def htmlentity_transform(matchobj):
 263     """Transforms an HTML entity to a character.
 264
 265     This function receives a match object and is intended to be used with
 266     the re.sub() function.
 267     """
 268     entity = matchobj.group(1)
 269
 270     # Known non-numeric HTML entity
 271     if entity in compat_html_entities.name2codepoint:
 272         return compat_chr(compat_html_entities.name2codepoint[entity])
 273
 274     mobj = re.match(u'(?u)#(x?\\d+)', entity)
 275     if mobj is not None:
 276         numstr = mobj.group(1)
 277         if numstr.startswith(u'x'):
 278             base = 16
 279             numstr = u'0%s' % numstr
 280         else:
 281             base = 10
 282         return compat_chr(int(numstr, base))
 283
 284     # Unknown entity in name, return its literal representation
 285     return (u'&%s;' % entity)
 286
 287 compat_html_parser.locatestarttagend = re.compile(r"""<[a-zA-Z][-.a-zA-Z0-9:_]*(?:\s+(?:(?<=['"\s])[^\s/>][^\s/=>]*(?:\s*=+\s*(?:'[^']*'|"[^"]*"|(?!['"])[^>\s]*))?\s*)*)?\s*""", re.VERBOSE) # backport bugfix
 288 class BaseHTMLParser(compat_html_parser.HTMLParser):
 289     def __init(self):
 290         compat_html_parser.HTMLParser.__init__(self)
 291         self.html = None
 292
 293     def loads(self, html):
 294         self.html = html
 295         self.feed(html)
 296         self.close()
 297
 298 class AttrParser(BaseHTMLParser):
 299     """Modified HTMLParser that isolates a tag with the specified attribute"""
 300     def __init__(self, attribute, value):
 301         self.attribute = attribute
 302         self.value = value
 303         self.result = None
 304         self.started = False
 305         self.depth = {}
 306         self.watch_startpos = False
 307         self.error_count = 0
 308         BaseHTMLParser.__init__(self)
 309
 310     def error(self, message):
 311         if self.error_count > 10 or self.started:
 312             raise compat_html_parser.HTMLParseError(message, self.getpos())
 313         self.rawdata = '\n'.join(self.html.split('\n')[self.getpos()[0]:]) # skip one line
 314         self.error_count += 1
 315         self.goahead(1)
 316
 317     def handle_starttag(self, tag, attrs):
 318         attrs = dict(attrs)
 319         if self.started:
 320             self.find_startpos(None)
 321         if self.attribute in attrs and attrs[self.attribute] == self.value:
 322             self.result = [tag]
 323             self.started = True
 324             self.watch_startpos = True
 325         if self.started:
 326             if not tag in self.depth: self.depth[tag] = 0
 327             self.depth[tag] += 1
 328
 329     def handle_endtag(self, tag):
 330         if self.started:
 331             if tag in self.depth: self.depth[tag] -= 1
 332             if self.depth[self.result[0]] == 0:
 333                 self.started = False
 334                 self.result.append(self.getpos())
 335
 336     def find_startpos(self, x):
 337         """Needed to put the start position of the result (self.result[1])
 338         after the opening tag with the requested id"""
 339         if self.watch_startpos:
 340             self.watch_startpos = False
 341             self.result.append(self.getpos())
 342     handle_entityref = handle_charref = handle_data = handle_comment = \
 343     handle_decl = handle_pi = unknown_decl = find_startpos
 344
 345     def get_result(self):
 346         if self.result is None:
 347             return None
 348         if len(self.result) != 3:
 349             return None
 350         lines = self.html.split('\n')
 351         lines = lines[self.result[1][0]-1:self.result[2][0]]
 352         lines[0] = lines[0][self.result[1][1]:]
 353         if len(lines) == 1:
 354             lines[-1] = lines[-1][:self.result[2][1]-self.result[1][1]]
 355         lines[-1] = lines[-1][:self.result[2][1]]
 356         return '\n'.join(lines).strip()
 357 # Hack for https://github.com/rg3/youtube-dl/issues/662
 358 if sys.version_info < (2, 7, 3):
 359     AttrParser.parse_endtag = (lambda self, i:
 360         i + len("</scr'+'ipt>")
 361         if self.rawdata[i:].startswith("</scr'+'ipt>")
 362         else compat_html_parser.HTMLParser.parse_endtag(self, i))
 363
 364 def get_element_by_id(id, html):
 365     """Return the content of the tag with the specified ID in the passed HTML document"""
 366     return get_element_by_attribute("id", id, html)
 367
 368 def get_element_by_attribute(attribute, value, html):
 369     """Return the content of the tag with the specified attribute in the passed HTML document"""
 370     parser = AttrParser(attribute, value)
 371     try:
 372         parser.loads(html)
 373     except compat_html_parser.HTMLParseError:
 374         pass
 375     return parser.get_result()
 376
 377 class MetaParser(BaseHTMLParser):
 378     """
 379     Modified HTMLParser that isolates a meta tag with the specified name
 380     attribute.
 381     """
 382     def __init__(self, name):
 383         BaseHTMLParser.__init__(self)
 384         self.name = name
 385         self.content = None
 386         self.result = None
 387
 388     def handle_starttag(self, tag, attrs):
 389         if tag != 'meta':
 390             return
 391         attrs = dict(attrs)
 392         if attrs.get('name') == self.name:
 393             self.result = attrs.get('content')
 394
 395     def get_result(self):
 396         return self.result
 397
 398 def get_meta_content(name, html):
 399     """
 400     Return the content attribute from the meta tag with the given name attribute.
 401     """
 402     parser = MetaParser(name)
 403     try:
 404         parser.loads(html)
 405     except compat_html_parser.HTMLParseError:
 406         pass
 407     return parser.get_result()
 408
 409
 410 def clean_html(html):
 411     """Clean an HTML snippet into a readable string"""
 412     # Newline vs <br />
 413     html = html.replace('\n', ' ')
 414     html = re.sub(r'\s*<\s*br\s*/?\s*>\s*', '\n', html)
 415     html = re.sub(r'<\s*/\s*p\s*>\s*<\s*p[^>]*>', '\n', html)
 416     # Strip html tags
 417     html = re.sub('<.*?>', '', html)
 418     # Replace html entities
 419     html = unescapeHTML(html)
 420     return html.strip()
 421
 422
 423 def sanitize_open(filename, open_mode):
 424     """Try to open the given filename, and slightly tweak it if this fails.
 425
 426     Attempts to open the given filename. If this fails, it tries to change
 427     the filename slightly, step by step, until it's either able to open it
 428     or it fails and raises a final exception, like the standard open()
 429     function.
 430
 431     It returns the tuple (stream, definitive_file_name).
 432     """
 433     try:
 434         if filename == u'-':
 435             if sys.platform == 'win32':
 436                 import msvcrt
 437                 msvcrt.setmode(sys.stdout.fileno(), os.O_BINARY)
 438             return (sys.stdout.buffer if hasattr(sys.stdout, 'buffer') else sys.stdout, filename)
 439         stream = open(encodeFilename(filename), open_mode)
 440         return (stream, filename)
 441     except (IOError, OSError) as err:
 442         if err.errno in (errno.EACCES,):
 443             raise
 444
 445         # In case of error, try to remove win32 forbidden chars
 446         alt_filename = os.path.join(
 447                         re.sub(u'[/<>:"\\|\\\\?\\*]', u'#', path_part)
 448                         for path_part in os.path.split(filename)
 449                        )
 450         if alt_filename == filename:
 451             raise
 452         else:
 453             # An exception here should be caught in the caller
 454             stream = open(encodeFilename(filename), open_mode)
 455             return (stream, alt_filename)
 456
 457
 458 def timeconvert(timestr):
 459     """Convert RFC 2822 defined time string into system timestamp"""
 460     timestamp = None
 461     timetuple = email.utils.parsedate_tz(timestr)
 462     if timetuple is not None:
 463         timestamp = email.utils.mktime_tz(timetuple)
 464     return timestamp
 465
 466 def sanitize_filename(s, restricted=False, is_id=False):
 467     """Sanitizes a string so it could be used as part of a filename.
 468     If restricted is set, use a stricter subset of allowed characters.
 469     Set is_id if this is not an arbitrary string, but an ID that should be kept if possible
 470     """
 471     def replace_insane(char):
 472         if char == '?' or ord(char) < 32 or ord(char) == 127:
 473             return ''
 474         elif char == '"':
 475             return '' if restricted else '\''
 476         elif char == ':':
 477             return '_-' if restricted else ' -'
 478         elif char in '\\/|*<>':
 479             return '_'
 480         if restricted and (char in '!&\'()[]{}$;`^,#' or char.isspace()):
 481             return '_'
 482         if restricted and ord(char) > 127:
 483             return '_'
 484         return char
 485
 486     result = u''.join(map(replace_insane, s))
 487     if not is_id:
 488         while '__' in result:
 489             result = result.replace('__', '_')
 490         result = result.strip('_')
 491         # Common case of "Foreign band name - English song title"
 492         if restricted and result.startswith('-_'):
 493             result = result[2:]
 494         if not result:
 495             result = '_'
 496     return result
 497
 498 def orderedSet(iterable):
 499     """ Remove all duplicates from the input iterable """
 500     res = []
 501     for el in iterable:
 502         if el not in res:
 503             res.append(el)
 504     return res
 505
 506
 507 def unescapeHTML(s):
 508     if s is None:
 509         return None
 510     assert type(s) == compat_str
 511
 512     result = re.sub(r'(?u)&(.+?);', htmlentity_transform, s)
 513     return result
 514
 515
 516 def encodeFilename(s, for_subprocess=False):
 517     """
 518     @param s The name of the file
 519     """
 520
 521     assert type(s) == compat_str
 522
 523     # Python 3 has a Unicode API
 524     if sys.version_info >= (3, 0):
 525         return s
 526
 527     if sys.platform == 'win32' and sys.getwindowsversion()[0] >= 5:
 528         # Pass u'' directly to use Unicode APIs on Windows 2000 and up
 529         # (Detecting Windows NT 4 is tricky because 'major >= 4' would
 530         # match Windows 9x series as well. Besides, NT 4 is obsolete.)
 531         if not for_subprocess:
 532             return s
 533         else:
 534             # For subprocess calls, encode with locale encoding
 535             # Refer to http://stackoverflow.com/a/9951851/35070
 536             encoding = preferredencoding()
 537     else:
 538         encoding = sys.getfilesystemencoding()
 539     if encoding is None:
 540         encoding = 'utf-8'
 541     return s.encode(encoding, 'ignore')
 542
 543
 544 def encodeArgument(s):
 545     if not isinstance(s, compat_str):
 546         # Legacy code that uses byte strings
 547         # Uncomment the following line after fixing all post processors
 548         #assert False, 'Internal error: %r should be of type %r, is %r' % (s, compat_str, type(s))
 549         s = s.decode('ascii')
 550     return encodeFilename(s, True)
 551
 552
 553 def decodeOption(optval):
 554     if optval is None:
 555         return optval
 556     if isinstance(optval, bytes):
 557         optval = optval.decode(preferredencoding())
 558
 559     assert isinstance(optval, compat_str)
 560     return optval
 561
 562 def formatSeconds(secs):
 563     if secs > 3600:
 564         return '%d:%02d:%02d' % (secs // 3600, (secs % 3600) // 60, secs % 60)
 565     elif secs > 60:
 566         return '%d:%02d' % (secs // 60, secs % 60)
 567     else:
 568         return '%d' % secs
 569
 570
 571 def make_HTTPS_handler(opts_no_check_certificate, **kwargs):
 572     if sys.version_info < (3, 2):
 573         import httplib
 574
 575         class HTTPSConnectionV3(httplib.HTTPSConnection):
 576             def __init__(self, *args, **kwargs):
 577                 httplib.HTTPSConnection.__init__(self, *args, **kwargs)
 578
 579             def connect(self):
 580                 sock = socket.create_connection((self.host, self.port), self.timeout)
 581                 if getattr(self, '_tunnel_host', False):
 582                     self.sock = sock
 583                     self._tunnel()
 584                 try:
 585                     self.sock = ssl.wrap_socket(sock, self.key_file, self.cert_file, ssl_version=ssl.PROTOCOL_SSLv3)
 586                 except ssl.SSLError:
 587                     self.sock = ssl.wrap_socket(sock, self.key_file, self.cert_file, ssl_version=ssl.PROTOCOL_SSLv23)
 588
 589         class HTTPSHandlerV3(compat_urllib_request.HTTPSHandler):
 590             def https_open(self, req):
 591                 return self.do_open(HTTPSConnectionV3, req)
 592         return HTTPSHandlerV3(**kwargs)
 593     else:
 594         context = ssl.SSLContext(ssl.PROTOCOL_SSLv3)
 595         context.verify_mode = (ssl.CERT_NONE
 596                                if opts_no_check_certificate
 597                                else ssl.CERT_REQUIRED)
 598         context.set_default_verify_paths()
 599         try:
 600             context.load_default_certs()
 601         except AttributeError:
 602             pass  # Python < 3.4
 603         return compat_urllib_request.HTTPSHandler(context=context, **kwargs)
 604
 605 class ExtractorError(Exception):
 606     """Error during info extraction."""
 607     def __init__(self, msg, tb=None, expected=False, cause=None, video_id=None):
 608         """ tb, if given, is the original traceback (so that it can be printed out).
 609         If expected is set, this is a normal error message and most likely not a bug in youtube-dl.
 610         """
 611
 612         if sys.exc_info()[0] in (compat_urllib_error.URLError, socket.timeout, UnavailableVideoError):
 613             expected = True
 614         if video_id is not None:
 615             msg = video_id + ': ' + msg
 616         if not expected:
 617             msg = msg + u'; please report this issue on https://yt-dl.org/bug . Be sure to call youtube-dl with the --verbose flag and include its complete output. Make sure you are using the latest version; type  youtube-dl -U  to update.'
 618         super(ExtractorError, self).__init__(msg)
 619
 620         self.traceback = tb
 621         self.exc_info = sys.exc_info()  # preserve original exception
 622         self.cause = cause
 623         self.video_id = video_id
 624
 625     def format_traceback(self):
 626         if self.traceback is None:
 627             return None
 628         return u''.join(traceback.format_tb(self.traceback))
 629
 630
 631 class RegexNotFoundError(ExtractorError):
 632     """Error when a regex didn't match"""
 633     pass
 634
 635
 636 class DownloadError(Exception):
 637     """Download Error exception.
 638
 639     This exception may be thrown by FileDownloader objects if they are not
 640     configured to continue on errors. They will contain the appropriate
 641     error message.
 642     """
 643     def __init__(self, msg, exc_info=None):
 644         """ exc_info, if given, is the original exception that caused the trouble (as returned by sys.exc_info()). """
 645         super(DownloadError, self).__init__(msg)
 646         self.exc_info = exc_info
 647
 648
 649 class SameFileError(Exception):
 650     """Same File exception.
 651
 652     This exception will be thrown by FileDownloader objects if they detect
 653     multiple files would have to be downloaded to the same file on disk.
 654     """
 655     pass
 656
 657
 658 class PostProcessingError(Exception):
 659     """Post Processing exception.
 660
 661     This exception may be raised by PostProcessor's .run() method to
 662     indicate an error in the postprocessing task.
 663     """
 664     def __init__(self, msg):
 665         self.msg = msg
 666
 667 class MaxDownloadsReached(Exception):
 668     """ --max-downloads limit has been reached. """
 669     pass
 670
 671
 672 class UnavailableVideoError(Exception):
 673     """Unavailable Format exception.
 674
 675     This exception will be thrown when a video is requested
 676     in a format that is not available for that video.
 677     """
 678     pass
 679
 680
 681 class ContentTooShortError(Exception):
 682     """Content Too Short exception.
 683
 684     This exception may be raised by FileDownloader objects when a file they
 685     download is too small for what the server announced first, indicating
 686     the connection was probably interrupted.
 687     """
 688     # Both in bytes
 689     downloaded = None
 690     expected = None
 691
 692     def __init__(self, downloaded, expected):
 693         self.downloaded = downloaded
 694         self.expected = expected
 695
 696 class YoutubeDLHandler(compat_urllib_request.HTTPHandler):
 697     """Handler for HTTP requests and responses.
 698
 699     This class, when installed with an OpenerDirector, automatically adds
 700     the standard headers to every HTTP request and handles gzipped and
 701     deflated responses from web servers. If compression is to be avoided in
 702     a particular request, the original request in the program code only has
 703     to include the HTTP header "Youtubedl-No-Compression", which will be
 704     removed before making the real request.
 705
 706     Part of this code was copied from:
 707
 708     http://techknack.net/python-urllib2-handlers/
 709
 710     Andrew Rowls, the author of that code, agreed to release it to the
 711     public domain.
 712     """
 713
 714     @staticmethod
 715     def deflate(data):
 716         try:
 717             return zlib.decompress(data, -zlib.MAX_WBITS)
 718         except zlib.error:
 719             return zlib.decompress(data)
 720
 721     @staticmethod
 722     def addinfourl_wrapper(stream, headers, url, code):
 723         if hasattr(compat_urllib_request.addinfourl, 'getcode'):
 724             return compat_urllib_request.addinfourl(stream, headers, url, code)
 725         ret = compat_urllib_request.addinfourl(stream, headers, url)
 726         ret.code = code
 727         return ret
 728
 729     def http_request(self, req):
 730         for h,v in std_headers.items():
 731             if h in req.headers:
 732                 del req.headers[h]
 733             req.add_header(h, v)
 734         if 'Youtubedl-no-compression' in req.headers:
 735             if 'Accept-encoding' in req.headers:
 736                 del req.headers['Accept-encoding']
 737             del req.headers['Youtubedl-no-compression']
 738         if 'Youtubedl-user-agent' in req.headers:
 739             if 'User-agent' in req.headers:
 740                 del req.headers['User-agent']
 741             req.headers['User-agent'] = req.headers['Youtubedl-user-agent']
 742             del req.headers['Youtubedl-user-agent']
 743         return req
 744
 745     def http_response(self, req, resp):
 746         old_resp = resp
 747         # gzip
 748         if resp.headers.get('Content-encoding', '') == 'gzip':
 749             content = resp.read()
 750             gz = gzip.GzipFile(fileobj=io.BytesIO(content), mode='rb')
 751             try:
 752                 uncompressed = io.BytesIO(gz.read())
 753             except IOError as original_ioerror:
 754                 # There may be junk add the end of the file
 755                 # See http://stackoverflow.com/q/4928560/35070 for details
 756                 for i in range(1, 1024):
 757                     try:
 758                         gz = gzip.GzipFile(fileobj=io.BytesIO(content[:-i]), mode='rb')
 759                         uncompressed = io.BytesIO(gz.read())
 760                     except IOError:
 761                         continue
 762                     break
 763                 else:
 764                     raise original_ioerror
 765             resp = self.addinfourl_wrapper(uncompressed, old_resp.headers, old_resp.url, old_resp.code)
 766             resp.msg = old_resp.msg
 767         # deflate
 768         if resp.headers.get('Content-encoding', '') == 'deflate':
 769             gz = io.BytesIO(self.deflate(resp.read()))
 770             resp = self.addinfourl_wrapper(gz, old_resp.headers, old_resp.url, old_resp.code)
 771             resp.msg = old_resp.msg
 772         return resp
 773
 774     https_request = http_request
 775     https_response = http_response
 776
 777
 778 def parse_iso8601(date_str):
 779     """ Return a UNIX timestamp from the given date """
 780
 781     if date_str is None:
 782         return None
 783
 784     m = re.search(
 785         r'Z$| ?(?P<sign>\+|-)(?P<hours>[0-9]{2}):?(?P<minutes>[0-9]{2})$',
 786         date_str)
 787     if not m:
 788         timezone = datetime.timedelta()
 789     else:
 790         date_str = date_str[:-len(m.group(0))]
 791         if not m.group('sign'):
 792             timezone = datetime.timedelta()
 793         else:
 794             sign = 1 if m.group('sign') == '+' else -1
 795             timezone = datetime.timedelta(
 796                 hours=sign * int(m.group('hours')),
 797                 minutes=sign * int(m.group('minutes')))
 798
 799     dt = datetime.datetime.strptime(date_str, '%Y-%m-%dT%H:%M:%S') - timezone
 800     return calendar.timegm(dt.timetuple())
 801
 802
 803 def unified_strdate(date_str):
 804     """Return a string with the date in the format YYYYMMDD"""
 805
 806     if date_str is None:
 807         return None
 808
 809     upload_date = None
 810     #Replace commas
 811     date_str = date_str.replace(',', ' ')
 812     # %z (UTC offset) is only supported in python>=3.2
 813     date_str = re.sub(r' ?(\+|-)[0-9]{2}:?[0-9]{2}$', '', date_str)
 814     format_expressions = [
 815         '%d %B %Y',
 816         '%d %b %Y',
 817         '%B %d %Y',
 818         '%b %d %Y',
 819         '%Y-%m-%d',
 820         '%d.%m.%Y',
 821         '%d/%m/%Y',
 822         '%Y/%m/%d %H:%M:%S',
 823         '%Y-%m-%d %H:%M:%S',
 824         '%d.%m.%Y %H:%M',
 825         '%d.%m.%Y %H.%M',
 826         '%Y-%m-%dT%H:%M:%SZ',
 827         '%Y-%m-%dT%H:%M:%S.%fZ',
 828         '%Y-%m-%dT%H:%M:%S.%f0Z',
 829         '%Y-%m-%dT%H:%M:%S',
 830         '%Y-%m-%dT%H:%M:%S.%f',
 831         '%Y-%m-%dT%H:%M',
 832     ]
 833     for expression in format_expressions:
 834         try:
 835             upload_date = datetime.datetime.strptime(date_str, expression).strftime('%Y%m%d')
 836         except ValueError:
 837             pass
 838     if upload_date is None:
 839         timetuple = email.utils.parsedate_tz(date_str)
 840         if timetuple:
 841             upload_date = datetime.datetime(*timetuple[:6]).strftime('%Y%m%d')
 842     return upload_date
 843
 844 def determine_ext(url, default_ext=u'unknown_video'):
 845     guess = url.partition(u'?')[0].rpartition(u'.')[2]
 846     if re.match(r'^[A-Za-z0-9]+$', guess):
 847         return guess
 848     else:
 849         return default_ext
 850
 851 def subtitles_filename(filename, sub_lang, sub_format):
 852     return filename.rsplit('.', 1)[0] + u'.' + sub_lang + u'.' + sub_format
 853
 854 def date_from_str(date_str):
 855     """
 856     Return a datetime object from a string in the format YYYYMMDD or
 857     (now|today)[+-][0-9](day|week|month|year)(s)?"""
 858     today = datetime.date.today()
 859     if date_str == 'now'or date_str == 'today':
 860         return today
 861     match = re.match('(now|today)(?P<sign>[+-])(?P<time>\d+)(?P<unit>day|week|month|year)(s)?', date_str)
 862     if match is not None:
 863         sign = match.group('sign')
 864         time = int(match.group('time'))
 865         if sign == '-':
 866             time = -time
 867         unit = match.group('unit')
 868         #A bad aproximation?
 869         if unit == 'month':
 870             unit = 'day'
 871             time *= 30
 872         elif unit == 'year':
 873             unit = 'day'
 874             time *= 365
 875         unit += 's'
 876         delta = datetime.timedelta(**{unit: time})
 877         return today + delta
 878     return datetime.datetime.strptime(date_str, "%Y%m%d").date()
 879
 880 def hyphenate_date(date_str):
 881     """
 882     Convert a date in 'YYYYMMDD' format to 'YYYY-MM-DD' format"""
 883     match = re.match(r'^(\d\d\d\d)(\d\d)(\d\d)$', date_str)
 884     if match is not None:
 885         return '-'.join(match.groups())
 886     else:
 887         return date_str
 888
 889 class DateRange(object):
 890     """Represents a time interval between two dates"""
 891     def __init__(self, start=None, end=None):
 892         """start and end must be strings in the format accepted by date"""
 893         if start is not None:
 894             self.start = date_from_str(start)
 895         else:
 896             self.start = datetime.datetime.min.date()
 897         if end is not None:
 898             self.end = date_from_str(end)
 899         else:
 900             self.end = datetime.datetime.max.date()
 901         if self.start > self.end:
 902             raise ValueError('Date range: "%s" , the start date must be before the end date' % self)
 903     @classmethod
 904     def day(cls, day):
 905         """Returns a range that only contains the given day"""
 906         return cls(day,day)
 907     def __contains__(self, date):
 908         """Check if the date is in the range"""
 909         if not isinstance(date, datetime.date):
 910             date = date_from_str(date)
 911         return self.start <= date <= self.end
 912     def __str__(self):
 913         return '%s - %s' % ( self.start.isoformat(), self.end.isoformat())
 914
 915
 916 def platform_name():
 917     """ Returns the platform name as a compat_str """
 918     res = platform.platform()
 919     if isinstance(res, bytes):
 920         res = res.decode(preferredencoding())
 921
 922     assert isinstance(res, compat_str)
 923     return res
 924
 925
 926 def _windows_write_string(s, out):
 927     """ Returns True if the string was written using special methods,
 928     False if it has yet to be written out."""
 929     # Adapted from http://stackoverflow.com/a/3259271/35070
 930
 931     import ctypes
 932     import ctypes.wintypes
 933
 934     WIN_OUTPUT_IDS = {
 935         1: -11,
 936         2: -12,
 937     }
 938
 939     try:
 940         fileno = out.fileno()
 941     except AttributeError:
 942         # If the output stream doesn't have a fileno, it's virtual
 943         return False
 944     if fileno not in WIN_OUTPUT_IDS:
 945         return False
 946
 947     GetStdHandle = ctypes.WINFUNCTYPE(
 948         ctypes.wintypes.HANDLE, ctypes.wintypes.DWORD)(
 949         ("GetStdHandle", ctypes.windll.kernel32))
 950     h = GetStdHandle(WIN_OUTPUT_IDS[fileno])
 951
 952     WriteConsoleW = ctypes.WINFUNCTYPE(
 953         ctypes.wintypes.BOOL, ctypes.wintypes.HANDLE, ctypes.wintypes.LPWSTR,
 954         ctypes.wintypes.DWORD, ctypes.POINTER(ctypes.wintypes.DWORD),
 955         ctypes.wintypes.LPVOID)(("WriteConsoleW", ctypes.windll.kernel32))
 956     written = ctypes.wintypes.DWORD(0)
 957
 958     GetFileType = ctypes.WINFUNCTYPE(ctypes.wintypes.DWORD, ctypes.wintypes.DWORD)(("GetFileType", ctypes.windll.kernel32))
 959     FILE_TYPE_CHAR = 0x0002
 960     FILE_TYPE_REMOTE = 0x8000
 961     GetConsoleMode = ctypes.WINFUNCTYPE(
 962         ctypes.wintypes.BOOL, ctypes.wintypes.HANDLE,
 963         ctypes.POINTER(ctypes.wintypes.DWORD))(
 964         ("GetConsoleMode", ctypes.windll.kernel32))
 965     INVALID_HANDLE_VALUE = ctypes.wintypes.DWORD(-1).value
 966
 967     def not_a_console(handle):
 968         if handle == INVALID_HANDLE_VALUE or handle is None:
 969             return True
 970         return ((GetFileType(handle) & ~FILE_TYPE_REMOTE) != FILE_TYPE_CHAR
 971                 or GetConsoleMode(handle, ctypes.byref(ctypes.wintypes.DWORD())) == 0)
 972
 973     if not_a_console(h):
 974         return False
 975
 976     def next_nonbmp_pos(s):
 977         try:
 978             return next(i for i, c in enumerate(s) if ord(c) > 0xffff)
 979         except StopIteration:
 980             return len(s)
 981
 982     while s:
 983         count = min(next_nonbmp_pos(s), 1024)
 984
 985         ret = WriteConsoleW(
 986             h, s, count if count else 2, ctypes.byref(written), None)
 987         if ret == 0:
 988             raise OSError('Failed to write string')
 989         if not count:  # We just wrote a non-BMP character
 990             assert written.value == 2
 991             s = s[1:]
 992         else:
 993             assert written.value > 0
 994             s = s[written.value:]
 995     return True
 996
 997
 998 def write_string(s, out=None, encoding=None):
 999     if out is None:
1000         out = sys.stderr
1001     assert type(s) == compat_str
1002
1003     if sys.platform == 'win32' and encoding is None and hasattr(out, 'fileno'):
1004         if _windows_write_string(s, out):
1005             return
1006
1007     if ('b' in getattr(out, 'mode', '') or
1008             sys.version_info[0] < 3):  # Python 2 lies about mode of sys.stderr
1009         byt = s.encode(encoding or preferredencoding(), 'ignore')
1010         out.write(byt)
1011     elif hasattr(out, 'buffer'):
1012         enc = encoding or getattr(out, 'encoding', None) or preferredencoding()
1013         byt = s.encode(enc, 'ignore')
1014         out.buffer.write(byt)
1015     else:
1016         out.write(s)
1017     out.flush()
1018
1019
1020 def bytes_to_intlist(bs):
1021     if not bs:
1022         return []
1023     if isinstance(bs[0], int):  # Python 3
1024         return list(bs)
1025     else:
1026         return [ord(c) for c in bs]
1027
1028
1029 def intlist_to_bytes(xs):
1030     if not xs:
1031         return b''
1032     if isinstance(chr(0), bytes):  # Python 2
1033         return ''.join([chr(x) for x in xs])
1034     else:
1035         return bytes(xs)
1036
1037
1038 def get_cachedir(params={}):
1039     cache_root = os.environ.get('XDG_CACHE_HOME',
1040                                 os.path.expanduser('~/.cache'))
1041     return params.get('cachedir', os.path.join(cache_root, 'youtube-dl'))
1042
1043
1044 # Cross-platform file locking
1045 if sys.platform == 'win32':
1046     import ctypes.wintypes
1047     import msvcrt
1048
1049     class OVERLAPPED(ctypes.Structure):
1050         _fields_ = [
1051             ('Internal', ctypes.wintypes.LPVOID),
1052             ('InternalHigh', ctypes.wintypes.LPVOID),
1053             ('Offset', ctypes.wintypes.DWORD),
1054             ('OffsetHigh', ctypes.wintypes.DWORD),
1055             ('hEvent', ctypes.wintypes.HANDLE),
1056         ]
1057
1058     kernel32 = ctypes.windll.kernel32
1059     LockFileEx = kernel32.LockFileEx
1060     LockFileEx.argtypes = [
1061         ctypes.wintypes.HANDLE,     # hFile
1062         ctypes.wintypes.DWORD,      # dwFlags
1063         ctypes.wintypes.DWORD,      # dwReserved
1064         ctypes.wintypes.DWORD,      # nNumberOfBytesToLockLow
1065         ctypes.wintypes.DWORD,      # nNumberOfBytesToLockHigh
1066         ctypes.POINTER(OVERLAPPED)  # Overlapped
1067     ]
1068     LockFileEx.restype = ctypes.wintypes.BOOL
1069     UnlockFileEx = kernel32.UnlockFileEx
1070     UnlockFileEx.argtypes = [
1071         ctypes.wintypes.HANDLE,     # hFile
1072         ctypes.wintypes.DWORD,      # dwReserved
1073         ctypes.wintypes.DWORD,      # nNumberOfBytesToLockLow
1074         ctypes.wintypes.DWORD,      # nNumberOfBytesToLockHigh
1075         ctypes.POINTER(OVERLAPPED)  # Overlapped
1076     ]
1077     UnlockFileEx.restype = ctypes.wintypes.BOOL
1078     whole_low = 0xffffffff
1079     whole_high = 0x7fffffff
1080
1081     def _lock_file(f, exclusive):
1082         overlapped = OVERLAPPED()
1083         overlapped.Offset = 0
1084         overlapped.OffsetHigh = 0
1085         overlapped.hEvent = 0
1086         f._lock_file_overlapped_p = ctypes.pointer(overlapped)
1087         handle = msvcrt.get_osfhandle(f.fileno())
1088         if not LockFileEx(handle, 0x2 if exclusive else 0x0, 0,
1089                           whole_low, whole_high, f._lock_file_overlapped_p):
1090             raise OSError('Locking file failed: %r' % ctypes.FormatError())
1091
1092     def _unlock_file(f):
1093         assert f._lock_file_overlapped_p
1094         handle = msvcrt.get_osfhandle(f.fileno())
1095         if not UnlockFileEx(handle, 0,
1096                             whole_low, whole_high, f._lock_file_overlapped_p):
1097             raise OSError('Unlocking file failed: %r' % ctypes.FormatError())
1098
1099 else:
1100     import fcntl
1101
1102     def _lock_file(f, exclusive):
1103         fcntl.lockf(f, fcntl.LOCK_EX if exclusive else fcntl.LOCK_SH)
1104
1105     def _unlock_file(f):
1106         fcntl.lockf(f, fcntl.LOCK_UN)
1107
1108
1109 class locked_file(object):
1110     def __init__(self, filename, mode, encoding=None):
1111         assert mode in ['r', 'a', 'w']
1112         self.f = io.open(filename, mode, encoding=encoding)
1113         self.mode = mode
1114
1115     def __enter__(self):
1116         exclusive = self.mode != 'r'
1117         try:
1118             _lock_file(self.f, exclusive)
1119         except IOError:
1120             self.f.close()
1121             raise
1122         return self
1123
1124     def __exit__(self, etype, value, traceback):
1125         try:
1126             _unlock_file(self.f)
1127         finally:
1128             self.f.close()
1129
1130     def __iter__(self):
1131         return iter(self.f)
1132
1133     def write(self, *args):
1134         return self.f.write(*args)
1135
1136     def read(self, *args):
1137         return self.f.read(*args)
1138
1139
1140 def shell_quote(args):
1141     quoted_args = []
1142     encoding = sys.getfilesystemencoding()
1143     if encoding is None:
1144         encoding = 'utf-8'
1145     for a in args:
1146         if isinstance(a, bytes):
1147             # We may get a filename encoded with 'encodeFilename'
1148             a = a.decode(encoding)
1149         quoted_args.append(pipes.quote(a))
1150     return u' '.join(quoted_args)
1151
1152
1153 def takewhile_inclusive(pred, seq):
1154     """ Like itertools.takewhile, but include the latest evaluated element
1155         (the first element so that Not pred(e)) """
1156     for e in seq:
1157         yield e
1158         if not pred(e):
1159             return
1160
1161
1162 def smuggle_url(url, data):
1163     """ Pass additional data in a URL for internal use. """
1164
1165     sdata = compat_urllib_parse.urlencode(
1166         {u'__youtubedl_smuggle': json.dumps(data)})
1167     return url + u'#' + sdata
1168
1169
1170 def unsmuggle_url(smug_url, default=None):
1171     if not '#__youtubedl_smuggle' in smug_url:
1172         return smug_url, default
1173     url, _, sdata = smug_url.rpartition(u'#')
1174     jsond = compat_parse_qs(sdata)[u'__youtubedl_smuggle'][0]
1175     data = json.loads(jsond)
1176     return url, data
1177
1178
1179 def format_bytes(bytes):
1180     if bytes is None:
1181         return u'N/A'
1182     if type(bytes) is str:
1183         bytes = float(bytes)
1184     if bytes == 0.0:
1185         exponent = 0
1186     else:
1187         exponent = int(math.log(bytes, 1024.0))
1188     suffix = [u'B', u'KiB', u'MiB', u'GiB', u'TiB', u'PiB', u'EiB', u'ZiB', u'YiB'][exponent]
1189     converted = float(bytes) / float(1024 ** exponent)
1190     return u'%.2f%s' % (converted, suffix)
1191
1192
1193 def str_to_int(int_str):
1194     int_str = re.sub(r'[,\.]', u'', int_str)
1195     return int(int_str)
1196
1197
1198 def get_term_width():
1199     columns = os.environ.get('COLUMNS', None)
1200     if columns:
1201         return int(columns)
1202
1203     try:
1204         sp = subprocess.Popen(
1205             ['stty', 'size'],
1206             stdout=subprocess.PIPE, stderr=subprocess.PIPE)
1207         out, err = sp.communicate()
1208         return int(out.split()[1])
1209     except:
1210         pass
1211     return None
1212
1213
1214 def month_by_name(name):
1215     """ Return the number of a month by (locale-independently) English name """
1216
1217     ENGLISH_NAMES = [
1218         u'January', u'February', u'March', u'April', u'May', u'June',
1219         u'July', u'August', u'September', u'October', u'November', u'December']
1220     try:
1221         return ENGLISH_NAMES.index(name) + 1
1222     except ValueError:
1223         return None
1224
1225
1226 def fix_xml_ampersands(xml_str):
1227     """Replace all the '&' by '&amp;' in XML"""
1228     return re.sub(
1229         r'&(?!amp;|lt;|gt;|apos;|quot;|#x[0-9a-fA-F]{,4};|#[0-9]{,4};)',
1230         u'&amp;',
1231         xml_str)
1232
1233
1234 def setproctitle(title):
1235     assert isinstance(title, compat_str)
1236     try:
1237         libc = ctypes.cdll.LoadLibrary("libc.so.6")
1238     except OSError:
1239         return
1240     title_bytes = title.encode('utf-8')
1241     buf = ctypes.create_string_buffer(len(title_bytes))
1242     buf.value = title_bytes
1243     try:
1244         libc.prctl(15, buf, 0, 0, 0)
1245     except AttributeError:
1246         return  # Strange libc, just skip this
1247
1248
1249 def remove_start(s, start):
1250     if s.startswith(start):
1251         return s[len(start):]
1252     return s
1253
1254
1255 def url_basename(url):
1256     path = compat_urlparse.urlparse(url).path
1257     return path.strip(u'/').split(u'/')[-1]
1258
1259
1260 class HEADRequest(compat_urllib_request.Request):
1261     def get_method(self):
1262         return "HEAD"
1263
1264
1265 def int_or_none(v, scale=1, default=None, get_attr=None):
1266     if get_attr:
1267         if v is not None:
1268             v = getattr(v, get_attr, None)
1269     return default if v is None else (int(v) // scale)
1270
1271
1272 def float_or_none(v, scale=1, default=None):
1273     return default if v is None else (float(v) / scale)
1274
1275
1276 def parse_duration(s):
1277     if s is None:
1278         return None
1279
1280     m = re.match(
1281         r'(?:(?:(?P<hours>[0-9]+)[:h])?(?P<mins>[0-9]+)[:m])?(?P<secs>[0-9]+)s?(?::[0-9]+)?$', s)
1282     if not m:
1283         return None
1284     res = int(m.group('secs'))
1285     if m.group('mins'):
1286         res += int(m.group('mins')) * 60
1287         if m.group('hours'):
1288             res += int(m.group('hours')) * 60 * 60
1289     return res
1290
1291
1292 def prepend_extension(filename, ext):
1293     name, real_ext = os.path.splitext(filename)
1294     return u'{0}.{1}{2}'.format(name, ext, real_ext)
1295
1296
1297 def check_executable(exe, args=[]):
1298     """ Checks if the given binary is installed somewhere in PATH, and returns its name.
1299     args can be a list of arguments for a short output (like -version) """
1300     try:
1301         subprocess.Popen([exe] + args, stdout=subprocess.PIPE, stderr=subprocess.PIPE).communicate()
1302     except OSError:
1303         return False
1304     return exe
1305
1306
1307 class PagedList(object):
1308     def __init__(self, pagefunc, pagesize):
1309         self._pagefunc = pagefunc
1310         self._pagesize = pagesize
1311
1312     def __len__(self):
1313         # This is only useful for tests
1314         return len(self.getslice())
1315
1316     def getslice(self, start=0, end=None):
1317         res = []
1318         for pagenum in itertools.count(start // self._pagesize):
1319             firstid = pagenum * self._pagesize
1320             nextfirstid = pagenum * self._pagesize + self._pagesize
1321             if start >= nextfirstid:
1322                 continue
1323
1324             page_results = list(self._pagefunc(pagenum))
1325
1326             startv = (
1327                 start % self._pagesize
1328                 if firstid <= start < nextfirstid
1329                 else 0)
1330
1331             endv = (
1332                 ((end - 1) % self._pagesize) + 1
1333                 if (end is not None and firstid <= end <= nextfirstid)
1334                 else None)
1335
1336             if startv != 0 or endv is not None:
1337                 page_results = page_results[startv:endv]
1338             res.extend(page_results)
1339
1340             # A little optimization - if current page is not "full", ie. does
1341             # not contain page_size videos then we can assume that this page
1342             # is the last one - there are no more ids on further pages -
1343             # i.e. no need to query again.
1344             if len(page_results) + startv < self._pagesize:
1345                 break
1346
1347             # If we got the whole page, but the next page is not interesting,
1348             # break out early as well
1349             if end == nextfirstid:
1350                 break
1351         return res
1352
1353
1354 def uppercase_escape(s):
1355     unicode_escape = codecs.getdecoder('unicode_escape')
1356     return re.sub(
1357         r'\\U[0-9a-fA-F]{8}',
1358         lambda m: unicode_escape(m.group(0))[0],
1359         s)
1360
1361 try:
1362     struct.pack(u'!I', 0)
1363 except TypeError:
1364     # In Python 2.6 (and some 2.7 versions), struct requires a bytes argument
1365     def struct_pack(spec, *args):
1366         if isinstance(spec, compat_str):
1367             spec = spec.encode('ascii')
1368         return struct.pack(spec, *args)
1369
1370     def struct_unpack(spec, *args):
1371         if isinstance(spec, compat_str):
1372             spec = spec.encode('ascii')
1373         return struct.unpack(spec, *args)
1374 else:
1375     struct_pack = struct.pack
1376     struct_unpack = struct.unpack
1377
1378
1379 def read_batch_urls(batch_fd):
1380     def fixup(url):
1381         if not isinstance(url, compat_str):
1382             url = url.decode('utf-8', 'replace')
1383         BOM_UTF8 = u'\xef\xbb\xbf'
1384         if url.startswith(BOM_UTF8):
1385             url = url[len(BOM_UTF8):]
1386         url = url.strip()
1387         if url.startswith(('#', ';', ']')):
1388             return False
1389         return url
1390
1391     with contextlib.closing(batch_fd) as fd:
1392         return [url for url in map(fixup, fd) if url]
1393
1394
1395 def urlencode_postdata(*args, **kargs):
1396     return compat_urllib_parse.urlencode(*args, **kargs).encode('ascii')
1397
1398
1399 def parse_xml(s):
1400     class TreeBuilder(xml.etree.ElementTree.TreeBuilder):
1401         def doctype(self, name, pubid, system):
1402             pass  # Ignore doctypes
1403
1404     parser = xml.etree.ElementTree.XMLParser(target=TreeBuilder())
1405     kwargs = {'parser': parser} if sys.version_info >= (2, 7) else {}
1406     return xml.etree.ElementTree.XML(s.encode('utf-8'), **kwargs)
1407
1408
1409 if sys.version_info < (3, 0) and sys.platform == 'win32':
1410     def compat_getpass(prompt, *args, **kwargs):
1411         if isinstance(prompt, compat_str):
1412             prompt = prompt.encode(preferredencoding())
1413         return getpass.getpass(prompt, *args, **kwargs)
1414 else:
1415     compat_getpass = getpass.getpass
1416
1417
1418 US_RATINGS = {
1419     'G': 0,
1420     'PG': 10,
1421     'PG-13': 13,
1422     'R': 16,
1423     'NC': 18,
1424 }
1425
1426
1427 def strip_jsonp(code):
1428     return re.sub(r'(?s)^[a-zA-Z_]+\s*\(\s*(.*)\);\s*?\s*$', r'\1', code)
1429
1430
1431 def qualities(quality_ids):
1432     """ Get a numeric quality value out of a list of possible values """
1433     def q(qid):
1434         try:
1435             return quality_ids.index(qid)
1436         except ValueError:
1437             return -1
1438     return q
1439
1440
1441 DEFAULT_OUTTMPL = '%(title)s-%(id)s.%(ext)s'
1442
1443 try:
1444     subprocess_check_output = subprocess.check_output
1445 except AttributeError:
1446     def subprocess_check_output(*args, **kwargs):
1447         assert 'input' not in kwargs
1448         p = subprocess.Popen(*args, stdout=subprocess.PIPE, **kwargs)
1449         output, _ = p.communicate()
1450         ret = p.poll()
1451         if ret:
1452             raise subprocess.CalledProcessError(ret, p.args, output=output)
1453         return output