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

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