_ Git - youtube-dl/blob - youtube_dl/FileDownloader.py

   1 #!/usr/bin/env python
   2 # -*- coding: utf-8 -*-
   3
   4 import httplib
   5 import math
   6 import os
   7 import re
   8 import socket
   9 import subprocess
  10 import sys
  11 import time
  12 import urllib2
  13
  14 if os.name == 'nt':
  15         import ctypes
  16
  17 from utils import *
  18
  19
  20 class FileDownloader(object):
  21         """File Downloader class.
  22
  23         File downloader objects are the ones responsible of downloading the
  24         actual video file and writing it to disk if the user has requested
  25         it, among some other tasks. In most cases there should be one per
  26         program. As, given a video URL, the downloader doesn't know how to
  27         extract all the needed information, task that InfoExtractors do, it
  28         has to pass the URL to one of them.
  29
  30         For this, file downloader objects have a method that allows
  31         InfoExtractors to be registered in a given order. When it is passed
  32         a URL, the file downloader handles it to the first InfoExtractor it
  33         finds that reports being able to handle it. The InfoExtractor extracts
  34         all the information about the video or videos the URL refers to, and
  35         asks the FileDownloader to process the video information, possibly
  36         downloading the video.
  37
  38         File downloaders accept a lot of parameters. In order not to saturate
  39         the object constructor with arguments, it receives a dictionary of
  40         options instead. These options are available through the params
  41         attribute for the InfoExtractors to use. The FileDownloader also
  42         registers itself as the downloader in charge for the InfoExtractors
  43         that are added to it, so this is a "mutual registration".
  44
  45         Available options:
  46
  47         username:         Username for authentication purposes.
  48         password:         Password for authentication purposes.
  49         usenetrc:         Use netrc for authentication instead.
  50         quiet:            Do not print messages to stdout.
  51         forceurl:         Force printing final URL.
  52         forcetitle:       Force printing title.
  53         forcethumbnail:   Force printing thumbnail URL.
  54         forcedescription: Force printing description.
  55         forcefilename:    Force printing final filename.
  56         simulate:         Do not download the video files.
  57         format:           Video format code.
  58         format_limit:     Highest quality format to try.
  59         outtmpl:          Template for output names.
  60         ignoreerrors:     Do not stop on download errors.
  61         ratelimit:        Download speed limit, in bytes/sec.
  62         nooverwrites:     Prevent overwriting files.
  63         retries:          Number of times to retry for HTTP error 5xx
  64         continuedl:       Try to continue downloads if possible.
  65         noprogress:       Do not print the progress bar.
  66         playliststart:    Playlist item to start at.
  67         playlistend:      Playlist item to end at.
  68         matchtitle:       Download only matching titles.
  69         rejecttitle:      Reject downloads for matching titles.
  70         logtostderr:      Log messages to stderr instead of stdout.
  71         consoletitle:     Display progress in console window's titlebar.
  72         nopart:           Do not use temporary .part files.
  73         updatetime:       Use the Last-modified header to set output file timestamps.
  74         writedescription: Write the video description to a .description file
  75         writeinfojson:    Write the video description to a .info.json file
  76         writesubtitles:   Write the video subtitles to a .srt file
  77         subtitleslang:    Language of the subtitles to download
  78         """
  79
  80         params = None
  81         _ies = []
  82         _pps = []
  83         _download_retcode = None
  84         _num_downloads = None
  85         _screen_file = None
  86
  87         def __init__(self, params):
  88                 """Create a FileDownloader object with the given options."""
  89                 self._ies = []
  90                 self._pps = []
  91                 self._download_retcode = 0
  92                 self._num_downloads = 0
  93                 self._screen_file = [sys.stdout, sys.stderr][params.get('logtostderr', False)]
  94                 self.params = params
  95
  96         @staticmethod
  97         def format_bytes(bytes):
  98                 if bytes is None:
  99                         return 'N/A'
 100                 if type(bytes) is str:
 101                         bytes = float(bytes)
 102                 if bytes == 0.0:
 103                         exponent = 0
 104                 else:
 105                         exponent = long(math.log(bytes, 1024.0))
 106                 suffix = 'bkMGTPEZY'[exponent]
 107                 converted = float(bytes) / float(1024 ** exponent)
 108                 return '%.2f%s' % (converted, suffix)
 109
 110         @staticmethod
 111         def calc_percent(byte_counter, data_len):
 112                 if data_len is None:
 113                         return '---.-%'
 114                 return '%6s' % ('%3.1f%%' % (float(byte_counter) / float(data_len) * 100.0))
 115
 116         @staticmethod
 117         def calc_eta(start, now, total, current):
 118                 if total is None:
 119                         return '--:--'
 120                 dif = now - start
 121                 if current == 0 or dif < 0.001: # One millisecond
 122                         return '--:--'
 123                 rate = float(current) / dif
 124                 eta = long((float(total) - float(current)) / rate)
 125                 (eta_mins, eta_secs) = divmod(eta, 60)
 126                 if eta_mins > 99:
 127                         return '--:--'
 128                 return '%02d:%02d' % (eta_mins, eta_secs)
 129
 130         @staticmethod
 131         def calc_speed(start, now, bytes):
 132                 dif = now - start
 133                 if bytes == 0 or dif < 0.001: # One millisecond
 134                         return '%10s' % '---b/s'
 135                 return '%10s' % ('%s/s' % FileDownloader.format_bytes(float(bytes) / dif))
 136
 137         @staticmethod
 138         def best_block_size(elapsed_time, bytes):
 139                 new_min = max(bytes / 2.0, 1.0)
 140                 new_max = min(max(bytes * 2.0, 1.0), 4194304) # Do not surpass 4 MB
 141                 if elapsed_time < 0.001:
 142                         return long(new_max)
 143                 rate = bytes / elapsed_time
 144                 if rate > new_max:
 145                         return long(new_max)
 146                 if rate < new_min:
 147                         return long(new_min)
 148                 return long(rate)
 149
 150         @staticmethod
 151         def parse_bytes(bytestr):
 152                 """Parse a string indicating a byte quantity into a long integer."""
 153                 matchobj = re.match(r'(?i)^(\d+(?:\.\d+)?)([kMGTPEZY]?)$', bytestr)
 154                 if matchobj is None:
 155                         return None
 156                 number = float(matchobj.group(1))
 157                 multiplier = 1024.0 ** 'bkmgtpezy'.index(matchobj.group(2).lower())
 158                 return long(round(number * multiplier))
 159
 160         def add_info_extractor(self, ie):
 161                 """Add an InfoExtractor object to the end of the list."""
 162                 self._ies.append(ie)
 163                 ie.set_downloader(self)
 164
 165         def add_post_processor(self, pp):
 166                 """Add a PostProcessor object to the end of the chain."""
 167                 self._pps.append(pp)
 168                 pp.set_downloader(self)
 169
 170         def to_screen(self, message, skip_eol=False):
 171                 """Print message to stdout if not in quiet mode."""
 172                 assert type(message) == type(u'')
 173                 if not self.params.get('quiet', False):
 174                         terminator = [u'\n', u''][skip_eol]
 175                         output = message + terminator
 176
 177                         if 'b' not in self._screen_file.mode or sys.version_info[0] < 3: # Python 2 lies about the mode of sys.stdout/sys.stderr
 178                                 output = output.encode(preferredencoding(), 'ignore')
 179                         self._screen_file.write(output)
 180                         self._screen_file.flush()
 181
 182         def to_stderr(self, message):
 183                 """Print message to stderr."""
 184                 print >>sys.stderr, message.encode(preferredencoding())
 185
 186         def to_cons_title(self, message):
 187                 """Set console/terminal window title to message."""
 188                 if not self.params.get('consoletitle', False):
 189                         return
 190                 if os.name == 'nt' and ctypes.windll.kernel32.GetConsoleWindow():
 191                         # c_wchar_p() might not be necessary if `message` is
 192                         # already of type unicode()
 193                         ctypes.windll.kernel32.SetConsoleTitleW(ctypes.c_wchar_p(message))
 194                 elif 'TERM' in os.environ:
 195                         sys.stderr.write('\033]0;%s\007' % message.encode(preferredencoding()))
 196
 197         def fixed_template(self):
 198                 """Checks if the output template is fixed."""
 199                 return (re.search(ur'(?u)%\(.+?\)s', self.params['outtmpl']) is None)
 200
 201         def trouble(self, message=None):
 202                 """Determine action to take when a download problem appears.
 203
 204                 Depending on if the downloader has been configured to ignore
 205                 download errors or not, this method may throw an exception or
 206                 not when errors are found, after printing the message.
 207                 """
 208                 if message is not None:
 209                         self.to_stderr(message)
 210                 if not self.params.get('ignoreerrors', False):
 211                         raise DownloadError(message)
 212                 self._download_retcode = 1
 213
 214         def slow_down(self, start_time, byte_counter):
 215                 """Sleep if the download speed is over the rate limit."""
 216                 rate_limit = self.params.get('ratelimit', None)
 217                 if rate_limit is None or byte_counter == 0:
 218                         return
 219                 now = time.time()
 220                 elapsed = now - start_time
 221                 if elapsed <= 0.0:
 222                         return
 223                 speed = float(byte_counter) / elapsed
 224                 if speed > rate_limit:
 225                         time.sleep((byte_counter - rate_limit * (now - start_time)) / rate_limit)
 226
 227         def temp_name(self, filename):
 228                 """Returns a temporary filename for the given filename."""
 229                 if self.params.get('nopart', False) or filename == u'-' or \
 230                                 (os.path.exists(encodeFilename(filename)) and not os.path.isfile(encodeFilename(filename))):
 231                         return filename
 232                 return filename + u'.part'
 233
 234         def undo_temp_name(self, filename):
 235                 if filename.endswith(u'.part'):
 236                         return filename[:-len(u'.part')]
 237                 return filename
 238
 239         def try_rename(self, old_filename, new_filename):
 240                 try:
 241                         if old_filename == new_filename:
 242                                 return
 243                         os.rename(encodeFilename(old_filename), encodeFilename(new_filename))
 244                 except (IOError, OSError), err:
 245                         self.trouble(u'ERROR: unable to rename file')
 246
 247         def try_utime(self, filename, last_modified_hdr):
 248                 """Try to set the last-modified time of the given file."""
 249                 if last_modified_hdr is None:
 250                         return
 251                 if not os.path.isfile(encodeFilename(filename)):
 252                         return
 253                 timestr = last_modified_hdr
 254                 if timestr is None:
 255                         return
 256                 filetime = timeconvert(timestr)
 257                 if filetime is None:
 258                         return filetime
 259                 try:
 260                         os.utime(filename, (time.time(), filetime))
 261                 except:
 262                         pass
 263                 return filetime
 264
 265         def report_writedescription(self, descfn):
 266                 """ Report that the description file is being written """
 267                 self.to_screen(u'[info] Writing video description to: ' + descfn)
 268
 269         def report_writesubtitles(self, srtfn):
 270                 """ Report that the subtitles file is being written """
 271                 self.to_screen(u'[info] Writing video subtitles to: ' + srtfn)
 272
 273         def report_writeinfojson(self, infofn):
 274                 """ Report that the metadata file has been written """
 275                 self.to_screen(u'[info] Video description metadata as JSON to: ' + infofn)
 276
 277         def report_destination(self, filename):
 278                 """Report destination filename."""
 279                 self.to_screen(u'[download] Destination: ' + filename)
 280
 281         def report_progress(self, percent_str, data_len_str, speed_str, eta_str):
 282                 """Report download progress."""
 283                 if self.params.get('noprogress', False):
 284                         return
 285                 self.to_screen(u'\r[download] %s of %s at %s ETA %s' %
 286                                 (percent_str, data_len_str, speed_str, eta_str), skip_eol=True)
 287                 self.to_cons_title(u'youtube-dl - %s of %s at %s ETA %s' %
 288                                 (percent_str.strip(), data_len_str.strip(), speed_str.strip(), eta_str.strip()))
 289
 290         def report_resuming_byte(self, resume_len):
 291                 """Report attempt to resume at given byte."""
 292                 self.to_screen(u'[download] Resuming download at byte %s' % resume_len)
 293
 294         def report_retry(self, count, retries):
 295                 """Report retry in case of HTTP error 5xx"""
 296                 self.to_screen(u'[download] Got server HTTP error. Retrying (attempt %d of %d)...' % (count, retries))
 297
 298         def report_file_already_downloaded(self, file_name):
 299                 """Report file has already been fully downloaded."""
 300                 try:
 301                         self.to_screen(u'[download] %s has already been downloaded' % file_name)
 302                 except (UnicodeEncodeError), err:
 303                         self.to_screen(u'[download] The file has already been downloaded')
 304
 305         def report_unable_to_resume(self):
 306                 """Report it was impossible to resume download."""
 307                 self.to_screen(u'[download] Unable to resume')
 308
 309         def report_finish(self):
 310                 """Report download finished."""
 311                 if self.params.get('noprogress', False):
 312                         self.to_screen(u'[download] Download completed')
 313                 else:
 314                         self.to_screen(u'')
 315
 316         def increment_downloads(self):
 317                 """Increment the ordinal that assigns a number to each file."""
 318                 self._num_downloads += 1
 319
 320         def prepare_filename(self, info_dict):
 321                 """Generate the output filename."""
 322                 try:
 323                         template_dict = dict(info_dict)
 324                         template_dict['epoch'] = unicode(long(time.time()))
 325                         template_dict['autonumber'] = unicode('%05d' % self._num_downloads)
 326                         filename = self.params['outtmpl'] % template_dict
 327                         return filename
 328                 except (ValueError, KeyError), err:
 329                         self.trouble(u'ERROR: invalid system charset or erroneous output template')
 330                         return None
 331
 332         def _match_entry(self, info_dict):
 333                 """ Returns None iff the file should be downloaded """
 334
 335                 title = info_dict['title']
 336                 matchtitle = self.params.get('matchtitle', False)
 337                 if matchtitle and not re.search(matchtitle, title, re.IGNORECASE):
 338                         return u'[download] "' + title + '" title did not match pattern "' + matchtitle + '"'
 339                 rejecttitle = self.params.get('rejecttitle', False)
 340                 if rejecttitle and re.search(rejecttitle, title, re.IGNORECASE):
 341                         return u'"' + title + '" title matched reject pattern "' + rejecttitle + '"'
 342                 return None
 343
 344         def process_info(self, info_dict):
 345                 """Process a single dictionary returned by an InfoExtractor."""
 346
 347                 info_dict['stitle'] = sanitize_filename(info_dict['title'])
 348
 349                 reason = self._match_entry(info_dict)
 350                 if reason is not None:
 351                         self.to_screen(u'[download] ' + reason)
 352                         return
 353
 354                 max_downloads = self.params.get('max_downloads')
 355                 if max_downloads is not None:
 356                         if self._num_downloads > int(max_downloads):
 357                                 raise MaxDownloadsReached()
 358
 359                 filename = self.prepare_filename(info_dict)
 360
 361                 # Forced printings
 362                 if self.params.get('forcetitle', False):
 363                         print info_dict['title'].encode(preferredencoding(), 'xmlcharrefreplace')
 364                 if self.params.get('forceurl', False):
 365                         print info_dict['url'].encode(preferredencoding(), 'xmlcharrefreplace')
 366                 if self.params.get('forcethumbnail', False) and 'thumbnail' in info_dict:
 367                         print info_dict['thumbnail'].encode(preferredencoding(), 'xmlcharrefreplace')
 368                 if self.params.get('forcedescription', False) and 'description' in info_dict:
 369                         print info_dict['description'].encode(preferredencoding(), 'xmlcharrefreplace')
 370                 if self.params.get('forcefilename', False) and filename is not None:
 371                         print filename.encode(preferredencoding(), 'xmlcharrefreplace')
 372                 if self.params.get('forceformat', False):
 373                         print info_dict['format'].encode(preferredencoding(), 'xmlcharrefreplace')
 374
 375                 # Do nothing else if in simulate mode
 376                 if self.params.get('simulate', False):
 377                         return
 378
 379                 if filename is None:
 380                         return
 381
 382                 try:
 383                         dn = os.path.dirname(encodeFilename(filename))
 384                         if dn != '' and not os.path.exists(dn): # dn is already encoded
 385                                 os.makedirs(dn)
 386                 except (OSError, IOError), err:
 387                         self.trouble(u'ERROR: unable to create directory ' + unicode(err))
 388                         return
 389
 390                 if self.params.get('writedescription', False):
 391                         try:
 392                                 descfn = filename + u'.description'
 393                                 self.report_writedescription(descfn)
 394                                 descfile = open(encodeFilename(descfn), 'wb')
 395                                 try:
 396                                         descfile.write(info_dict['description'].encode('utf-8'))
 397                                 finally:
 398                                         descfile.close()
 399                         except (OSError, IOError):
 400                                 self.trouble(u'ERROR: Cannot write description file ' + descfn)
 401                                 return
 402
 403                 if self.params.get('writesubtitles', False) and 'subtitles' in info_dict and info_dict['subtitles']:
 404                         # subtitles download errors are already managed as troubles in relevant IE
 405                         # that way it will silently go on when used with unsupporting IE
 406                         try:
 407                                 srtfn = filename.rsplit('.', 1)[0] + u'.srt'
 408                                 self.report_writesubtitles(srtfn)
 409                                 srtfile = open(encodeFilename(srtfn), 'wb')
 410                                 try:
 411                                         srtfile.write(info_dict['subtitles'].encode('utf-8'))
 412                                 finally:
 413                                         srtfile.close()
 414                         except (OSError, IOError):
 415                                 self.trouble(u'ERROR: Cannot write subtitles file ' + descfn)
 416                                 return
 417
 418                 if self.params.get('writeinfojson', False):
 419                         infofn = filename + u'.info.json'
 420                         self.report_writeinfojson(infofn)
 421                         try:
 422                                 json.dump
 423                         except (NameError,AttributeError):
 424                                 self.trouble(u'ERROR: No JSON encoder found. Update to Python 2.6+, setup a json module, or leave out --write-info-json.')
 425                                 return
 426                         try:
 427                                 infof = open(encodeFilename(infofn), 'wb')
 428                                 try:
 429                                         json_info_dict = dict((k,v) for k,v in info_dict.iteritems() if not k in ('urlhandle',))
 430                                         json.dump(json_info_dict, infof)
 431                                 finally:
 432                                         infof.close()
 433                         except (OSError, IOError):
 434                                 self.trouble(u'ERROR: Cannot write metadata to JSON file ' + infofn)
 435                                 return
 436
 437                 if not self.params.get('skip_download', False):
 438                         if self.params.get('nooverwrites', False) and os.path.exists(encodeFilename(filename)):
 439                                 success = True
 440                         else:
 441                                 try:
 442                                         success = self._do_download(filename, info_dict)
 443                                 except (OSError, IOError), err:
 444                                         raise UnavailableVideoError
 445                                 except (urllib2.URLError, httplib.HTTPException, socket.error), err:
 446                                         self.trouble(u'ERROR: unable to download video data: %s' % str(err))
 447                                         return
 448                                 except (ContentTooShortError, ), err:
 449                                         self.trouble(u'ERROR: content too short (expected %s bytes and served %s)' % (err.expected, err.downloaded))
 450                                         return
 451
 452                         if success:
 453                                 try:
 454                                         self.post_process(filename, info_dict)
 455                                 except (PostProcessingError), err:
 456                                         self.trouble(u'ERROR: postprocessing: %s' % str(err))
 457                                         return
 458
 459         def download(self, url_list):
 460                 """Download a given list of URLs."""
 461                 if len(url_list) > 1 and self.fixed_template():
 462                         raise SameFileError(self.params['outtmpl'])
 463
 464                 for url in url_list:
 465                         suitable_found = False
 466                         for ie in self._ies:
 467                                 # Go to next InfoExtractor if not suitable
 468                                 if not ie.suitable(url):
 469                                         continue
 470
 471                                 # Suitable InfoExtractor found
 472                                 suitable_found = True
 473
 474                                 # Extract information from URL and process it
 475                                 videos = ie.extract(url)
 476                                 for video in videos or []:
 477                                         video['extractor'] = ie.IE_NAME
 478                                         try:
 479                                                 self.increment_downloads()
 480                                                 self.process_info(video)
 481                                         except UnavailableVideoError:
 482                                                 self.trouble(u'\nERROR: unable to download video')
 483
 484                                 # Suitable InfoExtractor had been found; go to next URL
 485                                 break
 486
 487                         if not suitable_found:
 488                                 self.trouble(u'ERROR: no suitable InfoExtractor: %s' % url)
 489
 490                 return self._download_retcode
 491
 492         def post_process(self, filename, ie_info):
 493                 """Run the postprocessing chain on the given file."""
 494                 info = dict(ie_info)
 495                 info['filepath'] = filename
 496                 for pp in self._pps:
 497                         info = pp.run(info)
 498                         if info is None:
 499                                 break
 500
 501         def _download_with_rtmpdump(self, filename, url, player_url):
 502                 self.report_destination(filename)
 503                 tmpfilename = self.temp_name(filename)
 504
 505                 # Check for rtmpdump first
 506                 try:
 507                         subprocess.call(['rtmpdump', '-h'], stdout=(file(os.path.devnull, 'w')), stderr=subprocess.STDOUT)
 508                 except (OSError, IOError):
 509                         self.trouble(u'ERROR: RTMP download detected but "rtmpdump" could not be run')
 510                         return False
 511
 512                 # Download using rtmpdump. rtmpdump returns exit code 2 when
 513                 # the connection was interrumpted and resuming appears to be
 514                 # possible. This is part of rtmpdump's normal usage, AFAIK.
 515                 basic_args = ['rtmpdump', '-q'] + [[], ['-W', player_url]][player_url is not None] + ['-r', url, '-o', tmpfilename]
 516                 args = basic_args + [[], ['-e', '-k', '1']][self.params.get('continuedl', False)]
 517                 if self.params.get('verbose', False):
 518                         try:
 519                                 import pipes
 520                                 shell_quote = lambda args: ' '.join(map(pipes.quote, args))
 521                         except ImportError:
 522                                 shell_quote = repr
 523                         self.to_screen(u'[debug] rtmpdump command line: ' + shell_quote(args))
 524                 retval = subprocess.call(args)
 525                 while retval == 2 or retval == 1:
 526                         prevsize = os.path.getsize(encodeFilename(tmpfilename))
 527                         self.to_screen(u'\r[rtmpdump] %s bytes' % prevsize, skip_eol=True)
 528                         time.sleep(5.0) # This seems to be needed
 529                         retval = subprocess.call(basic_args + ['-e'] + [[], ['-k', '1']][retval == 1])
 530                         cursize = os.path.getsize(encodeFilename(tmpfilename))
 531                         if prevsize == cursize and retval == 1:
 532                                 break
 533                          # Some rtmp streams seem abort after ~ 99.8%. Don't complain for those
 534                         if prevsize == cursize and retval == 2 and cursize > 1024:
 535                                 self.to_screen(u'\r[rtmpdump] Could not download the whole video. This can happen for some advertisements.')
 536                                 retval = 0
 537                                 break
 538                 if retval == 0:
 539                         self.to_screen(u'\r[rtmpdump] %s bytes' % os.path.getsize(encodeFilename(tmpfilename)))
 540                         self.try_rename(tmpfilename, filename)
 541                         return True
 542                 else:
 543                         self.trouble(u'\nERROR: rtmpdump exited with code %d' % retval)
 544                         return False
 545
 546         def _do_download(self, filename, info_dict):
 547                 url = info_dict['url']
 548                 player_url = info_dict.get('player_url', None)
 549
 550                 # Check file already present
 551                 if self.params.get('continuedl', False) and os.path.isfile(encodeFilename(filename)) and not self.params.get('nopart', False):
 552                         self.report_file_already_downloaded(filename)
 553                         return True
 554
 555                 # Attempt to download using rtmpdump
 556                 if url.startswith('rtmp'):
 557                         return self._download_with_rtmpdump(filename, url, player_url)
 558
 559                 tmpfilename = self.temp_name(filename)
 560                 stream = None
 561
 562                 # Do not include the Accept-Encoding header
 563                 headers = {'Youtubedl-no-compression': 'True'}
 564                 basic_request = urllib2.Request(url, None, headers)
 565                 request = urllib2.Request(url, None, headers)
 566
 567                 # Establish possible resume length
 568                 if os.path.isfile(encodeFilename(tmpfilename)):
 569                         resume_len = os.path.getsize(encodeFilename(tmpfilename))
 570                 else:
 571                         resume_len = 0
 572
 573                 open_mode = 'wb'
 574                 if resume_len != 0:
 575                         if self.params.get('continuedl', False):
 576                                 self.report_resuming_byte(resume_len)
 577                                 request.add_header('Range','bytes=%d-' % resume_len)
 578                                 open_mode = 'ab'
 579                         else:
 580                                 resume_len = 0
 581
 582                 count = 0
 583                 retries = self.params.get('retries', 0)
 584                 while count <= retries:
 585                         # Establish connection
 586                         try:
 587                                 if count == 0 and 'urlhandle' in info_dict:
 588                                         data = info_dict['urlhandle']
 589                                 data = urllib2.urlopen(request)
 590                                 break
 591                         except (urllib2.HTTPError, ), err:
 592                                 if (err.code < 500 or err.code >= 600) and err.code != 416:
 593                                         # Unexpected HTTP error
 594                                         raise
 595                                 elif err.code == 416:
 596                                         # Unable to resume (requested range not satisfiable)
 597                                         try:
 598                                                 # Open the connection again without the range header
 599                                                 data = urllib2.urlopen(basic_request)
 600                                                 content_length = data.info()['Content-Length']
 601                                         except (urllib2.HTTPError, ), err:
 602                                                 if err.code < 500 or err.code >= 600:
 603                                                         raise
 604                                         else:
 605                                                 # Examine the reported length
 606                                                 if (content_length is not None and
 607                                                                 (resume_len - 100 < long(content_length) < resume_len + 100)):
 608                                                         # The file had already been fully downloaded.
 609                                                         # Explanation to the above condition: in issue #175 it was revealed that
 610                                                         # YouTube sometimes adds or removes a few bytes from the end of the file,
 611                                                         # changing the file size slightly and causing problems for some users. So
 612                                                         # I decided to implement a suggested change and consider the file
 613                                                         # completely downloaded if the file size differs less than 100 bytes from
 614                                                         # the one in the hard drive.
 615                                                         self.report_file_already_downloaded(filename)
 616                                                         self.try_rename(tmpfilename, filename)
 617                                                         return True
 618                                                 else:
 619                                                         # The length does not match, we start the download over
 620                                                         self.report_unable_to_resume()
 621                                                         open_mode = 'wb'
 622                                                         break
 623                         # Retry
 624                         count += 1
 625                         if count <= retries:
 626                                 self.report_retry(count, retries)
 627
 628                 if count > retries:
 629                         self.trouble(u'ERROR: giving up after %s retries' % retries)
 630                         return False
 631
 632                 data_len = data.info().get('Content-length', None)
 633                 if data_len is not None:
 634                         data_len = long(data_len) + resume_len
 635                 data_len_str = self.format_bytes(data_len)
 636                 byte_counter = 0 + resume_len
 637                 block_size = 1024
 638                 start = time.time()
 639                 while True:
 640                         # Download and write
 641                         before = time.time()
 642                         data_block = data.read(block_size)
 643                         after = time.time()
 644                         if len(data_block) == 0:
 645                                 break
 646                         byte_counter += len(data_block)
 647
 648                         # Open file just in time
 649                         if stream is None:
 650                                 try:
 651                                         (stream, tmpfilename) = sanitize_open(tmpfilename, open_mode)
 652                                         assert stream is not None
 653                                         filename = self.undo_temp_name(tmpfilename)
 654                                         self.report_destination(filename)
 655                                 except (OSError, IOError), err:
 656                                         self.trouble(u'ERROR: unable to open for writing: %s' % str(err))
 657                                         return False
 658                         try:
 659                                 stream.write(data_block)
 660                         except (IOError, OSError), err:
 661                                 self.trouble(u'\nERROR: unable to write data: %s' % str(err))
 662                                 return False
 663                         block_size = self.best_block_size(after - before, len(data_block))
 664
 665                         # Progress message
 666                         speed_str = self.calc_speed(start, time.time(), byte_counter - resume_len)
 667                         if data_len is None:
 668                                 self.report_progress('Unknown %', data_len_str, speed_str, 'Unknown ETA')
 669                         else:
 670                                 percent_str = self.calc_percent(byte_counter, data_len)
 671                                 eta_str = self.calc_eta(start, time.time(), data_len - resume_len, byte_counter - resume_len)
 672                                 self.report_progress(percent_str, data_len_str, speed_str, eta_str)
 673
 674                         # Apply rate limit
 675                         self.slow_down(start, byte_counter - resume_len)
 676
 677                 if stream is None:
 678                         self.trouble(u'\nERROR: Did not get any data blocks')
 679                         return False
 680                 stream.close()
 681                 self.report_finish()
 682                 if data_len is not None and byte_counter != data_len:
 683                         raise ContentTooShortError(byte_counter, long(data_len))
 684                 self.try_rename(tmpfilename, filename)
 685
 686                 # Update file modification time
 687                 if self.params.get('updatetime', True):
 688                         info_dict['filetime'] = self.try_utime(filename, data.info().get('last-modified', None))
 689
 690                 return True