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

   1 import os
   2 import re
   3 import subprocess
   4 import sys
   5 import time
   6
   7 from .utils import (
   8     compat_urllib_error,
   9     compat_urllib_request,
  10     ContentTooShortError,
  11     determine_ext,
  12     encodeFilename,
  13     format_bytes,
  14     sanitize_open,
  15     timeconvert,
  16 )
  17
  18
  19 class FileDownloader(object):
  20     """File Downloader class.
  21
  22     File downloader objects are the ones responsible of downloading the
  23     actual video file and writing it to disk.
  24
  25     File downloaders accept a lot of parameters. In order not to saturate
  26     the object constructor with arguments, it receives a dictionary of
  27     options instead.
  28
  29     Available options:
  30
  31     verbose:           Print additional info to stdout.
  32     quiet:             Do not print messages to stdout.
  33     ratelimit:         Download speed limit, in bytes/sec.
  34     retries:           Number of times to retry for HTTP error 5xx
  35     buffersize:        Size of download buffer in bytes.
  36     noresizebuffer:    Do not automatically resize the download buffer.
  37     continuedl:        Try to continue downloads if possible.
  38     noprogress:        Do not print the progress bar.
  39     logtostderr:       Log messages to stderr instead of stdout.
  40     consoletitle:      Display progress in console window's titlebar.
  41     nopart:            Do not use temporary .part files.
  42     updatetime:        Use the Last-modified header to set output file timestamps.
  43     test:              Download only first bytes to test the downloader.
  44     min_filesize:      Skip files smaller than this size
  45     max_filesize:      Skip files larger than this size
  46     """
  47
  48     params = None
  49
  50     def __init__(self, ydl, params):
  51         """Create a FileDownloader object with the given options."""
  52         self.ydl = ydl
  53         self._progress_hooks = []
  54         self.params = params
  55
  56     @staticmethod
  57     def format_seconds(seconds):
  58         (mins, secs) = divmod(seconds, 60)
  59         (hours, mins) = divmod(mins, 60)
  60         if hours > 99:
  61             return '--:--:--'
  62         if hours == 0:
  63             return '%02d:%02d' % (mins, secs)
  64         else:
  65             return '%02d:%02d:%02d' % (hours, mins, secs)
  66
  67     @staticmethod
  68     def calc_percent(byte_counter, data_len):
  69         if data_len is None:
  70             return None
  71         return float(byte_counter) / float(data_len) * 100.0
  72
  73     @staticmethod
  74     def format_percent(percent):
  75         if percent is None:
  76             return '---.-%'
  77         return '%6s' % ('%3.1f%%' % percent)
  78
  79     @staticmethod
  80     def calc_eta(start, now, total, current):
  81         if total is None:
  82             return None
  83         dif = now - start
  84         if current == 0 or dif < 0.001: # One millisecond
  85             return None
  86         rate = float(current) / dif
  87         return int((float(total) - float(current)) / rate)
  88
  89     @staticmethod
  90     def format_eta(eta):
  91         if eta is None:
  92             return '--:--'
  93         return FileDownloader.format_seconds(eta)
  94
  95     @staticmethod
  96     def calc_speed(start, now, bytes):
  97         dif = now - start
  98         if bytes == 0 or dif < 0.001: # One millisecond
  99             return None
 100         return float(bytes) / dif
 101
 102     @staticmethod
 103     def format_speed(speed):
 104         if speed is None:
 105             return '%10s' % '---b/s'
 106         return '%10s' % ('%s/s' % format_bytes(speed))
 107
 108     @staticmethod
 109     def best_block_size(elapsed_time, bytes):
 110         new_min = max(bytes / 2.0, 1.0)
 111         new_max = min(max(bytes * 2.0, 1.0), 4194304) # Do not surpass 4 MB
 112         if elapsed_time < 0.001:
 113             return int(new_max)
 114         rate = bytes / elapsed_time
 115         if rate > new_max:
 116             return int(new_max)
 117         if rate < new_min:
 118             return int(new_min)
 119         return int(rate)
 120
 121     @staticmethod
 122     def parse_bytes(bytestr):
 123         """Parse a string indicating a byte quantity into an integer."""
 124         matchobj = re.match(r'(?i)^(\d+(?:\.\d+)?)([kMGTPEZY]?)$', bytestr)
 125         if matchobj is None:
 126             return None
 127         number = float(matchobj.group(1))
 128         multiplier = 1024.0 ** 'bkmgtpezy'.index(matchobj.group(2).lower())
 129         return int(round(number * multiplier))
 130
 131     def to_screen(self, *args, **kargs):
 132         self.ydl.to_screen(*args, **kargs)
 133
 134     def to_stderr(self, message):
 135         self.ydl.to_screen(message)
 136
 137     def to_console_title(self, message):
 138         self.ydl.to_console_title(message)
 139
 140     def trouble(self, *args, **kargs):
 141         self.ydl.trouble(*args, **kargs)
 142
 143     def report_warning(self, *args, **kargs):
 144         self.ydl.report_warning(*args, **kargs)
 145
 146     def report_error(self, *args, **kargs):
 147         self.ydl.report_error(*args, **kargs)
 148
 149     def slow_down(self, start_time, byte_counter):
 150         """Sleep if the download speed is over the rate limit."""
 151         rate_limit = self.params.get('ratelimit', None)
 152         if rate_limit is None or byte_counter == 0:
 153             return
 154         now = time.time()
 155         elapsed = now - start_time
 156         if elapsed <= 0.0:
 157             return
 158         speed = float(byte_counter) / elapsed
 159         if speed > rate_limit:
 160             time.sleep((byte_counter - rate_limit * (now - start_time)) / rate_limit)
 161
 162     def temp_name(self, filename):
 163         """Returns a temporary filename for the given filename."""
 164         if self.params.get('nopart', False) or filename == u'-' or \
 165                 (os.path.exists(encodeFilename(filename)) and not os.path.isfile(encodeFilename(filename))):
 166             return filename
 167         return filename + u'.part'
 168
 169     def undo_temp_name(self, filename):
 170         if filename.endswith(u'.part'):
 171             return filename[:-len(u'.part')]
 172         return filename
 173
 174     def try_rename(self, old_filename, new_filename):
 175         try:
 176             if old_filename == new_filename:
 177                 return
 178             os.rename(encodeFilename(old_filename), encodeFilename(new_filename))
 179         except (IOError, OSError):
 180             self.report_error(u'unable to rename file')
 181
 182     def try_utime(self, filename, last_modified_hdr):
 183         """Try to set the last-modified time of the given file."""
 184         if last_modified_hdr is None:
 185             return
 186         if not os.path.isfile(encodeFilename(filename)):
 187             return
 188         timestr = last_modified_hdr
 189         if timestr is None:
 190             return
 191         filetime = timeconvert(timestr)
 192         if filetime is None:
 193             return filetime
 194         # Ignore obviously invalid dates
 195         if filetime == 0:
 196             return
 197         try:
 198             os.utime(filename, (time.time(), filetime))
 199         except:
 200             pass
 201         return filetime
 202
 203     def report_destination(self, filename):
 204         """Report destination filename."""
 205         self.to_screen(u'[download] Destination: ' + filename)
 206
 207     def _report_progress_status(self, msg, is_last_line=False):
 208         fullmsg = u'[download] ' + msg
 209         if self.params.get('progress_with_newline', False):
 210             self.to_screen(fullmsg)
 211         else:
 212             if os.name == 'nt':
 213                 prev_len = getattr(self, '_report_progress_prev_line_length',
 214                                    0)
 215                 if prev_len > len(fullmsg):
 216                     fullmsg += u' ' * (prev_len - len(fullmsg))
 217                 self._report_progress_prev_line_length = len(fullmsg)
 218                 clear_line = u'\r'
 219             else:
 220                 clear_line = (u'\r\x1b[K' if sys.stderr.isatty() else u'\r')
 221             self.to_screen(clear_line + fullmsg, skip_eol=not is_last_line)
 222         self.to_console_title(u'youtube-dl ' + msg)
 223
 224     def report_progress(self, percent, data_len_str, speed, eta):
 225         """Report download progress."""
 226         if self.params.get('noprogress', False):
 227             return
 228         if eta is not None:
 229             eta_str = self.format_eta(eta)
 230         else:
 231             eta_str = 'Unknown ETA'
 232         if percent is not None:
 233             percent_str = self.format_percent(percent)
 234         else:
 235             percent_str = 'Unknown %'
 236         speed_str = self.format_speed(speed)
 237
 238         msg = (u'%s of %s at %s ETA %s' %
 239                (percent_str, data_len_str, speed_str, eta_str))
 240         self._report_progress_status(msg)
 241
 242     def report_progress_live_stream(self, downloaded_data_len, speed, elapsed):
 243         if self.params.get('noprogress', False):
 244             return
 245         downloaded_str = format_bytes(downloaded_data_len)
 246         speed_str = self.format_speed(speed)
 247         elapsed_str = FileDownloader.format_seconds(elapsed)
 248         msg = u'%s at %s (%s)' % (downloaded_str, speed_str, elapsed_str)
 249         self._report_progress_status(msg)
 250
 251     def report_finish(self, data_len_str, tot_time):
 252         """Report download finished."""
 253         if self.params.get('noprogress', False):
 254             self.to_screen(u'[download] Download completed')
 255         else:
 256             self._report_progress_status(
 257                 (u'100%% of %s in %s' %
 258                  (data_len_str, self.format_seconds(tot_time))),
 259                 is_last_line=True)
 260
 261     def report_resuming_byte(self, resume_len):
 262         """Report attempt to resume at given byte."""
 263         self.to_screen(u'[download] Resuming download at byte %s' % resume_len)
 264
 265     def report_retry(self, count, retries):
 266         """Report retry in case of HTTP error 5xx"""
 267         self.to_screen(u'[download] Got server HTTP error. Retrying (attempt %d of %d)...' % (count, retries))
 268
 269     def report_file_already_downloaded(self, file_name):
 270         """Report file has already been fully downloaded."""
 271         try:
 272             self.to_screen(u'[download] %s has already been downloaded' % file_name)
 273         except UnicodeEncodeError:
 274             self.to_screen(u'[download] The file has already been downloaded')
 275
 276     def report_unable_to_resume(self):
 277         """Report it was impossible to resume download."""
 278         self.to_screen(u'[download] Unable to resume')
 279
 280     def _download_with_rtmpdump(self, filename, url, player_url, page_url, play_path, tc_url, live, conn):
 281         def run_rtmpdump(args):
 282             start = time.time()
 283             resume_percent = None
 284             resume_downloaded_data_len = None
 285             proc = subprocess.Popen(args, stderr=subprocess.PIPE)
 286             cursor_in_new_line = True
 287             proc_stderr_closed = False
 288             while not proc_stderr_closed:
 289                 # read line from stderr
 290                 line = u''
 291                 while True:
 292                     char = proc.stderr.read(1)
 293                     if not char:
 294                         proc_stderr_closed = True
 295                         break
 296                     if char in [b'\r', b'\n']:
 297                         break
 298                     line += char.decode('ascii', 'replace')
 299                 if not line:
 300                     # proc_stderr_closed is True
 301                     continue
 302                 mobj = re.search(r'([0-9]+\.[0-9]{3}) kB / [0-9]+\.[0-9]{2} sec \(([0-9]{1,2}\.[0-9])%\)', line)
 303                 if mobj:
 304                     downloaded_data_len = int(float(mobj.group(1))*1024)
 305                     percent = float(mobj.group(2))
 306                     if not resume_percent:
 307                         resume_percent = percent
 308                         resume_downloaded_data_len = downloaded_data_len
 309                     eta = self.calc_eta(start, time.time(), 100-resume_percent, percent-resume_percent)
 310                     speed = self.calc_speed(start, time.time(), downloaded_data_len-resume_downloaded_data_len)
 311                     data_len = None
 312                     if percent > 0:
 313                         data_len = int(downloaded_data_len * 100 / percent)
 314                     data_len_str = u'~' + format_bytes(data_len)
 315                     self.report_progress(percent, data_len_str, speed, eta)
 316                     cursor_in_new_line = False
 317                     self._hook_progress({
 318                         'downloaded_bytes': downloaded_data_len,
 319                         'total_bytes': data_len,
 320                         'tmpfilename': tmpfilename,
 321                         'filename': filename,
 322                         'status': 'downloading',
 323                         'eta': eta,
 324                         'speed': speed,
 325                     })
 326                 else:
 327                     # no percent for live streams
 328                     mobj = re.search(r'([0-9]+\.[0-9]{3}) kB / [0-9]+\.[0-9]{2} sec', line)
 329                     if mobj:
 330                         downloaded_data_len = int(float(mobj.group(1))*1024)
 331                         time_now = time.time()
 332                         speed = self.calc_speed(start, time_now, downloaded_data_len)
 333                         self.report_progress_live_stream(downloaded_data_len, speed, time_now - start)
 334                         cursor_in_new_line = False
 335                         self._hook_progress({
 336                             'downloaded_bytes': downloaded_data_len,
 337                             'tmpfilename': tmpfilename,
 338                             'filename': filename,
 339                             'status': 'downloading',
 340                             'speed': speed,
 341                         })
 342                     elif self.params.get('verbose', False):
 343                         if not cursor_in_new_line:
 344                             self.to_screen(u'')
 345                         cursor_in_new_line = True
 346                         self.to_screen(u'[rtmpdump] '+line)
 347             proc.wait()
 348             if not cursor_in_new_line:
 349                 self.to_screen(u'')
 350             return proc.returncode
 351
 352         self.report_destination(filename)
 353         tmpfilename = self.temp_name(filename)
 354         test = self.params.get('test', False)
 355
 356         # Check for rtmpdump first
 357         try:
 358             subprocess.call(['rtmpdump', '-h'], stdout=(open(os.path.devnull, 'w')), stderr=subprocess.STDOUT)
 359         except (OSError, IOError):
 360             self.report_error(u'RTMP download detected but "rtmpdump" could not be run')
 361             return False
 362
 363         # Download using rtmpdump. rtmpdump returns exit code 2 when
 364         # the connection was interrumpted and resuming appears to be
 365         # possible. This is part of rtmpdump's normal usage, AFAIK.
 366         basic_args = ['rtmpdump', '--verbose', '-r', url, '-o', tmpfilename]
 367         if player_url is not None:
 368             basic_args += ['--swfVfy', player_url]
 369         if page_url is not None:
 370             basic_args += ['--pageUrl', page_url]
 371         if play_path is not None:
 372             basic_args += ['--playpath', play_path]
 373         if tc_url is not None:
 374             basic_args += ['--tcUrl', url]
 375         if test:
 376             basic_args += ['--stop', '1']
 377         if live:
 378             basic_args += ['--live']
 379         if conn:
 380             basic_args += ['--conn', conn]
 381         args = basic_args + [[], ['--resume', '--skip', '1']][self.params.get('continuedl', False)]
 382
 383         if sys.platform == 'win32' and sys.version_info < (3, 0):
 384             # Windows subprocess module does not actually support Unicode
 385             # on Python 2.x
 386             # See http://stackoverflow.com/a/9951851/35070
 387             subprocess_encoding = sys.getfilesystemencoding()
 388             args = [a.encode(subprocess_encoding, 'ignore') for a in args]
 389         else:
 390             subprocess_encoding = None
 391
 392         if self.params.get('verbose', False):
 393             if subprocess_encoding:
 394                 str_args = [
 395                     a.decode(subprocess_encoding) if isinstance(a, bytes) else a
 396                     for a in args]
 397             else:
 398                 str_args = args
 399             try:
 400                 import pipes
 401                 shell_quote = lambda args: ' '.join(map(pipes.quote, str_args))
 402             except ImportError:
 403                 shell_quote = repr
 404             self.to_screen(u'[debug] rtmpdump command line: ' + shell_quote(str_args))
 405
 406         retval = run_rtmpdump(args)
 407
 408         while (retval == 2 or retval == 1) and not test:
 409             prevsize = os.path.getsize(encodeFilename(tmpfilename))
 410             self.to_screen(u'[rtmpdump] %s bytes' % prevsize)
 411             time.sleep(5.0) # This seems to be needed
 412             retval = run_rtmpdump(basic_args + ['-e'] + [[], ['-k', '1']][retval == 1])
 413             cursize = os.path.getsize(encodeFilename(tmpfilename))
 414             if prevsize == cursize and retval == 1:
 415                 break
 416              # Some rtmp streams seem abort after ~ 99.8%. Don't complain for those
 417             if prevsize == cursize and retval == 2 and cursize > 1024:
 418                 self.to_screen(u'[rtmpdump] Could not download the whole video. This can happen for some advertisements.')
 419                 retval = 0
 420                 break
 421         if retval == 0 or (test and retval == 2):
 422             fsize = os.path.getsize(encodeFilename(tmpfilename))
 423             self.to_screen(u'[rtmpdump] %s bytes' % fsize)
 424             self.try_rename(tmpfilename, filename)
 425             self._hook_progress({
 426                 'downloaded_bytes': fsize,
 427                 'total_bytes': fsize,
 428                 'filename': filename,
 429                 'status': 'finished',
 430             })
 431             return True
 432         else:
 433             self.to_stderr(u"\n")
 434             self.report_error(u'rtmpdump exited with code %d' % retval)
 435             return False
 436
 437     def _download_with_mplayer(self, filename, url):
 438         self.report_destination(filename)
 439         tmpfilename = self.temp_name(filename)
 440
 441         args = ['mplayer', '-really-quiet', '-vo', 'null', '-vc', 'dummy', '-dumpstream', '-dumpfile', tmpfilename, url]
 442         # Check for mplayer first
 443         try:
 444             subprocess.call(['mplayer', '-h'], stdout=(open(os.path.devnull, 'w')), stderr=subprocess.STDOUT)
 445         except (OSError, IOError):
 446             self.report_error(u'MMS or RTSP download detected but "%s" could not be run' % args[0] )
 447             return False
 448
 449         # Download using mplayer.
 450         retval = subprocess.call(args)
 451         if retval == 0:
 452             fsize = os.path.getsize(encodeFilename(tmpfilename))
 453             self.to_screen(u'\r[%s] %s bytes' % (args[0], fsize))
 454             self.try_rename(tmpfilename, filename)
 455             self._hook_progress({
 456                 'downloaded_bytes': fsize,
 457                 'total_bytes': fsize,
 458                 'filename': filename,
 459                 'status': 'finished',
 460             })
 461             return True
 462         else:
 463             self.to_stderr(u"\n")
 464             self.report_error(u'mplayer exited with code %d' % retval)
 465             return False
 466
 467     def _download_m3u8_with_ffmpeg(self, filename, url):
 468         self.report_destination(filename)
 469         tmpfilename = self.temp_name(filename)
 470
 471         args = ['-y', '-i', url, '-f', 'mp4', '-c', 'copy',
 472             '-bsf:a', 'aac_adtstoasc', tmpfilename]
 473
 474         for program in ['avconv', 'ffmpeg']:
 475             try:
 476                 subprocess.call([program, '-version'], stdout=(open(os.path.devnull, 'w')), stderr=subprocess.STDOUT)
 477                 break
 478             except (OSError, IOError):
 479                 pass
 480         else:
 481             self.report_error(u'm3u8 download detected but ffmpeg or avconv could not be found')
 482         cmd = [program] + args
 483
 484         retval = subprocess.call(cmd)
 485         if retval == 0:
 486             fsize = os.path.getsize(encodeFilename(tmpfilename))
 487             self.to_screen(u'\r[%s] %s bytes' % (args[0], fsize))
 488             self.try_rename(tmpfilename, filename)
 489             self._hook_progress({
 490                 'downloaded_bytes': fsize,
 491                 'total_bytes': fsize,
 492                 'filename': filename,
 493                 'status': 'finished',
 494             })
 495             return True
 496         else:
 497             self.to_stderr(u"\n")
 498             self.report_error(u'ffmpeg exited with code %d' % retval)
 499             return False
 500
 501
 502     def _do_download(self, filename, info_dict):
 503         url = info_dict['url']
 504
 505         # Check file already present
 506         if self.params.get('continuedl', False) and os.path.isfile(encodeFilename(filename)) and not self.params.get('nopart', False):
 507             self.report_file_already_downloaded(filename)
 508             self._hook_progress({
 509                 'filename': filename,
 510                 'status': 'finished',
 511                 'total_bytes': os.path.getsize(encodeFilename(filename)),
 512             })
 513             return True
 514
 515         # Attempt to download using rtmpdump
 516         if url.startswith('rtmp'):
 517             return self._download_with_rtmpdump(filename, url,
 518                                                 info_dict.get('player_url', None),
 519                                                 info_dict.get('page_url', None),
 520                                                 info_dict.get('play_path', None),
 521                                                 info_dict.get('tc_url', None),
 522                                                 info_dict.get('rtmp_live', False),
 523                                                 info_dict.get('rtmp_conn', None))
 524
 525         # Attempt to download using mplayer
 526         if url.startswith('mms') or url.startswith('rtsp'):
 527             return self._download_with_mplayer(filename, url)
 528
 529         # m3u8 manifest are downloaded with ffmpeg
 530         if determine_ext(url) == u'm3u8':
 531             return self._download_m3u8_with_ffmpeg(filename, url)
 532
 533         tmpfilename = self.temp_name(filename)
 534         stream = None
 535
 536         # Do not include the Accept-Encoding header
 537         headers = {'Youtubedl-no-compression': 'True'}
 538         if 'user_agent' in info_dict:
 539             headers['Youtubedl-user-agent'] = info_dict['user_agent']
 540         basic_request = compat_urllib_request.Request(url, None, headers)
 541         request = compat_urllib_request.Request(url, None, headers)
 542
 543         if self.params.get('test', False):
 544             request.add_header('Range','bytes=0-10240')
 545
 546         # Establish possible resume length
 547         if os.path.isfile(encodeFilename(tmpfilename)):
 548             resume_len = os.path.getsize(encodeFilename(tmpfilename))
 549         else:
 550             resume_len = 0
 551
 552         open_mode = 'wb'
 553         if resume_len != 0:
 554             if self.params.get('continuedl', False):
 555                 self.report_resuming_byte(resume_len)
 556                 request.add_header('Range','bytes=%d-' % resume_len)
 557                 open_mode = 'ab'
 558             else:
 559                 resume_len = 0
 560
 561         count = 0
 562         retries = self.params.get('retries', 0)
 563         while count <= retries:
 564             # Establish connection
 565             try:
 566                 if count == 0 and 'urlhandle' in info_dict:
 567                     data = info_dict['urlhandle']
 568                 data = compat_urllib_request.urlopen(request)
 569                 break
 570             except (compat_urllib_error.HTTPError, ) as err:
 571                 if (err.code < 500 or err.code >= 600) and err.code != 416:
 572                     # Unexpected HTTP error
 573                     raise
 574                 elif err.code == 416:
 575                     # Unable to resume (requested range not satisfiable)
 576                     try:
 577                         # Open the connection again without the range header
 578                         data = compat_urllib_request.urlopen(basic_request)
 579                         content_length = data.info()['Content-Length']
 580                     except (compat_urllib_error.HTTPError, ) as err:
 581                         if err.code < 500 or err.code >= 600:
 582                             raise
 583                     else:
 584                         # Examine the reported length
 585                         if (content_length is not None and
 586                                 (resume_len - 100 < int(content_length) < resume_len + 100)):
 587                             # The file had already been fully downloaded.
 588                             # Explanation to the above condition: in issue #175 it was revealed that
 589                             # YouTube sometimes adds or removes a few bytes from the end of the file,
 590                             # changing the file size slightly and causing problems for some users. So
 591                             # I decided to implement a suggested change and consider the file
 592                             # completely downloaded if the file size differs less than 100 bytes from
 593                             # the one in the hard drive.
 594                             self.report_file_already_downloaded(filename)
 595                             self.try_rename(tmpfilename, filename)
 596                             self._hook_progress({
 597                                 'filename': filename,
 598                                 'status': 'finished',
 599                             })
 600                             return True
 601                         else:
 602                             # The length does not match, we start the download over
 603                             self.report_unable_to_resume()
 604                             open_mode = 'wb'
 605                             break
 606             # Retry
 607             count += 1
 608             if count <= retries:
 609                 self.report_retry(count, retries)
 610
 611         if count > retries:
 612             self.report_error(u'giving up after %s retries' % retries)
 613             return False
 614
 615         data_len = data.info().get('Content-length', None)
 616         if data_len is not None:
 617             data_len = int(data_len) + resume_len
 618             min_data_len = self.params.get("min_filesize", None)
 619             max_data_len =  self.params.get("max_filesize", None)
 620             if min_data_len is not None and data_len < min_data_len:
 621                 self.to_screen(u'\r[download] File is smaller than min-filesize (%s bytes < %s bytes). Aborting.' % (data_len, min_data_len))
 622                 return False
 623             if max_data_len is not None and data_len > max_data_len:
 624                 self.to_screen(u'\r[download] File is larger than max-filesize (%s bytes > %s bytes). Aborting.' % (data_len, max_data_len))
 625                 return False
 626
 627         data_len_str = format_bytes(data_len)
 628         byte_counter = 0 + resume_len
 629         block_size = self.params.get('buffersize', 1024)
 630         start = time.time()
 631         while True:
 632             # Download and write
 633             before = time.time()
 634             data_block = data.read(block_size)
 635             after = time.time()
 636             if len(data_block) == 0:
 637                 break
 638             byte_counter += len(data_block)
 639
 640             # Open file just in time
 641             if stream is None:
 642                 try:
 643                     (stream, tmpfilename) = sanitize_open(tmpfilename, open_mode)
 644                     assert stream is not None
 645                     filename = self.undo_temp_name(tmpfilename)
 646                     self.report_destination(filename)
 647                 except (OSError, IOError) as err:
 648                     self.report_error(u'unable to open for writing: %s' % str(err))
 649                     return False
 650             try:
 651                 stream.write(data_block)
 652             except (IOError, OSError) as err:
 653                 self.to_stderr(u"\n")
 654                 self.report_error(u'unable to write data: %s' % str(err))
 655                 return False
 656             if not self.params.get('noresizebuffer', False):
 657                 block_size = self.best_block_size(after - before, len(data_block))
 658
 659             # Progress message
 660             speed = self.calc_speed(start, time.time(), byte_counter - resume_len)
 661             if data_len is None:
 662                 eta = percent = None
 663             else:
 664                 percent = self.calc_percent(byte_counter, data_len)
 665                 eta = self.calc_eta(start, time.time(), data_len - resume_len, byte_counter - resume_len)
 666             self.report_progress(percent, data_len_str, speed, eta)
 667
 668             self._hook_progress({
 669                 'downloaded_bytes': byte_counter,
 670                 'total_bytes': data_len,
 671                 'tmpfilename': tmpfilename,
 672                 'filename': filename,
 673                 'status': 'downloading',
 674                 'eta': eta,
 675                 'speed': speed,
 676             })
 677
 678             # Apply rate limit
 679             self.slow_down(start, byte_counter - resume_len)
 680
 681         if stream is None:
 682             self.to_stderr(u"\n")
 683             self.report_error(u'Did not get any data blocks')
 684             return False
 685         stream.close()
 686         self.report_finish(data_len_str, (time.time() - start))
 687         if data_len is not None and byte_counter != data_len:
 688             raise ContentTooShortError(byte_counter, int(data_len))
 689         self.try_rename(tmpfilename, filename)
 690
 691         # Update file modification time
 692         if self.params.get('updatetime', True):
 693             info_dict['filetime'] = self.try_utime(filename, data.info().get('last-modified', None))
 694
 695         self._hook_progress({
 696             'downloaded_bytes': byte_counter,
 697             'total_bytes': byte_counter,
 698             'filename': filename,
 699             'status': 'finished',
 700         })
 701
 702         return True
 703
 704     def _hook_progress(self, status):
 705         for ph in self._progress_hooks:
 706             ph(status)
 707
 708     def add_progress_hook(self, ph):
 709         """ ph gets called on download progress, with a dictionary with the entries
 710         * filename: The final filename
 711         * status: One of "downloading" and "finished"
 712
 713         It can also have some of the following entries:
 714
 715         * downloaded_bytes: Bytes on disks
 716         * total_bytes: Total bytes, None if unknown
 717         * tmpfilename: The filename we're currently writing to
 718         * eta: The estimated time in seconds, None if unknown
 719         * speed: The download speed in bytes/second, None if unknown
 720
 721         Hooks are guaranteed to be called at least once (with status "finished")
 722         if the download is successful.
 723         """
 724         self._progress_hooks.append(ph)