_ 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(self, percent, data_len_str, speed, eta):
 208         """Report download progress."""
 209         if self.params.get('noprogress', False):
 210             return
 211         clear_line = (u'\x1b[K' if sys.stderr.isatty() and os.name != 'nt' else u'')
 212         if eta is not None:
 213             eta_str = self.format_eta(eta)
 214         else:
 215             eta_str = 'Unknown ETA'
 216         if percent is not None:
 217             percent_str = self.format_percent(percent)
 218         else:
 219             percent_str = 'Unknown %'
 220         speed_str = self.format_speed(speed)
 221         if self.params.get('progress_with_newline', False):
 222             self.to_screen(u'[download] %s of %s at %s ETA %s' %
 223                 (percent_str, data_len_str, speed_str, eta_str))
 224         else:
 225             self.to_screen(u'\r%s[download] %s of %s at %s ETA %s' %
 226                 (clear_line, percent_str, data_len_str, speed_str, eta_str), skip_eol=True)
 227         self.to_console_title(u'youtube-dl - %s of %s at %s ETA %s' %
 228                 (percent_str.strip(), data_len_str.strip(), speed_str.strip(), eta_str.strip()))
 229
 230     def report_progress_live_stream(self, downloaded_data_len, speed, elapsed):
 231         if self.params.get('noprogress', False):
 232             return
 233         clear_line = (u'\x1b[K' if sys.stderr.isatty() and os.name != 'nt' else u'')
 234         downloaded_str = format_bytes(downloaded_data_len)
 235         speed_str = self.format_speed(speed)
 236         elapsed_str = FileDownloader.format_seconds(elapsed)
 237         if self.params.get('progress_with_newline', False):
 238             self.to_screen(u'[download] %s at %s' %
 239                 (downloaded_str, speed_str))
 240         else:
 241             self.to_screen(u'\r%s[download] %s at %s ET %s' %
 242                 (clear_line, downloaded_str, speed_str, elapsed_str), skip_eol=True)
 243         self.to_console_title(u'youtube-dl - %s at %s ET %s' %
 244                 (downloaded_str.strip(), speed_str.strip(), elapsed_str.strip()))
 245
 246     def report_resuming_byte(self, resume_len):
 247         """Report attempt to resume at given byte."""
 248         self.to_screen(u'[download] Resuming download at byte %s' % resume_len)
 249
 250     def report_retry(self, count, retries):
 251         """Report retry in case of HTTP error 5xx"""
 252         self.to_screen(u'[download] Got server HTTP error. Retrying (attempt %d of %d)...' % (count, retries))
 253
 254     def report_file_already_downloaded(self, file_name):
 255         """Report file has already been fully downloaded."""
 256         try:
 257             self.to_screen(u'[download] %s has already been downloaded' % file_name)
 258         except UnicodeEncodeError:
 259             self.to_screen(u'[download] The file has already been downloaded')
 260
 261     def report_unable_to_resume(self):
 262         """Report it was impossible to resume download."""
 263         self.to_screen(u'[download] Unable to resume')
 264
 265     def report_finish(self, data_len_str, tot_time):
 266         """Report download finished."""
 267         if self.params.get('noprogress', False):
 268             self.to_screen(u'[download] Download completed')
 269         else:
 270             clear_line = (u'\x1b[K' if sys.stderr.isatty() and os.name != 'nt' else u'')
 271             self.to_screen(u'\r%s[download] 100%% of %s in %s' %
 272                 (clear_line, data_len_str, self.format_seconds(tot_time)))
 273
 274     def _download_with_rtmpdump(self, filename, url, player_url, page_url, play_path, tc_url, live, conn):
 275         def run_rtmpdump(args):
 276             start = time.time()
 277             resume_percent = None
 278             resume_downloaded_data_len = None
 279             proc = subprocess.Popen(args, stderr=subprocess.PIPE)
 280             cursor_in_new_line = True
 281             proc_stderr_closed = False
 282             while not proc_stderr_closed:
 283                 # read line from stderr
 284                 line = u''
 285                 while True:
 286                     char = proc.stderr.read(1)
 287                     if not char:
 288                         proc_stderr_closed = True
 289                         break
 290                     if char in [b'\r', b'\n']:
 291                         break
 292                     line += char.decode('ascii', 'replace')
 293                 if not line:
 294                     # proc_stderr_closed is True
 295                     continue
 296                 mobj = re.search(r'([0-9]+\.[0-9]{3}) kB / [0-9]+\.[0-9]{2} sec \(([0-9]{1,2}\.[0-9])%\)', line)
 297                 if mobj:
 298                     downloaded_data_len = int(float(mobj.group(1))*1024)
 299                     percent = float(mobj.group(2))
 300                     if not resume_percent:
 301                         resume_percent = percent
 302                         resume_downloaded_data_len = downloaded_data_len
 303                     eta = self.calc_eta(start, time.time(), 100-resume_percent, percent-resume_percent)
 304                     speed = self.calc_speed(start, time.time(), downloaded_data_len-resume_downloaded_data_len)
 305                     data_len = None
 306                     if percent > 0:
 307                         data_len = int(downloaded_data_len * 100 / percent)
 308                     data_len_str = u'~' + format_bytes(data_len)
 309                     self.report_progress(percent, data_len_str, speed, eta)
 310                     cursor_in_new_line = False
 311                     self._hook_progress({
 312                         'downloaded_bytes': downloaded_data_len,
 313                         'total_bytes': data_len,
 314                         'tmpfilename': tmpfilename,
 315                         'filename': filename,
 316                         'status': 'downloading',
 317                         'eta': eta,
 318                         'speed': speed,
 319                     })
 320                 else:
 321                     # no percent for live streams
 322                     mobj = re.search(r'([0-9]+\.[0-9]{3}) kB / [0-9]+\.[0-9]{2} sec', line)
 323                     if mobj:
 324                         downloaded_data_len = int(float(mobj.group(1))*1024)
 325                         time_now = time.time()
 326                         speed = self.calc_speed(start, time_now, downloaded_data_len)
 327                         self.report_progress_live_stream(downloaded_data_len, speed, time_now - start)
 328                         cursor_in_new_line = False
 329                         self._hook_progress({
 330                             'downloaded_bytes': downloaded_data_len,
 331                             'tmpfilename': tmpfilename,
 332                             'filename': filename,
 333                             'status': 'downloading',
 334                             'speed': speed,
 335                         })
 336                     elif self.params.get('verbose', False):
 337                         if not cursor_in_new_line:
 338                             self.to_screen(u'')
 339                         cursor_in_new_line = True
 340                         self.to_screen(u'[rtmpdump] '+line)
 341             proc.wait()
 342             if not cursor_in_new_line:
 343                 self.to_screen(u'')
 344             return proc.returncode
 345
 346         self.report_destination(filename)
 347         tmpfilename = self.temp_name(filename)
 348         test = self.params.get('test', False)
 349
 350         # Check for rtmpdump first
 351         try:
 352             subprocess.call(['rtmpdump', '-h'], stdout=(open(os.path.devnull, 'w')), stderr=subprocess.STDOUT)
 353         except (OSError, IOError):
 354             self.report_error(u'RTMP download detected but "rtmpdump" could not be run')
 355             return False
 356
 357         # Download using rtmpdump. rtmpdump returns exit code 2 when
 358         # the connection was interrumpted and resuming appears to be
 359         # possible. This is part of rtmpdump's normal usage, AFAIK.
 360         basic_args = ['rtmpdump', '--verbose', '-r', url, '-o', tmpfilename]
 361         if player_url is not None:
 362             basic_args += ['--swfVfy', player_url]
 363         if page_url is not None:
 364             basic_args += ['--pageUrl', page_url]
 365         if play_path is not None:
 366             basic_args += ['--playpath', play_path]
 367         if tc_url is not None:
 368             basic_args += ['--tcUrl', url]
 369         if test:
 370             basic_args += ['--stop', '1']
 371         if live:
 372             basic_args += ['--live']
 373         if conn:
 374             basic_args += ['--conn', conn]
 375         args = basic_args + [[], ['--resume', '--skip', '1']][self.params.get('continuedl', False)]
 376
 377         if sys.platform == 'win32' and sys.version_info < (3, 0):
 378             # Windows subprocess module does not actually support Unicode
 379             # on Python 2.x
 380             # See http://stackoverflow.com/a/9951851/35070
 381             subprocess_encoding = sys.getfilesystemencoding()
 382             args = [a.encode(subprocess_encoding, 'ignore') for a in args]
 383         else:
 384             subprocess_encoding = None
 385
 386         if self.params.get('verbose', False):
 387             if subprocess_encoding:
 388                 str_args = [
 389                     a.decode(subprocess_encoding) if isinstance(a, bytes) else a
 390                     for a in args]
 391             else:
 392                 str_args = args
 393             try:
 394                 import pipes
 395                 shell_quote = lambda args: ' '.join(map(pipes.quote, str_args))
 396             except ImportError:
 397                 shell_quote = repr
 398             self.to_screen(u'[debug] rtmpdump command line: ' + shell_quote(str_args))
 399
 400         retval = run_rtmpdump(args)
 401
 402         while (retval == 2 or retval == 1) and not test:
 403             prevsize = os.path.getsize(encodeFilename(tmpfilename))
 404             self.to_screen(u'[rtmpdump] %s bytes' % prevsize)
 405             time.sleep(5.0) # This seems to be needed
 406             retval = run_rtmpdump(basic_args + ['-e'] + [[], ['-k', '1']][retval == 1])
 407             cursize = os.path.getsize(encodeFilename(tmpfilename))
 408             if prevsize == cursize and retval == 1:
 409                 break
 410              # Some rtmp streams seem abort after ~ 99.8%. Don't complain for those
 411             if prevsize == cursize and retval == 2 and cursize > 1024:
 412                 self.to_screen(u'[rtmpdump] Could not download the whole video. This can happen for some advertisements.')
 413                 retval = 0
 414                 break
 415         if retval == 0 or (test and retval == 2):
 416             fsize = os.path.getsize(encodeFilename(tmpfilename))
 417             self.to_screen(u'[rtmpdump] %s bytes' % fsize)
 418             self.try_rename(tmpfilename, filename)
 419             self._hook_progress({
 420                 'downloaded_bytes': fsize,
 421                 'total_bytes': fsize,
 422                 'filename': filename,
 423                 'status': 'finished',
 424             })
 425             return True
 426         else:
 427             self.to_stderr(u"\n")
 428             self.report_error(u'rtmpdump exited with code %d' % retval)
 429             return False
 430
 431     def _download_with_mplayer(self, filename, url):
 432         self.report_destination(filename)
 433         tmpfilename = self.temp_name(filename)
 434
 435         args = ['mplayer', '-really-quiet', '-vo', 'null', '-vc', 'dummy', '-dumpstream', '-dumpfile', tmpfilename, url]
 436         # Check for mplayer first
 437         try:
 438             subprocess.call(['mplayer', '-h'], stdout=(open(os.path.devnull, 'w')), stderr=subprocess.STDOUT)
 439         except (OSError, IOError):
 440             self.report_error(u'MMS or RTSP download detected but "%s" could not be run' % args[0] )
 441             return False
 442
 443         # Download using mplayer.
 444         retval = subprocess.call(args)
 445         if retval == 0:
 446             fsize = os.path.getsize(encodeFilename(tmpfilename))
 447             self.to_screen(u'\r[%s] %s bytes' % (args[0], fsize))
 448             self.try_rename(tmpfilename, filename)
 449             self._hook_progress({
 450                 'downloaded_bytes': fsize,
 451                 'total_bytes': fsize,
 452                 'filename': filename,
 453                 'status': 'finished',
 454             })
 455             return True
 456         else:
 457             self.to_stderr(u"\n")
 458             self.report_error(u'mplayer exited with code %d' % retval)
 459             return False
 460
 461     def _download_m3u8_with_ffmpeg(self, filename, url):
 462         self.report_destination(filename)
 463         tmpfilename = self.temp_name(filename)
 464
 465         args = ['-y', '-i', url, '-f', 'mp4', '-c', 'copy',
 466             '-bsf:a', 'aac_adtstoasc', tmpfilename]
 467
 468         for program in ['avconv', 'ffmpeg']:
 469             try:
 470                 subprocess.call([program, '-version'], stdout=(open(os.path.devnull, 'w')), stderr=subprocess.STDOUT)
 471                 break
 472             except (OSError, IOError):
 473                 pass
 474         else:
 475             self.report_error(u'm3u8 download detected but ffmpeg or avconv could not be found')
 476         cmd = [program] + args
 477
 478         retval = subprocess.call(cmd)
 479         if retval == 0:
 480             fsize = os.path.getsize(encodeFilename(tmpfilename))
 481             self.to_screen(u'\r[%s] %s bytes' % (args[0], fsize))
 482             self.try_rename(tmpfilename, filename)
 483             self._hook_progress({
 484                 'downloaded_bytes': fsize,
 485                 'total_bytes': fsize,
 486                 'filename': filename,
 487                 'status': 'finished',
 488             })
 489             return True
 490         else:
 491             self.to_stderr(u"\n")
 492             self.report_error(u'ffmpeg exited with code %d' % retval)
 493             return False
 494
 495
 496     def _do_download(self, filename, info_dict):
 497         url = info_dict['url']
 498
 499         # Check file already present
 500         if self.params.get('continuedl', False) and os.path.isfile(encodeFilename(filename)) and not self.params.get('nopart', False):
 501             self.report_file_already_downloaded(filename)
 502             self._hook_progress({
 503                 'filename': filename,
 504                 'status': 'finished',
 505                 'total_bytes': os.path.getsize(encodeFilename(filename)),
 506             })
 507             return True
 508
 509         # Attempt to download using rtmpdump
 510         if url.startswith('rtmp'):
 511             return self._download_with_rtmpdump(filename, url,
 512                                                 info_dict.get('player_url', None),
 513                                                 info_dict.get('page_url', None),
 514                                                 info_dict.get('play_path', None),
 515                                                 info_dict.get('tc_url', None),
 516                                                 info_dict.get('rtmp_live', False),
 517                                                 info_dict.get('rtmp_conn', None))
 518
 519         # Attempt to download using mplayer
 520         if url.startswith('mms') or url.startswith('rtsp'):
 521             return self._download_with_mplayer(filename, url)
 522
 523         # m3u8 manifest are downloaded with ffmpeg
 524         if determine_ext(url) == u'm3u8':
 525             return self._download_m3u8_with_ffmpeg(filename, url)
 526
 527         tmpfilename = self.temp_name(filename)
 528         stream = None
 529
 530         # Do not include the Accept-Encoding header
 531         headers = {'Youtubedl-no-compression': 'True'}
 532         if 'user_agent' in info_dict:
 533             headers['Youtubedl-user-agent'] = info_dict['user_agent']
 534         basic_request = compat_urllib_request.Request(url, None, headers)
 535         request = compat_urllib_request.Request(url, None, headers)
 536
 537         if self.params.get('test', False):
 538             request.add_header('Range','bytes=0-10240')
 539
 540         # Establish possible resume length
 541         if os.path.isfile(encodeFilename(tmpfilename)):
 542             resume_len = os.path.getsize(encodeFilename(tmpfilename))
 543         else:
 544             resume_len = 0
 545
 546         open_mode = 'wb'
 547         if resume_len != 0:
 548             if self.params.get('continuedl', False):
 549                 self.report_resuming_byte(resume_len)
 550                 request.add_header('Range','bytes=%d-' % resume_len)
 551                 open_mode = 'ab'
 552             else:
 553                 resume_len = 0
 554
 555         count = 0
 556         retries = self.params.get('retries', 0)
 557         while count <= retries:
 558             # Establish connection
 559             try:
 560                 if count == 0 and 'urlhandle' in info_dict:
 561                     data = info_dict['urlhandle']
 562                 data = compat_urllib_request.urlopen(request)
 563                 break
 564             except (compat_urllib_error.HTTPError, ) as err:
 565                 if (err.code < 500 or err.code >= 600) and err.code != 416:
 566                     # Unexpected HTTP error
 567                     raise
 568                 elif err.code == 416:
 569                     # Unable to resume (requested range not satisfiable)
 570                     try:
 571                         # Open the connection again without the range header
 572                         data = compat_urllib_request.urlopen(basic_request)
 573                         content_length = data.info()['Content-Length']
 574                     except (compat_urllib_error.HTTPError, ) as err:
 575                         if err.code < 500 or err.code >= 600:
 576                             raise
 577                     else:
 578                         # Examine the reported length
 579                         if (content_length is not None and
 580                                 (resume_len - 100 < int(content_length) < resume_len + 100)):
 581                             # The file had already been fully downloaded.
 582                             # Explanation to the above condition: in issue #175 it was revealed that
 583                             # YouTube sometimes adds or removes a few bytes from the end of the file,
 584                             # changing the file size slightly and causing problems for some users. So
 585                             # I decided to implement a suggested change and consider the file
 586                             # completely downloaded if the file size differs less than 100 bytes from
 587                             # the one in the hard drive.
 588                             self.report_file_already_downloaded(filename)
 589                             self.try_rename(tmpfilename, filename)
 590                             self._hook_progress({
 591                                 'filename': filename,
 592                                 'status': 'finished',
 593                             })
 594                             return True
 595                         else:
 596                             # The length does not match, we start the download over
 597                             self.report_unable_to_resume()
 598                             open_mode = 'wb'
 599                             break
 600             # Retry
 601             count += 1
 602             if count <= retries:
 603                 self.report_retry(count, retries)
 604
 605         if count > retries:
 606             self.report_error(u'giving up after %s retries' % retries)
 607             return False
 608
 609         data_len = data.info().get('Content-length', None)
 610         if data_len is not None:
 611             data_len = int(data_len) + resume_len
 612             min_data_len = self.params.get("min_filesize", None)
 613             max_data_len =  self.params.get("max_filesize", None)
 614             if min_data_len is not None and data_len < min_data_len:
 615                 self.to_screen(u'\r[download] File is smaller than min-filesize (%s bytes < %s bytes). Aborting.' % (data_len, min_data_len))
 616                 return False
 617             if max_data_len is not None and data_len > max_data_len:
 618                 self.to_screen(u'\r[download] File is larger than max-filesize (%s bytes > %s bytes). Aborting.' % (data_len, max_data_len))
 619                 return False
 620
 621         data_len_str = format_bytes(data_len)
 622         byte_counter = 0 + resume_len
 623         block_size = self.params.get('buffersize', 1024)
 624         start = time.time()
 625         while True:
 626             # Download and write
 627             before = time.time()
 628             data_block = data.read(block_size)
 629             after = time.time()
 630             if len(data_block) == 0:
 631                 break
 632             byte_counter += len(data_block)
 633
 634             # Open file just in time
 635             if stream is None:
 636                 try:
 637                     (stream, tmpfilename) = sanitize_open(tmpfilename, open_mode)
 638                     assert stream is not None
 639                     filename = self.undo_temp_name(tmpfilename)
 640                     self.report_destination(filename)
 641                 except (OSError, IOError) as err:
 642                     self.report_error(u'unable to open for writing: %s' % str(err))
 643                     return False
 644             try:
 645                 stream.write(data_block)
 646             except (IOError, OSError) as err:
 647                 self.to_stderr(u"\n")
 648                 self.report_error(u'unable to write data: %s' % str(err))
 649                 return False
 650             if not self.params.get('noresizebuffer', False):
 651                 block_size = self.best_block_size(after - before, len(data_block))
 652
 653             # Progress message
 654             speed = self.calc_speed(start, time.time(), byte_counter - resume_len)
 655             if data_len is None:
 656                 eta = percent = None
 657             else:
 658                 percent = self.calc_percent(byte_counter, data_len)
 659                 eta = self.calc_eta(start, time.time(), data_len - resume_len, byte_counter - resume_len)
 660             self.report_progress(percent, data_len_str, speed, eta)
 661
 662             self._hook_progress({
 663                 'downloaded_bytes': byte_counter,
 664                 'total_bytes': data_len,
 665                 'tmpfilename': tmpfilename,
 666                 'filename': filename,
 667                 'status': 'downloading',
 668                 'eta': eta,
 669                 'speed': speed,
 670             })
 671
 672             # Apply rate limit
 673             self.slow_down(start, byte_counter - resume_len)
 674
 675         if stream is None:
 676             self.to_stderr(u"\n")
 677             self.report_error(u'Did not get any data blocks')
 678             return False
 679         stream.close()
 680         self.report_finish(data_len_str, (time.time() - start))
 681         if data_len is not None and byte_counter != data_len:
 682             raise ContentTooShortError(byte_counter, int(data_len))
 683         self.try_rename(tmpfilename, filename)
 684
 685         # Update file modification time
 686         if self.params.get('updatetime', True):
 687             info_dict['filetime'] = self.try_utime(filename, data.info().get('last-modified', None))
 688
 689         self._hook_progress({
 690             'downloaded_bytes': byte_counter,
 691             'total_bytes': byte_counter,
 692             'filename': filename,
 693             'status': 'finished',
 694         })
 695
 696         return True
 697
 698     def _hook_progress(self, status):
 699         for ph in self._progress_hooks:
 700             ph(status)
 701
 702     def add_progress_hook(self, ph):
 703         """ ph gets called on download progress, with a dictionary with the entries
 704         * filename: The final filename
 705         * status: One of "downloading" and "finished"
 706
 707         It can also have some of the following entries:
 708
 709         * downloaded_bytes: Bytes on disks
 710         * total_bytes: Total bytes, None if unknown
 711         * tmpfilename: The filename we're currently writing to
 712         * eta: The estimated time in seconds, None if unknown
 713         * speed: The download speed in bytes/second, None if unknown
 714
 715         Hooks are guaranteed to be called at least once (with status "finished")
 716         if the download is successful.
 717         """
 718         self._progress_hooks.append(ph)