2 # -*- coding: utf-8 -*-
4 from __future__ import absolute_import
22 class FileDownloader(object):
23 """File Downloader class.
25 File downloader objects are the ones responsible of downloading the
26 actual video file and writing it to disk if the user has requested
27 it, among some other tasks. In most cases there should be one per
28 program. As, given a video URL, the downloader doesn't know how to
29 extract all the needed information, task that InfoExtractors do, it
30 has to pass the URL to one of them.
32 For this, file downloader objects have a method that allows
33 InfoExtractors to be registered in a given order. When it is passed
34 a URL, the file downloader handles it to the first InfoExtractor it
35 finds that reports being able to handle it. The InfoExtractor extracts
36 all the information about the video or videos the URL refers to, and
37 asks the FileDownloader to process the video information, possibly
38 downloading the video.
40 File downloaders accept a lot of parameters. In order not to saturate
41 the object constructor with arguments, it receives a dictionary of
42 options instead. These options are available through the params
43 attribute for the InfoExtractors to use. The FileDownloader also
44 registers itself as the downloader in charge for the InfoExtractors
45 that are added to it, so this is a "mutual registration".
49 username: Username for authentication purposes.
50 password: Password for authentication purposes.
51 usenetrc: Use netrc for authentication instead.
52 quiet: Do not print messages to stdout.
53 forceurl: Force printing final URL.
54 forcetitle: Force printing title.
55 forcethumbnail: Force printing thumbnail URL.
56 forcedescription: Force printing description.
57 forcefilename: Force printing final filename.
58 simulate: Do not download the video files.
59 format: Video format code.
60 format_limit: Highest quality format to try.
61 outtmpl: Template for output names.
62 restrictfilenames: Do not allow "&" and spaces in file names
63 ignoreerrors: Do not stop on download errors.
64 ratelimit: Download speed limit, in bytes/sec.
65 nooverwrites: Prevent overwriting files.
66 retries: Number of times to retry for HTTP error 5xx
67 buffersize: Size of download buffer in bytes.
68 noresizebuffer: Do not automatically resize the download buffer.
69 continuedl: Try to continue downloads if possible.
70 noprogress: Do not print the progress bar.
71 playliststart: Playlist item to start at.
72 playlistend: Playlist item to end at.
73 matchtitle: Download only matching titles.
74 rejecttitle: Reject downloads for matching titles.
75 logtostderr: Log messages to stderr instead of stdout.
76 consoletitle: Display progress in console window's titlebar.
77 nopart: Do not use temporary .part files.
78 updatetime: Use the Last-modified header to set output file timestamps.
79 writedescription: Write the video description to a .description file
80 writeinfojson: Write the video description to a .info.json file
81 writesubtitles: Write the video subtitles to a .srt file
82 subtitleslang: Language of the subtitles to download
83 test: Download only first bytes to test the downloader.
89 _download_retcode = None
93 def __init__(self, params):
94 """Create a FileDownloader object with the given options."""
97 self._download_retcode = 0
98 self._num_downloads = 0
99 self._screen_file = [sys.stdout, sys.stderr][params.get('logtostderr', False)]
102 if '%(stitle)s' in self.params['outtmpl']:
103 self.to_stderr(u'WARNING: %(stitle)s is deprecated. Use the %(title)s and the --restrict-filenames flag(which also secures %(uploader)s et al) instead.')
106 def format_bytes(bytes):
109 if type(bytes) is str:
114 exponent = int(math.log(bytes, 1024.0))
115 suffix = 'bkMGTPEZY'[exponent]
116 converted = float(bytes) / float(1024 ** exponent)
117 return '%.2f%s' % (converted, suffix)
120 def calc_percent(byte_counter, data_len):
123 return '%6s' % ('%3.1f%%' % (float(byte_counter) / float(data_len) * 100.0))
126 def calc_eta(start, now, total, current):
130 if current == 0 or dif < 0.001: # One millisecond
132 rate = float(current) / dif
133 eta = int((float(total) - float(current)) / rate)
134 (eta_mins, eta_secs) = divmod(eta, 60)
137 return '%02d:%02d' % (eta_mins, eta_secs)
140 def calc_speed(start, now, bytes):
142 if bytes == 0 or dif < 0.001: # One millisecond
143 return '%10s' % '---b/s'
144 return '%10s' % ('%s/s' % FileDownloader.format_bytes(float(bytes) / dif))
147 def best_block_size(elapsed_time, bytes):
148 new_min = max(bytes / 2.0, 1.0)
149 new_max = min(max(bytes * 2.0, 1.0), 4194304) # Do not surpass 4 MB
150 if elapsed_time < 0.001:
152 rate = bytes / elapsed_time
160 def parse_bytes(bytestr):
161 """Parse a string indicating a byte quantity into an integer."""
162 matchobj = re.match(r'(?i)^(\d+(?:\.\d+)?)([kMGTPEZY]?)$', bytestr)
165 number = float(matchobj.group(1))
166 multiplier = 1024.0 ** 'bkmgtpezy'.index(matchobj.group(2).lower())
167 return int(round(number * multiplier))
169 def add_info_extractor(self, ie):
170 """Add an InfoExtractor object to the end of the list."""
172 ie.set_downloader(self)
174 def add_post_processor(self, pp):
175 """Add a PostProcessor object to the end of the chain."""
177 pp.set_downloader(self)
179 def to_screen(self, message, skip_eol=False):
180 """Print message to stdout if not in quiet mode."""
181 assert type(message) == type(u'')
182 if not self.params.get('quiet', False):
183 terminator = [u'\n', u''][skip_eol]
184 output = message + terminator
185 if 'b' in getattr(self._screen_file, 'mode', '') or sys.version_info[0] < 3: # Python 2 lies about the mode of sys.stdout/sys.stderr
186 output = output.encode(preferredencoding(), 'ignore')
187 self._screen_file.write(output)
188 self._screen_file.flush()
190 def to_stderr(self, message):
191 """Print message to stderr."""
192 assert type(message) == type(u'')
193 output = message + u'\n'
194 if 'b' in getattr(self._screen_file, 'mode', '') or sys.version_info[0] < 3: # Python 2 lies about the mode of sys.stdout/sys.stderr
195 output = output.encode(preferredencoding())
196 sys.stderr.write(output)
198 def to_cons_title(self, message):
199 """Set console/terminal window title to message."""
200 if not self.params.get('consoletitle', False):
202 if os.name == 'nt' and ctypes.windll.kernel32.GetConsoleWindow():
203 # c_wchar_p() might not be necessary if `message` is
204 # already of type unicode()
205 ctypes.windll.kernel32.SetConsoleTitleW(ctypes.c_wchar_p(message))
206 elif 'TERM' in os.environ:
207 sys.stderr.write('\033]0;%s\007' % message.encode(preferredencoding()))
209 def fixed_template(self):
210 """Checks if the output template is fixed."""
211 return (re.search(u'(?u)%\\(.+?\\)s', self.params['outtmpl']) is None)
213 def trouble(self, message=None, tb=None):
214 """Determine action to take when a download problem appears.
216 Depending on if the downloader has been configured to ignore
217 download errors or not, this method may throw an exception or
218 not when errors are found, after printing the message.
220 tb, if given, is additional traceback information.
222 if message is not None:
223 self.to_stderr(message)
224 if self.params.get('verbose'):
226 tb_data = traceback.format_list(traceback.extract_stack())
227 tb = u''.join(tb_data)
229 if not self.params.get('ignoreerrors', False):
230 raise DownloadError(message)
231 self._download_retcode = 1
233 def slow_down(self, start_time, byte_counter):
234 """Sleep if the download speed is over the rate limit."""
235 rate_limit = self.params.get('ratelimit', None)
236 if rate_limit is None or byte_counter == 0:
239 elapsed = now - start_time
242 speed = float(byte_counter) / elapsed
243 if speed > rate_limit:
244 time.sleep((byte_counter - rate_limit * (now - start_time)) / rate_limit)
246 def temp_name(self, filename):
247 """Returns a temporary filename for the given filename."""
248 if self.params.get('nopart', False) or filename == u'-' or \
249 (os.path.exists(encodeFilename(filename)) and not os.path.isfile(encodeFilename(filename))):
251 return filename + u'.part'
253 def undo_temp_name(self, filename):
254 if filename.endswith(u'.part'):
255 return filename[:-len(u'.part')]
258 def try_rename(self, old_filename, new_filename):
260 if old_filename == new_filename:
262 os.rename(encodeFilename(old_filename), encodeFilename(new_filename))
263 except (IOError, OSError) as err:
264 self.trouble(u'ERROR: unable to rename file')
266 def try_utime(self, filename, last_modified_hdr):
267 """Try to set the last-modified time of the given file."""
268 if last_modified_hdr is None:
270 if not os.path.isfile(encodeFilename(filename)):
272 timestr = last_modified_hdr
275 filetime = timeconvert(timestr)
279 os.utime(filename, (time.time(), filetime))
284 def report_writedescription(self, descfn):
285 """ Report that the description file is being written """
286 self.to_screen(u'[info] Writing video description to: ' + descfn)
288 def report_writesubtitles(self, srtfn):
289 """ Report that the subtitles file is being written """
290 self.to_screen(u'[info] Writing video subtitles to: ' + srtfn)
292 def report_writeinfojson(self, infofn):
293 """ Report that the metadata file has been written """
294 self.to_screen(u'[info] Video description metadata as JSON to: ' + infofn)
296 def report_destination(self, filename):
297 """Report destination filename."""
298 self.to_screen(u'[download] Destination: ' + filename)
300 def report_progress(self, percent_str, data_len_str, speed_str, eta_str):
301 """Report download progress."""
302 if self.params.get('noprogress', False):
304 self.to_screen(u'\r[download] %s of %s at %s ETA %s' %
305 (percent_str, data_len_str, speed_str, eta_str), skip_eol=True)
306 self.to_cons_title(u'youtube-dl - %s of %s at %s ETA %s' %
307 (percent_str.strip(), data_len_str.strip(), speed_str.strip(), eta_str.strip()))
309 def report_resuming_byte(self, resume_len):
310 """Report attempt to resume at given byte."""
311 self.to_screen(u'[download] Resuming download at byte %s' % resume_len)
313 def report_retry(self, count, retries):
314 """Report retry in case of HTTP error 5xx"""
315 self.to_screen(u'[download] Got server HTTP error. Retrying (attempt %d of %d)...' % (count, retries))
317 def report_file_already_downloaded(self, file_name):
318 """Report file has already been fully downloaded."""
320 self.to_screen(u'[download] %s has already been downloaded' % file_name)
321 except (UnicodeEncodeError) as err:
322 self.to_screen(u'[download] The file has already been downloaded')
324 def report_unable_to_resume(self):
325 """Report it was impossible to resume download."""
326 self.to_screen(u'[download] Unable to resume')
328 def report_finish(self):
329 """Report download finished."""
330 if self.params.get('noprogress', False):
331 self.to_screen(u'[download] Download completed')
335 def increment_downloads(self):
336 """Increment the ordinal that assigns a number to each file."""
337 self._num_downloads += 1
339 def prepare_filename(self, info_dict):
340 """Generate the output filename."""
342 template_dict = dict(info_dict)
344 template_dict['epoch'] = int(time.time())
345 template_dict['autonumber'] = u'%05d' % self._num_downloads
347 sanitize = lambda k,v: sanitize_filename(
348 u'NA' if v is None else compat_str(v),
349 restricted=self.params.get('restrictfilenames'),
351 template_dict = dict((k, sanitize(k, v)) for k,v in template_dict.items())
353 filename = self.params['outtmpl'] % template_dict
355 except (ValueError, KeyError) as err:
356 self.trouble(u'ERROR: invalid system charset or erroneous output template')
359 def _match_entry(self, info_dict):
360 """ Returns None iff the file should be downloaded """
362 title = info_dict['title']
363 matchtitle = self.params.get('matchtitle', False)
365 matchtitle = matchtitle.decode('utf8')
366 if not re.search(matchtitle, title, re.IGNORECASE):
367 return u'[download] "' + title + '" title did not match pattern "' + matchtitle + '"'
368 rejecttitle = self.params.get('rejecttitle', False)
370 rejecttitle = rejecttitle.decode('utf8')
371 if re.search(rejecttitle, title, re.IGNORECASE):
372 return u'"' + title + '" title matched reject pattern "' + rejecttitle + '"'
375 def process_info(self, info_dict):
376 """Process a single dictionary returned by an InfoExtractor."""
378 # Keep for backwards compatibility
379 info_dict['stitle'] = info_dict['title']
381 if not 'format' in info_dict:
382 info_dict['format'] = info_dict['ext']
384 reason = self._match_entry(info_dict)
385 if reason is not None:
386 self.to_screen(u'[download] ' + reason)
389 max_downloads = self.params.get('max_downloads')
390 if max_downloads is not None:
391 if self._num_downloads > int(max_downloads):
392 raise MaxDownloadsReached()
394 filename = self.prepare_filename(info_dict)
397 if self.params.get('forcetitle', False):
398 compat_print(info_dict['title'])
399 if self.params.get('forceurl', False):
400 compat_print(info_dict['url'])
401 if self.params.get('forcethumbnail', False) and 'thumbnail' in info_dict:
402 compat_print(info_dict['thumbnail'])
403 if self.params.get('forcedescription', False) and 'description' in info_dict:
404 compat_print(info_dict['description'])
405 if self.params.get('forcefilename', False) and filename is not None:
406 compat_print(filename)
407 if self.params.get('forceformat', False):
408 compat_print(info_dict['format'])
410 # Do nothing else if in simulate mode
411 if self.params.get('simulate', False):
418 dn = os.path.dirname(encodeFilename(filename))
419 if dn != '' and not os.path.exists(dn): # dn is already encoded
421 except (OSError, IOError) as err:
422 self.trouble(u'ERROR: unable to create directory ' + compat_str(err))
425 if self.params.get('writedescription', False):
427 descfn = filename + u'.description'
428 self.report_writedescription(descfn)
429 with io.open(encodeFilename(descfn), 'w', encoding='utf-8') as descfile:
430 descfile.write(info_dict['description'])
431 except (OSError, IOError):
432 self.trouble(u'ERROR: Cannot write description file ' + descfn)
435 if self.params.get('writesubtitles', False) and 'subtitles' in info_dict and info_dict['subtitles']:
436 # subtitles download errors are already managed as troubles in relevant IE
437 # that way it will silently go on when used with unsupporting IE
439 srtfn = filename.rsplit('.', 1)[0] + u'.srt'
440 self.report_writesubtitles(srtfn)
441 with io.open(encodeFilename(srtfn), 'w', encoding='utf-8') as srtfile:
442 srtfile.write(info_dict['subtitles'])
443 except (OSError, IOError):
444 self.trouble(u'ERROR: Cannot write subtitles file ' + descfn)
447 if self.params.get('writeinfojson', False):
448 infofn = filename + u'.info.json'
449 self.report_writeinfojson(infofn)
451 json_info_dict = dict((k, v) for k,v in info_dict.items() if not k in ['urlhandle'])
452 write_json_file(json_info_dict, encodeFilename(infofn))
453 except (OSError, IOError):
454 self.trouble(u'ERROR: Cannot write metadata to JSON file ' + infofn)
457 if not self.params.get('skip_download', False):
458 if self.params.get('nooverwrites', False) and os.path.exists(encodeFilename(filename)):
462 success = self._do_download(filename, info_dict)
463 except (OSError, IOError) as err:
464 raise UnavailableVideoError()
465 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
466 self.trouble(u'ERROR: unable to download video data: %s' % str(err))
468 except (ContentTooShortError, ) as err:
469 self.trouble(u'ERROR: content too short (expected %s bytes and served %s)' % (err.expected, err.downloaded))
474 self.post_process(filename, info_dict)
475 except (PostProcessingError) as err:
476 self.trouble(u'ERROR: postprocessing: %s' % str(err))
479 def download(self, url_list):
480 """Download a given list of URLs."""
481 if len(url_list) > 1 and self.fixed_template():
482 raise SameFileError(self.params['outtmpl'])
485 suitable_found = False
487 # Go to next InfoExtractor if not suitable
488 if not ie.suitable(url):
491 # Warn if the _WORKING attribute is False
493 self.to_stderr(u'WARNING: the program functionality for this site has been marked as broken, '
494 u'and will probably not work. If you want to go on, use the -i option.')
496 # Suitable InfoExtractor found
497 suitable_found = True
499 # Extract information from URL and process it
501 videos = ie.extract(url)
502 except ExtractorError as de: # An error we somewhat expected
503 self.trouble(u'ERROR: ' + compat_str(de), de.format_traceback())
505 except Exception as e:
506 if self.params.get('ignoreerrors', False):
507 self.trouble(u'ERROR: ' + compat_str(e), tb=compat_str(traceback.format_exc()))
512 if len(videos or []) > 1 and self.fixed_template():
513 raise SameFileError(self.params['outtmpl'])
515 for video in videos or []:
516 video['extractor'] = ie.IE_NAME
518 self.increment_downloads()
519 self.process_info(video)
520 except UnavailableVideoError:
521 self.trouble(u'\nERROR: unable to download video')
523 # Suitable InfoExtractor had been found; go to next URL
526 if not suitable_found:
527 self.trouble(u'ERROR: no suitable InfoExtractor: %s' % url)
529 return self._download_retcode
531 def post_process(self, filename, ie_info):
532 """Run the postprocessing chain on the given file."""
534 info['filepath'] = filename
540 def _download_with_rtmpdump(self, filename, url, player_url, page_url):
541 self.report_destination(filename)
542 tmpfilename = self.temp_name(filename)
544 # Check for rtmpdump first
546 subprocess.call(['rtmpdump', '-h'], stdout=(file(os.path.devnull, 'w')), stderr=subprocess.STDOUT)
547 except (OSError, IOError):
548 self.trouble(u'ERROR: RTMP download detected but "rtmpdump" could not be run')
551 # Download using rtmpdump. rtmpdump returns exit code 2 when
552 # the connection was interrumpted and resuming appears to be
553 # possible. This is part of rtmpdump's normal usage, AFAIK.
554 basic_args = ['rtmpdump', '-q', '-r', url, '-o', tmpfilename]
555 if player_url is not None:
556 basic_args += ['-W', player_url]
557 if page_url is not None:
558 basic_args += ['--pageUrl', page_url]
559 args = basic_args + [[], ['-e', '-k', '1']][self.params.get('continuedl', False)]
560 if self.params.get('verbose', False):
563 shell_quote = lambda args: ' '.join(map(pipes.quote, args))
566 self.to_screen(u'[debug] rtmpdump command line: ' + shell_quote(args))
567 retval = subprocess.call(args)
568 while retval == 2 or retval == 1:
569 prevsize = os.path.getsize(encodeFilename(tmpfilename))
570 self.to_screen(u'\r[rtmpdump] %s bytes' % prevsize, skip_eol=True)
571 time.sleep(5.0) # This seems to be needed
572 retval = subprocess.call(basic_args + ['-e'] + [[], ['-k', '1']][retval == 1])
573 cursize = os.path.getsize(encodeFilename(tmpfilename))
574 if prevsize == cursize and retval == 1:
576 # Some rtmp streams seem abort after ~ 99.8%. Don't complain for those
577 if prevsize == cursize and retval == 2 and cursize > 1024:
578 self.to_screen(u'\r[rtmpdump] Could not download the whole video. This can happen for some advertisements.')
582 self.to_screen(u'\r[rtmpdump] %s bytes' % os.path.getsize(encodeFilename(tmpfilename)))
583 self.try_rename(tmpfilename, filename)
586 self.trouble(u'\nERROR: rtmpdump exited with code %d' % retval)
589 def _do_download(self, filename, info_dict):
590 url = info_dict['url']
592 # Check file already present
593 if self.params.get('continuedl', False) and os.path.isfile(encodeFilename(filename)) and not self.params.get('nopart', False):
594 self.report_file_already_downloaded(filename)
597 # Attempt to download using rtmpdump
598 if url.startswith('rtmp'):
599 return self._download_with_rtmpdump(filename, url,
600 info_dict.get('player_url', None),
601 info_dict.get('page_url', None))
603 tmpfilename = self.temp_name(filename)
606 # Do not include the Accept-Encoding header
607 headers = {'Youtubedl-no-compression': 'True'}
608 basic_request = compat_urllib_request.Request(url, None, headers)
609 request = compat_urllib_request.Request(url, None, headers)
611 if self.params.get('test', False):
612 request.add_header('Range','bytes=0-10240')
614 # Establish possible resume length
615 if os.path.isfile(encodeFilename(tmpfilename)):
616 resume_len = os.path.getsize(encodeFilename(tmpfilename))
622 if self.params.get('continuedl', False):
623 self.report_resuming_byte(resume_len)
624 request.add_header('Range','bytes=%d-' % resume_len)
630 retries = self.params.get('retries', 0)
631 while count <= retries:
632 # Establish connection
634 if count == 0 and 'urlhandle' in info_dict:
635 data = info_dict['urlhandle']
636 data = compat_urllib_request.urlopen(request)
638 except (compat_urllib_error.HTTPError, ) as err:
639 if (err.code < 500 or err.code >= 600) and err.code != 416:
640 # Unexpected HTTP error
642 elif err.code == 416:
643 # Unable to resume (requested range not satisfiable)
645 # Open the connection again without the range header
646 data = compat_urllib_request.urlopen(basic_request)
647 content_length = data.info()['Content-Length']
648 except (compat_urllib_error.HTTPError, ) as err:
649 if err.code < 500 or err.code >= 600:
652 # Examine the reported length
653 if (content_length is not None and
654 (resume_len - 100 < int(content_length) < resume_len + 100)):
655 # The file had already been fully downloaded.
656 # Explanation to the above condition: in issue #175 it was revealed that
657 # YouTube sometimes adds or removes a few bytes from the end of the file,
658 # changing the file size slightly and causing problems for some users. So
659 # I decided to implement a suggested change and consider the file
660 # completely downloaded if the file size differs less than 100 bytes from
661 # the one in the hard drive.
662 self.report_file_already_downloaded(filename)
663 self.try_rename(tmpfilename, filename)
666 # The length does not match, we start the download over
667 self.report_unable_to_resume()
673 self.report_retry(count, retries)
676 self.trouble(u'ERROR: giving up after %s retries' % retries)
679 data_len = data.info().get('Content-length', None)
680 if data_len is not None:
681 data_len = int(data_len) + resume_len
682 data_len_str = self.format_bytes(data_len)
683 byte_counter = 0 + resume_len
684 block_size = self.params.get('buffersize', 1024)
689 data_block = data.read(block_size)
691 if len(data_block) == 0:
693 byte_counter += len(data_block)
695 # Open file just in time
698 (stream, tmpfilename) = sanitize_open(tmpfilename, open_mode)
699 assert stream is not None
700 filename = self.undo_temp_name(tmpfilename)
701 self.report_destination(filename)
702 except (OSError, IOError) as err:
703 self.trouble(u'ERROR: unable to open for writing: %s' % str(err))
706 stream.write(data_block)
707 except (IOError, OSError) as err:
708 self.trouble(u'\nERROR: unable to write data: %s' % str(err))
710 if not self.params.get('noresizebuffer', False):
711 block_size = self.best_block_size(after - before, len(data_block))
714 speed_str = self.calc_speed(start, time.time(), byte_counter - resume_len)
716 self.report_progress('Unknown %', data_len_str, speed_str, 'Unknown ETA')
718 percent_str = self.calc_percent(byte_counter, data_len)
719 eta_str = self.calc_eta(start, time.time(), data_len - resume_len, byte_counter - resume_len)
720 self.report_progress(percent_str, data_len_str, speed_str, eta_str)
723 self.slow_down(start, byte_counter - resume_len)
726 self.trouble(u'\nERROR: Did not get any data blocks')
730 if data_len is not None and byte_counter != data_len:
731 raise ContentTooShortError(byte_counter, int(data_len))
732 self.try_rename(tmpfilename, filename)
734 # Update file modification time
735 if self.params.get('updatetime', True):
736 info_dict['filetime'] = self.try_utime(filename, data.info().get('last-modified', None))