2 # -*- coding: utf-8 -*-
4 from __future__ import absolute_import
20 from .InfoExtractors import get_info_extractor
23 class FileDownloader(object):
24 """File Downloader class.
26 File downloader objects are the ones responsible of downloading the
27 actual video file and writing it to disk if the user has requested
28 it, among some other tasks. In most cases there should be one per
29 program. As, given a video URL, the downloader doesn't know how to
30 extract all the needed information, task that InfoExtractors do, it
31 has to pass the URL to one of them.
33 For this, file downloader objects have a method that allows
34 InfoExtractors to be registered in a given order. When it is passed
35 a URL, the file downloader handles it to the first InfoExtractor it
36 finds that reports being able to handle it. The InfoExtractor extracts
37 all the information about the video or videos the URL refers to, and
38 asks the FileDownloader to process the video information, possibly
39 downloading the video.
41 File downloaders accept a lot of parameters. In order not to saturate
42 the object constructor with arguments, it receives a dictionary of
43 options instead. These options are available through the params
44 attribute for the InfoExtractors to use. The FileDownloader also
45 registers itself as the downloader in charge for the InfoExtractors
46 that are added to it, so this is a "mutual registration".
50 username: Username for authentication purposes.
51 password: Password for authentication purposes.
52 usenetrc: Use netrc for authentication instead.
53 quiet: Do not print messages to stdout.
54 forceurl: Force printing final URL.
55 forcetitle: Force printing title.
56 forcethumbnail: Force printing thumbnail URL.
57 forcedescription: Force printing description.
58 forcefilename: Force printing final filename.
59 simulate: Do not download the video files.
60 format: Video format code.
61 format_limit: Highest quality format to try.
62 outtmpl: Template for output names.
63 restrictfilenames: Do not allow "&" and spaces in file names
64 ignoreerrors: Do not stop on download errors.
65 ratelimit: Download speed limit, in bytes/sec.
66 nooverwrites: Prevent overwriting files.
67 retries: Number of times to retry for HTTP error 5xx
68 buffersize: Size of download buffer in bytes.
69 noresizebuffer: Do not automatically resize the download buffer.
70 continuedl: Try to continue downloads if possible.
71 noprogress: Do not print the progress bar.
72 playliststart: Playlist item to start at.
73 playlistend: Playlist item to end at.
74 matchtitle: Download only matching titles.
75 rejecttitle: Reject downloads for matching titles.
76 logtostderr: Log messages to stderr instead of stdout.
77 consoletitle: Display progress in console window's titlebar.
78 nopart: Do not use temporary .part files.
79 updatetime: Use the Last-modified header to set output file timestamps.
80 writedescription: Write the video description to a .description file
81 writeinfojson: Write the video description to a .info.json file
82 writesubtitles: Write the video subtitles to a file
83 onlysubtitles: Downloads only the subtitles of the video
84 allsubtitles: Downloads all the subtitles of the video
85 listsubtitles: Lists all available subtitles for the video
86 subtitlesformat: Subtitle format [sbv/srt] (default=srt)
87 subtitleslang: Language of the subtitles to download
88 test: Download only first bytes to test the downloader.
89 keepvideo: Keep the video file after post-processing
90 min_filesize: Skip files smaller than this size
91 max_filesize: Skip files larger than this size
92 daterange: A DateRange object, download only if the upload_date is in the range.
98 _download_retcode = None
102 def __init__(self, params):
103 """Create a FileDownloader object with the given options."""
106 self._progress_hooks = []
107 self._download_retcode = 0
108 self._num_downloads = 0
109 self._screen_file = [sys.stdout, sys.stderr][params.get('logtostderr', False)]
112 if '%(stitle)s' in self.params['outtmpl']:
113 self.report_warning(u'%(stitle)s is deprecated. Use the %(title)s and the --restrict-filenames flag(which also secures %(uploader)s et al) instead.')
116 def format_bytes(bytes):
119 if type(bytes) is str:
124 exponent = int(math.log(bytes, 1024.0))
125 suffix = ['B','KiB','MiB','GiB','TiB','PiB','EiB','ZiB','YiB'][exponent]
126 converted = float(bytes) / float(1024 ** exponent)
127 return '%.2f%s' % (converted, suffix)
130 def calc_percent(byte_counter, data_len):
133 return '%6s' % ('%3.1f%%' % (float(byte_counter) / float(data_len) * 100.0))
136 def calc_eta(start, now, total, current):
140 if current == 0 or dif < 0.001: # One millisecond
142 rate = float(current) / dif
143 eta = int((float(total) - float(current)) / rate)
144 (eta_mins, eta_secs) = divmod(eta, 60)
147 return '%02d:%02d' % (eta_mins, eta_secs)
150 def calc_speed(start, now, bytes):
152 if bytes == 0 or dif < 0.001: # One millisecond
153 return '%10s' % '---b/s'
154 return '%10s' % ('%s/s' % FileDownloader.format_bytes(float(bytes) / dif))
157 def best_block_size(elapsed_time, bytes):
158 new_min = max(bytes / 2.0, 1.0)
159 new_max = min(max(bytes * 2.0, 1.0), 4194304) # Do not surpass 4 MB
160 if elapsed_time < 0.001:
162 rate = bytes / elapsed_time
170 def parse_bytes(bytestr):
171 """Parse a string indicating a byte quantity into an integer."""
172 matchobj = re.match(r'(?i)^(\d+(?:\.\d+)?)([kMGTPEZY]?)$', bytestr)
175 number = float(matchobj.group(1))
176 multiplier = 1024.0 ** 'bkmgtpezy'.index(matchobj.group(2).lower())
177 return int(round(number * multiplier))
179 def add_info_extractor(self, ie):
180 """Add an InfoExtractor object to the end of the list."""
182 ie.set_downloader(self)
184 def add_post_processor(self, pp):
185 """Add a PostProcessor object to the end of the chain."""
187 pp.set_downloader(self)
189 def to_screen(self, message, skip_eol=False):
190 """Print message to stdout if not in quiet mode."""
191 assert type(message) == type(u'')
192 if not self.params.get('quiet', False):
193 terminator = [u'\n', u''][skip_eol]
194 output = message + terminator
195 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
196 output = output.encode(preferredencoding(), 'ignore')
197 self._screen_file.write(output)
198 self._screen_file.flush()
200 def to_stderr(self, message):
201 """Print message to stderr."""
202 assert type(message) == type(u'')
203 output = message + u'\n'
204 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
205 output = output.encode(preferredencoding())
206 sys.stderr.write(output)
208 def to_cons_title(self, message):
209 """Set console/terminal window title to message."""
210 if not self.params.get('consoletitle', False):
212 if os.name == 'nt' and ctypes.windll.kernel32.GetConsoleWindow():
213 # c_wchar_p() might not be necessary if `message` is
214 # already of type unicode()
215 ctypes.windll.kernel32.SetConsoleTitleW(ctypes.c_wchar_p(message))
216 elif 'TERM' in os.environ:
217 self.to_screen('\033]0;%s\007' % message, skip_eol=True)
219 def fixed_template(self):
220 """Checks if the output template is fixed."""
221 return (re.search(u'(?u)%\\(.+?\\)s', self.params['outtmpl']) is None)
223 def trouble(self, message=None, tb=None):
224 """Determine action to take when a download problem appears.
226 Depending on if the downloader has been configured to ignore
227 download errors or not, this method may throw an exception or
228 not when errors are found, after printing the message.
230 tb, if given, is additional traceback information.
232 if message is not None:
233 self.to_stderr(message)
234 if self.params.get('verbose'):
236 if sys.exc_info()[0]: # if .trouble has been called from an except block
238 if hasattr(sys.exc_info()[1], 'exc_info') and sys.exc_info()[1].exc_info[0]:
239 tb += u''.join(traceback.format_exception(*sys.exc_info()[1].exc_info))
240 tb += compat_str(traceback.format_exc())
242 tb_data = traceback.format_list(traceback.extract_stack())
243 tb = u''.join(tb_data)
245 if not self.params.get('ignoreerrors', False):
246 if sys.exc_info()[0] and hasattr(sys.exc_info()[1], 'exc_info') and sys.exc_info()[1].exc_info[0]:
247 exc_info = sys.exc_info()[1].exc_info
249 exc_info = sys.exc_info()
250 raise DownloadError(message, exc_info)
251 self._download_retcode = 1
253 def report_warning(self, message):
255 Print the message to stderr, it will be prefixed with 'WARNING:'
256 If stderr is a tty file the 'WARNING:' will be colored
258 if sys.stderr.isatty() and os.name != 'nt':
259 _msg_header=u'\033[0;33mWARNING:\033[0m'
261 _msg_header=u'WARNING:'
262 warning_message=u'%s %s' % (_msg_header,message)
263 self.to_stderr(warning_message)
265 def report_error(self, message, tb=None):
267 Do the same as trouble, but prefixes the message with 'ERROR:', colored
268 in red if stderr is a tty file.
270 if sys.stderr.isatty() and os.name != 'nt':
271 _msg_header = u'\033[0;31mERROR:\033[0m'
273 _msg_header = u'ERROR:'
274 error_message = u'%s %s' % (_msg_header, message)
275 self.trouble(error_message, tb)
277 def slow_down(self, start_time, byte_counter):
278 """Sleep if the download speed is over the rate limit."""
279 rate_limit = self.params.get('ratelimit', None)
280 if rate_limit is None or byte_counter == 0:
283 elapsed = now - start_time
286 speed = float(byte_counter) / elapsed
287 if speed > rate_limit:
288 time.sleep((byte_counter - rate_limit * (now - start_time)) / rate_limit)
290 def temp_name(self, filename):
291 """Returns a temporary filename for the given filename."""
292 if self.params.get('nopart', False) or filename == u'-' or \
293 (os.path.exists(encodeFilename(filename)) and not os.path.isfile(encodeFilename(filename))):
295 return filename + u'.part'
297 def undo_temp_name(self, filename):
298 if filename.endswith(u'.part'):
299 return filename[:-len(u'.part')]
302 def try_rename(self, old_filename, new_filename):
304 if old_filename == new_filename:
306 os.rename(encodeFilename(old_filename), encodeFilename(new_filename))
307 except (IOError, OSError) as err:
308 self.report_error(u'unable to rename file')
310 def try_utime(self, filename, last_modified_hdr):
311 """Try to set the last-modified time of the given file."""
312 if last_modified_hdr is None:
314 if not os.path.isfile(encodeFilename(filename)):
316 timestr = last_modified_hdr
319 filetime = timeconvert(timestr)
323 os.utime(filename, (time.time(), filetime))
328 def report_writedescription(self, descfn):
329 """ Report that the description file is being written """
330 self.to_screen(u'[info] Writing video description to: ' + descfn)
332 def report_writesubtitles(self, sub_filename):
333 """ Report that the subtitles file is being written """
334 self.to_screen(u'[info] Writing video subtitles to: ' + sub_filename)
336 def report_writeinfojson(self, infofn):
337 """ Report that the metadata file has been written """
338 self.to_screen(u'[info] Video description metadata as JSON to: ' + infofn)
340 def report_destination(self, filename):
341 """Report destination filename."""
342 self.to_screen(u'[download] Destination: ' + filename)
344 def report_progress(self, percent_str, data_len_str, speed_str, eta_str):
345 """Report download progress."""
346 if self.params.get('noprogress', False):
348 clear_line = (u'\x1b[K' if sys.stderr.isatty() and os.name != 'nt' else u'')
349 if self.params.get('progress_with_newline', False):
350 self.to_screen(u'[download] %s of %s at %s ETA %s' %
351 (percent_str, data_len_str, speed_str, eta_str))
353 self.to_screen(u'\r%s[download] %s of %s at %s ETA %s' %
354 (clear_line, percent_str, data_len_str, speed_str, eta_str), skip_eol=True)
355 self.to_cons_title(u'youtube-dl - %s of %s at %s ETA %s' %
356 (percent_str.strip(), data_len_str.strip(), speed_str.strip(), eta_str.strip()))
358 def report_resuming_byte(self, resume_len):
359 """Report attempt to resume at given byte."""
360 self.to_screen(u'[download] Resuming download at byte %s' % resume_len)
362 def report_retry(self, count, retries):
363 """Report retry in case of HTTP error 5xx"""
364 self.to_screen(u'[download] Got server HTTP error. Retrying (attempt %d of %d)...' % (count, retries))
366 def report_file_already_downloaded(self, file_name):
367 """Report file has already been fully downloaded."""
369 self.to_screen(u'[download] %s has already been downloaded' % file_name)
370 except (UnicodeEncodeError) as err:
371 self.to_screen(u'[download] The file has already been downloaded')
373 def report_unable_to_resume(self):
374 """Report it was impossible to resume download."""
375 self.to_screen(u'[download] Unable to resume')
377 def report_finish(self):
378 """Report download finished."""
379 if self.params.get('noprogress', False):
380 self.to_screen(u'[download] Download completed')
384 def increment_downloads(self):
385 """Increment the ordinal that assigns a number to each file."""
386 self._num_downloads += 1
388 def prepare_filename(self, info_dict):
389 """Generate the output filename."""
391 template_dict = dict(info_dict)
393 template_dict['epoch'] = int(time.time())
394 autonumber_size = self.params.get('autonumber_size')
395 if autonumber_size is None:
397 autonumber_templ = u'%0' + str(autonumber_size) + u'd'
398 template_dict['autonumber'] = autonumber_templ % self._num_downloads
399 if template_dict['playlist_index'] is not None:
400 template_dict['playlist_index'] = u'%05d' % template_dict['playlist_index']
402 sanitize = lambda k,v: sanitize_filename(
403 u'NA' if v is None else compat_str(v),
404 restricted=self.params.get('restrictfilenames'),
406 template_dict = dict((k, sanitize(k, v)) for k,v in template_dict.items())
408 filename = self.params['outtmpl'] % template_dict
410 except KeyError as err:
411 self.report_error(u'Erroneous output template')
413 except ValueError as err:
414 self.report_error(u'Insufficient system charset ' + repr(preferredencoding()))
417 def _match_entry(self, info_dict):
418 """ Returns None iff the file should be downloaded """
420 title = info_dict['title']
421 matchtitle = self.params.get('matchtitle', False)
423 if not re.search(matchtitle, title, re.IGNORECASE):
424 return u'[download] "' + title + '" title did not match pattern "' + matchtitle + '"'
425 rejecttitle = self.params.get('rejecttitle', False)
427 if re.search(rejecttitle, title, re.IGNORECASE):
428 return u'"' + title + '" title matched reject pattern "' + rejecttitle + '"'
429 date = info_dict.get('upload_date', None)
431 dateRange = self.params.get('daterange', DateRange())
432 if date not in dateRange:
433 return u'[download] %s upload date is not in range %s' % (date_from_str(date).isoformat(), dateRange)
436 def extract_info(self, url, download = True, ie_name = None):
438 Returns a list with a dictionary for each video we find.
439 If 'download', also downloads the videos.
441 suitable_found = False
443 #We copy the original list
444 ies = list(self._ies)
446 if ie_name is not None:
447 #We put in the first place the given info extractor
448 first_ie = get_info_extractor(ie_name)()
449 first_ie.set_downloader(self)
450 ies.insert(0, first_ie)
453 # Go to next InfoExtractor if not suitable
454 if not ie.suitable(url):
457 # Warn if the _WORKING attribute is False
459 self.report_warning(u'the program functionality for this site has been marked as broken, '
460 u'and will probably not work. If you want to go on, use the -i option.')
462 # Suitable InfoExtractor found
463 suitable_found = True
465 # Extract information from URL and process it
467 ie_results = ie.extract(url)
468 if ie_results is None: # Finished already (backwards compatibility; listformats and friends should be moved here)
471 for ie_result in ie_results:
472 if not 'extractor' in ie_result:
473 #The extractor has already been set somewhere else
474 ie_result['extractor'] = ie.IE_NAME
475 results.append(self.process_ie_result(ie_result, download))
477 except ExtractorError as de: # An error we somewhat expected
478 self.report_error(compat_str(de), de.format_traceback())
480 except Exception as e:
481 if self.params.get('ignoreerrors', False):
482 self.report_error(compat_str(e), tb=compat_str(traceback.format_exc()))
486 if not suitable_found:
487 self.report_error(u'no suitable InfoExtractor: %s' % url)
489 def process_ie_result(self, ie_result, download = True):
491 Take the result of the ie and return a list of videos.
492 For url elements it will search the suitable ie and get the videos
493 For playlist elements it will process each of the elements of the 'entries' key
495 It will also download the videos if 'download'.
497 result_type = ie_result.get('_type', 'video') #If not given we suppose it's a video, support the dafault old system
498 if result_type == 'video':
499 if 'playlist' not in ie_result:
500 #It isn't part of a playlist
501 ie_result['playlist'] = None
502 ie_result['playlist_index'] = None
505 self.process_info(ie_result)
507 elif result_type == 'url':
508 #We get the video pointed by the url
509 result = self.extract_info(ie_result['url'], download, ie_name = ie_result['ie_key'])[0]
511 elif result_type == 'playlist':
512 #We process each entry in the playlist
513 playlist = ie_result.get('title', None) or ie_result.get('id', None)
514 self.to_screen(u'[download] Downloading playlist: %s' % playlist)
516 playlist_results = []
518 n_all_entries = len(ie_result['entries'])
519 playliststart = self.params.get('playliststart', 1) - 1
520 playlistend = self.params.get('playlistend', -1)
522 if playlistend == -1:
523 entries = ie_result['entries'][playliststart:]
525 entries = ie_result['entries'][playliststart:playlistend]
527 n_entries = len(entries)
529 self.to_screen(u"[%s] playlist '%s': Collected %d video ids (downloading %d of them)" %
530 (ie_result['extractor'], playlist, n_all_entries, n_entries))
532 for i,entry in enumerate(entries,1):
533 self.to_screen(u'[download] Downloading video #%s of %s' %(i, n_entries))
534 entry_result = self.process_ie_result(entry, False)
535 entry_result['playlist'] = playlist
536 entry_result['playlist_index'] = i + playliststart
537 #We must do the download here to correctly set the 'playlist' key
539 self.process_info(entry_result)
540 playlist_results.append(entry_result)
541 result = ie_result.copy()
542 result['entries'] = playlist_results
545 def process_info(self, info_dict):
546 """Process a single dictionary returned by an InfoExtractor."""
548 #We increment the download the download count here to match the previous behaviour.
549 self.increment_downloads()
551 info_dict['fulltitle'] = info_dict['title']
552 if len(info_dict['title']) > 200:
553 info_dict['title'] = info_dict['title'][:197] + u'...'
555 # Keep for backwards compatibility
556 info_dict['stitle'] = info_dict['title']
558 if not 'format' in info_dict:
559 info_dict['format'] = info_dict['ext']
561 reason = self._match_entry(info_dict)
562 if reason is not None:
563 self.to_screen(u'[download] ' + reason)
566 max_downloads = self.params.get('max_downloads')
567 if max_downloads is not None:
568 if self._num_downloads > int(max_downloads):
569 raise MaxDownloadsReached()
571 filename = self.prepare_filename(info_dict)
574 if self.params.get('forcetitle', False):
575 compat_print(info_dict['title'])
576 if self.params.get('forceurl', False):
577 compat_print(info_dict['url'])
578 if self.params.get('forcethumbnail', False) and 'thumbnail' in info_dict:
579 compat_print(info_dict['thumbnail'])
580 if self.params.get('forcedescription', False) and 'description' in info_dict:
581 compat_print(info_dict['description'])
582 if self.params.get('forcefilename', False) and filename is not None:
583 compat_print(filename)
584 if self.params.get('forceformat', False):
585 compat_print(info_dict['format'])
587 # Do nothing else if in simulate mode
588 if self.params.get('simulate', False):
595 dn = os.path.dirname(encodeFilename(filename))
596 if dn != '' and not os.path.exists(dn): # dn is already encoded
598 except (OSError, IOError) as err:
599 self.report_error(u'unable to create directory ' + compat_str(err))
602 if self.params.get('writedescription', False):
604 descfn = filename + u'.description'
605 self.report_writedescription(descfn)
606 with io.open(encodeFilename(descfn), 'w', encoding='utf-8') as descfile:
607 descfile.write(info_dict['description'])
608 except (OSError, IOError):
609 self.report_error(u'Cannot write description file ' + descfn)
612 if self.params.get('writesubtitles', False) and 'subtitles' in info_dict and info_dict['subtitles']:
613 # subtitles download errors are already managed as troubles in relevant IE
614 # that way it will silently go on when used with unsupporting IE
615 subtitle = info_dict['subtitles'][0]
616 (sub_error, sub_lang, sub) = subtitle
617 sub_format = self.params.get('subtitlesformat')
619 self.report_warning("Some error while getting the subtitles")
622 sub_filename = filename.rsplit('.', 1)[0] + u'.' + sub_lang + u'.' + sub_format
623 self.report_writesubtitles(sub_filename)
624 with io.open(encodeFilename(sub_filename), 'w', encoding='utf-8') as subfile:
626 except (OSError, IOError):
627 self.report_error(u'Cannot write subtitles file ' + descfn)
629 if self.params.get('onlysubtitles', False):
632 if self.params.get('allsubtitles', False) and 'subtitles' in info_dict and info_dict['subtitles']:
633 subtitles = info_dict['subtitles']
634 sub_format = self.params.get('subtitlesformat')
635 for subtitle in subtitles:
636 (sub_error, sub_lang, sub) = subtitle
638 self.report_warning("Some error while getting the subtitles")
641 sub_filename = filename.rsplit('.', 1)[0] + u'.' + sub_lang + u'.' + sub_format
642 self.report_writesubtitles(sub_filename)
643 with io.open(encodeFilename(sub_filename), 'w', encoding='utf-8') as subfile:
645 except (OSError, IOError):
646 self.report_error(u'Cannot write subtitles file ' + descfn)
648 if self.params.get('onlysubtitles', False):
651 if self.params.get('writeinfojson', False):
652 infofn = filename + u'.info.json'
653 self.report_writeinfojson(infofn)
655 json_info_dict = dict((k, v) for k,v in info_dict.items() if not k in ['urlhandle'])
656 write_json_file(json_info_dict, encodeFilename(infofn))
657 except (OSError, IOError):
658 self.report_error(u'Cannot write metadata to JSON file ' + infofn)
661 if not self.params.get('skip_download', False):
662 if self.params.get('nooverwrites', False) and os.path.exists(encodeFilename(filename)):
666 success = self._do_download(filename, info_dict)
667 except (OSError, IOError) as err:
668 raise UnavailableVideoError()
669 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
670 self.report_error(u'unable to download video data: %s' % str(err))
672 except (ContentTooShortError, ) as err:
673 self.report_error(u'content too short (expected %s bytes and served %s)' % (err.expected, err.downloaded))
678 self.post_process(filename, info_dict)
679 except (PostProcessingError) as err:
680 self.report_error(u'postprocessing: %s' % str(err))
683 def download(self, url_list):
684 """Download a given list of URLs."""
685 if len(url_list) > 1 and self.fixed_template():
686 raise SameFileError(self.params['outtmpl'])
690 #It also downloads the videos
691 videos = self.extract_info(url)
692 except UnavailableVideoError:
693 self.report_error(u'unable to download video')
694 except MaxDownloadsReached:
695 self.to_screen(u'[info] Maximum number of downloaded files reached.')
698 return self._download_retcode
700 def post_process(self, filename, ie_info):
701 """Run all the postprocessors on the given file."""
703 info['filepath'] = filename
707 keep_video_wish,new_info = pp.run(info)
708 if keep_video_wish is not None:
710 keep_video = keep_video_wish
711 elif keep_video is None:
712 # No clear decision yet, let IE decide
713 keep_video = keep_video_wish
714 except PostProcessingError as e:
715 self.to_stderr(u'ERROR: ' + e.msg)
716 if keep_video is False and not self.params.get('keepvideo', False):
718 self.to_screen(u'Deleting original file %s (pass -k to keep)' % filename)
719 os.remove(encodeFilename(filename))
720 except (IOError, OSError):
721 self.report_warning(u'Unable to remove downloaded video file')
723 def _download_with_rtmpdump(self, filename, url, player_url, page_url, play_path):
724 self.report_destination(filename)
725 tmpfilename = self.temp_name(filename)
727 # Check for rtmpdump first
729 subprocess.call(['rtmpdump', '-h'], stdout=(open(os.path.devnull, 'w')), stderr=subprocess.STDOUT)
730 except (OSError, IOError):
731 self.report_error(u'RTMP download detected but "rtmpdump" could not be run')
734 # Download using rtmpdump. rtmpdump returns exit code 2 when
735 # the connection was interrumpted and resuming appears to be
736 # possible. This is part of rtmpdump's normal usage, AFAIK.
737 basic_args = ['rtmpdump', '-q', '-r', url, '-o', tmpfilename]
738 if player_url is not None:
739 basic_args += ['-W', player_url]
740 if page_url is not None:
741 basic_args += ['--pageUrl', page_url]
742 if play_path is not None:
743 basic_args += ['-y', play_path]
744 args = basic_args + [[], ['-e', '-k', '1']][self.params.get('continuedl', False)]
745 if self.params.get('verbose', False):
748 shell_quote = lambda args: ' '.join(map(pipes.quote, args))
751 self.to_screen(u'[debug] rtmpdump command line: ' + shell_quote(args))
752 retval = subprocess.call(args)
753 while retval == 2 or retval == 1:
754 prevsize = os.path.getsize(encodeFilename(tmpfilename))
755 self.to_screen(u'\r[rtmpdump] %s bytes' % prevsize, skip_eol=True)
756 time.sleep(5.0) # This seems to be needed
757 retval = subprocess.call(basic_args + ['-e'] + [[], ['-k', '1']][retval == 1])
758 cursize = os.path.getsize(encodeFilename(tmpfilename))
759 if prevsize == cursize and retval == 1:
761 # Some rtmp streams seem abort after ~ 99.8%. Don't complain for those
762 if prevsize == cursize and retval == 2 and cursize > 1024:
763 self.to_screen(u'\r[rtmpdump] Could not download the whole video. This can happen for some advertisements.')
767 fsize = os.path.getsize(encodeFilename(tmpfilename))
768 self.to_screen(u'\r[rtmpdump] %s bytes' % fsize)
769 self.try_rename(tmpfilename, filename)
770 self._hook_progress({
771 'downloaded_bytes': fsize,
772 'total_bytes': fsize,
773 'filename': filename,
774 'status': 'finished',
778 self.to_stderr(u"\n")
779 self.report_error(u'rtmpdump exited with code %d' % retval)
782 def _do_download(self, filename, info_dict):
783 url = info_dict['url']
785 # Check file already present
786 if self.params.get('continuedl', False) and os.path.isfile(encodeFilename(filename)) and not self.params.get('nopart', False):
787 self.report_file_already_downloaded(filename)
788 self._hook_progress({
789 'filename': filename,
790 'status': 'finished',
794 # Attempt to download using rtmpdump
795 if url.startswith('rtmp'):
796 return self._download_with_rtmpdump(filename, url,
797 info_dict.get('player_url', None),
798 info_dict.get('page_url', None),
799 info_dict.get('play_path', None))
801 tmpfilename = self.temp_name(filename)
804 # Do not include the Accept-Encoding header
805 headers = {'Youtubedl-no-compression': 'True'}
806 if 'user_agent' in info_dict:
807 headers['Youtubedl-user-agent'] = info_dict['user_agent']
808 basic_request = compat_urllib_request.Request(url, None, headers)
809 request = compat_urllib_request.Request(url, None, headers)
811 if self.params.get('test', False):
812 request.add_header('Range','bytes=0-10240')
814 # Establish possible resume length
815 if os.path.isfile(encodeFilename(tmpfilename)):
816 resume_len = os.path.getsize(encodeFilename(tmpfilename))
822 if self.params.get('continuedl', False):
823 self.report_resuming_byte(resume_len)
824 request.add_header('Range','bytes=%d-' % resume_len)
830 retries = self.params.get('retries', 0)
831 while count <= retries:
832 # Establish connection
834 if count == 0 and 'urlhandle' in info_dict:
835 data = info_dict['urlhandle']
836 data = compat_urllib_request.urlopen(request)
838 except (compat_urllib_error.HTTPError, ) as err:
839 if (err.code < 500 or err.code >= 600) and err.code != 416:
840 # Unexpected HTTP error
842 elif err.code == 416:
843 # Unable to resume (requested range not satisfiable)
845 # Open the connection again without the range header
846 data = compat_urllib_request.urlopen(basic_request)
847 content_length = data.info()['Content-Length']
848 except (compat_urllib_error.HTTPError, ) as err:
849 if err.code < 500 or err.code >= 600:
852 # Examine the reported length
853 if (content_length is not None and
854 (resume_len - 100 < int(content_length) < resume_len + 100)):
855 # The file had already been fully downloaded.
856 # Explanation to the above condition: in issue #175 it was revealed that
857 # YouTube sometimes adds or removes a few bytes from the end of the file,
858 # changing the file size slightly and causing problems for some users. So
859 # I decided to implement a suggested change and consider the file
860 # completely downloaded if the file size differs less than 100 bytes from
861 # the one in the hard drive.
862 self.report_file_already_downloaded(filename)
863 self.try_rename(tmpfilename, filename)
864 self._hook_progress({
865 'filename': filename,
866 'status': 'finished',
870 # The length does not match, we start the download over
871 self.report_unable_to_resume()
877 self.report_retry(count, retries)
880 self.report_error(u'giving up after %s retries' % retries)
883 data_len = data.info().get('Content-length', None)
884 if data_len is not None:
885 data_len = int(data_len) + resume_len
886 min_data_len = self.params.get("min_filesize", None)
887 max_data_len = self.params.get("max_filesize", None)
888 if min_data_len is not None and data_len < min_data_len:
889 self.to_screen(u'\r[download] File is smaller than min-filesize (%s bytes < %s bytes). Aborting.' % (data_len, min_data_len))
891 if max_data_len is not None and data_len > max_data_len:
892 self.to_screen(u'\r[download] File is larger than max-filesize (%s bytes > %s bytes). Aborting.' % (data_len, max_data_len))
895 data_len_str = self.format_bytes(data_len)
896 byte_counter = 0 + resume_len
897 block_size = self.params.get('buffersize', 1024)
902 data_block = data.read(block_size)
904 if len(data_block) == 0:
906 byte_counter += len(data_block)
908 # Open file just in time
911 (stream, tmpfilename) = sanitize_open(tmpfilename, open_mode)
912 assert stream is not None
913 filename = self.undo_temp_name(tmpfilename)
914 self.report_destination(filename)
915 except (OSError, IOError) as err:
916 self.report_error(u'unable to open for writing: %s' % str(err))
919 stream.write(data_block)
920 except (IOError, OSError) as err:
921 self.to_stderr(u"\n")
922 self.report_error(u'unable to write data: %s' % str(err))
924 if not self.params.get('noresizebuffer', False):
925 block_size = self.best_block_size(after - before, len(data_block))
928 speed_str = self.calc_speed(start, time.time(), byte_counter - resume_len)
930 self.report_progress('Unknown %', data_len_str, speed_str, 'Unknown ETA')
932 percent_str = self.calc_percent(byte_counter, data_len)
933 eta_str = self.calc_eta(start, time.time(), data_len - resume_len, byte_counter - resume_len)
934 self.report_progress(percent_str, data_len_str, speed_str, eta_str)
936 self._hook_progress({
937 'downloaded_bytes': byte_counter,
938 'total_bytes': data_len,
939 'tmpfilename': tmpfilename,
940 'filename': filename,
941 'status': 'downloading',
945 self.slow_down(start, byte_counter - resume_len)
948 self.to_stderr(u"\n")
949 self.report_error(u'Did not get any data blocks')
953 if data_len is not None and byte_counter != data_len:
954 raise ContentTooShortError(byte_counter, int(data_len))
955 self.try_rename(tmpfilename, filename)
957 # Update file modification time
958 if self.params.get('updatetime', True):
959 info_dict['filetime'] = self.try_utime(filename, data.info().get('last-modified', None))
961 self._hook_progress({
962 'downloaded_bytes': byte_counter,
963 'total_bytes': byte_counter,
964 'filename': filename,
965 'status': 'finished',
970 def _hook_progress(self, status):
971 for ph in self._progress_hooks:
974 def add_progress_hook(self, ph):
975 """ ph gets called on download progress, with a dictionary with the entries
976 * filename: The final filename
977 * status: One of "downloading" and "finished"
979 It can also have some of the following entries:
981 * downloaded_bytes: Bytes on disks
982 * total_bytes: Total bytes, None if unknown
983 * tmpfilename: The filename we're currently writing to
985 Hooks are guaranteed to be called at least once (with status "finished")
986 if the download is successful.
988 self._progress_hooks.append(ph)