2 # -*- coding: utf-8 -*-
4 from __future__ import absolute_import
21 from .InfoExtractors import get_info_extractor
24 class FileDownloader(object):
25 """File Downloader class.
27 File downloader objects are the ones responsible of downloading the
28 actual video file and writing it to disk if the user has requested
29 it, among some other tasks. In most cases there should be one per
30 program. As, given a video URL, the downloader doesn't know how to
31 extract all the needed information, task that InfoExtractors do, it
32 has to pass the URL to one of them.
34 For this, file downloader objects have a method that allows
35 InfoExtractors to be registered in a given order. When it is passed
36 a URL, the file downloader handles it to the first InfoExtractor it
37 finds that reports being able to handle it. The InfoExtractor extracts
38 all the information about the video or videos the URL refers to, and
39 asks the FileDownloader to process the video information, possibly
40 downloading the video.
42 File downloaders accept a lot of parameters. In order not to saturate
43 the object constructor with arguments, it receives a dictionary of
44 options instead. These options are available through the params
45 attribute for the InfoExtractors to use. The FileDownloader also
46 registers itself as the downloader in charge for the InfoExtractors
47 that are added to it, so this is a "mutual registration".
51 username: Username for authentication purposes.
52 password: Password for authentication purposes.
53 usenetrc: Use netrc for authentication instead.
54 quiet: Do not print messages to stdout.
55 forceurl: Force printing final URL.
56 forcetitle: Force printing title.
57 forcethumbnail: Force printing thumbnail URL.
58 forcedescription: Force printing description.
59 forcefilename: Force printing final filename.
60 simulate: Do not download the video files.
61 format: Video format code.
62 format_limit: Highest quality format to try.
63 outtmpl: Template for output names.
64 restrictfilenames: Do not allow "&" and spaces in file names
65 ignoreerrors: Do not stop on download errors.
66 ratelimit: Download speed limit, in bytes/sec.
67 nooverwrites: Prevent overwriting files.
68 retries: Number of times to retry for HTTP error 5xx
69 buffersize: Size of download buffer in bytes.
70 noresizebuffer: Do not automatically resize the download buffer.
71 continuedl: Try to continue downloads if possible.
72 noprogress: Do not print the progress bar.
73 playliststart: Playlist item to start at.
74 playlistend: Playlist item to end at.
75 matchtitle: Download only matching titles.
76 rejecttitle: Reject downloads for matching titles.
77 logtostderr: Log messages to stderr instead of stdout.
78 consoletitle: Display progress in console window's titlebar.
79 nopart: Do not use temporary .part files.
80 updatetime: Use the Last-modified header to set output file timestamps.
81 writedescription: Write the video description to a .description file
82 writeinfojson: Write the video description to a .info.json file
83 writethumbnail: Write the thumbnail image to a file
84 writesubtitles: Write the video subtitles to a file
85 onlysubtitles: Downloads only the subtitles of the video
86 allsubtitles: Downloads all the subtitles of the video
87 listsubtitles: Lists all available subtitles for the video
88 subtitlesformat: Subtitle format [sbv/srt] (default=srt)
89 subtitleslang: Language of the subtitles to download
90 test: Download only first bytes to test the downloader.
91 keepvideo: Keep the video file after post-processing
92 min_filesize: Skip files smaller than this size
93 max_filesize: Skip files larger than this size
94 daterange: A DateRange object, download only if the upload_date is in the range.
100 _download_retcode = None
101 _num_downloads = None
104 def __init__(self, params):
105 """Create a FileDownloader object with the given options."""
108 self._progress_hooks = []
109 self._download_retcode = 0
110 self._num_downloads = 0
111 self._screen_file = [sys.stdout, sys.stderr][params.get('logtostderr', False)]
114 if '%(stitle)s' in self.params['outtmpl']:
115 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.')
118 def format_bytes(bytes):
121 if type(bytes) is str:
126 exponent = int(math.log(bytes, 1024.0))
127 suffix = ['B','KiB','MiB','GiB','TiB','PiB','EiB','ZiB','YiB'][exponent]
128 converted = float(bytes) / float(1024 ** exponent)
129 return '%.2f%s' % (converted, suffix)
132 def calc_percent(byte_counter, data_len):
135 return '%6s' % ('%3.1f%%' % (float(byte_counter) / float(data_len) * 100.0))
138 def calc_eta(start, now, total, current):
142 if current == 0 or dif < 0.001: # One millisecond
144 rate = float(current) / dif
145 eta = int((float(total) - float(current)) / rate)
146 (eta_mins, eta_secs) = divmod(eta, 60)
149 return '%02d:%02d' % (eta_mins, eta_secs)
152 def calc_speed(start, now, bytes):
154 if bytes == 0 or dif < 0.001: # One millisecond
155 return '%10s' % '---b/s'
156 return '%10s' % ('%s/s' % FileDownloader.format_bytes(float(bytes) / dif))
159 def best_block_size(elapsed_time, bytes):
160 new_min = max(bytes / 2.0, 1.0)
161 new_max = min(max(bytes * 2.0, 1.0), 4194304) # Do not surpass 4 MB
162 if elapsed_time < 0.001:
164 rate = bytes / elapsed_time
172 def parse_bytes(bytestr):
173 """Parse a string indicating a byte quantity into an integer."""
174 matchobj = re.match(r'(?i)^(\d+(?:\.\d+)?)([kMGTPEZY]?)$', bytestr)
177 number = float(matchobj.group(1))
178 multiplier = 1024.0 ** 'bkmgtpezy'.index(matchobj.group(2).lower())
179 return int(round(number * multiplier))
181 def add_info_extractor(self, ie):
182 """Add an InfoExtractor object to the end of the list."""
184 ie.set_downloader(self)
186 def add_post_processor(self, pp):
187 """Add a PostProcessor object to the end of the chain."""
189 pp.set_downloader(self)
191 def to_screen(self, message, skip_eol=False):
192 """Print message to stdout if not in quiet mode."""
193 assert type(message) == type(u'')
194 if not self.params.get('quiet', False):
195 terminator = [u'\n', u''][skip_eol]
196 output = message + terminator
197 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
198 output = output.encode(preferredencoding(), 'ignore')
199 self._screen_file.write(output)
200 self._screen_file.flush()
202 def to_stderr(self, message):
203 """Print message to stderr."""
204 assert type(message) == type(u'')
205 output = message + u'\n'
206 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
207 output = output.encode(preferredencoding())
208 sys.stderr.write(output)
210 def to_cons_title(self, message):
211 """Set console/terminal window title to message."""
212 if not self.params.get('consoletitle', False):
214 if os.name == 'nt' and ctypes.windll.kernel32.GetConsoleWindow():
215 # c_wchar_p() might not be necessary if `message` is
216 # already of type unicode()
217 ctypes.windll.kernel32.SetConsoleTitleW(ctypes.c_wchar_p(message))
218 elif 'TERM' in os.environ:
219 self.to_screen('\033]0;%s\007' % message, skip_eol=True)
221 def fixed_template(self):
222 """Checks if the output template is fixed."""
223 return (re.search(u'(?u)%\\(.+?\\)s', self.params['outtmpl']) is None)
225 def trouble(self, message=None, tb=None):
226 """Determine action to take when a download problem appears.
228 Depending on if the downloader has been configured to ignore
229 download errors or not, this method may throw an exception or
230 not when errors are found, after printing the message.
232 tb, if given, is additional traceback information.
234 if message is not None:
235 self.to_stderr(message)
236 if self.params.get('verbose'):
238 if sys.exc_info()[0]: # if .trouble has been called from an except block
240 if hasattr(sys.exc_info()[1], 'exc_info') and sys.exc_info()[1].exc_info[0]:
241 tb += u''.join(traceback.format_exception(*sys.exc_info()[1].exc_info))
242 tb += compat_str(traceback.format_exc())
244 tb_data = traceback.format_list(traceback.extract_stack())
245 tb = u''.join(tb_data)
247 if not self.params.get('ignoreerrors', False):
248 if sys.exc_info()[0] and hasattr(sys.exc_info()[1], 'exc_info') and sys.exc_info()[1].exc_info[0]:
249 exc_info = sys.exc_info()[1].exc_info
251 exc_info = sys.exc_info()
252 raise DownloadError(message, exc_info)
253 self._download_retcode = 1
255 def report_warning(self, message):
257 Print the message to stderr, it will be prefixed with 'WARNING:'
258 If stderr is a tty file the 'WARNING:' will be colored
260 if sys.stderr.isatty() and os.name != 'nt':
261 _msg_header=u'\033[0;33mWARNING:\033[0m'
263 _msg_header=u'WARNING:'
264 warning_message=u'%s %s' % (_msg_header,message)
265 self.to_stderr(warning_message)
267 def report_error(self, message, tb=None):
269 Do the same as trouble, but prefixes the message with 'ERROR:', colored
270 in red if stderr is a tty file.
272 if sys.stderr.isatty() and os.name != 'nt':
273 _msg_header = u'\033[0;31mERROR:\033[0m'
275 _msg_header = u'ERROR:'
276 error_message = u'%s %s' % (_msg_header, message)
277 self.trouble(error_message, tb)
279 def slow_down(self, start_time, byte_counter):
280 """Sleep if the download speed is over the rate limit."""
281 rate_limit = self.params.get('ratelimit', None)
282 if rate_limit is None or byte_counter == 0:
285 elapsed = now - start_time
288 speed = float(byte_counter) / elapsed
289 if speed > rate_limit:
290 time.sleep((byte_counter - rate_limit * (now - start_time)) / rate_limit)
292 def temp_name(self, filename):
293 """Returns a temporary filename for the given filename."""
294 if self.params.get('nopart', False) or filename == u'-' or \
295 (os.path.exists(encodeFilename(filename)) and not os.path.isfile(encodeFilename(filename))):
297 return filename + u'.part'
299 def undo_temp_name(self, filename):
300 if filename.endswith(u'.part'):
301 return filename[:-len(u'.part')]
304 def try_rename(self, old_filename, new_filename):
306 if old_filename == new_filename:
308 os.rename(encodeFilename(old_filename), encodeFilename(new_filename))
309 except (IOError, OSError) as err:
310 self.report_error(u'unable to rename file')
312 def try_utime(self, filename, last_modified_hdr):
313 """Try to set the last-modified time of the given file."""
314 if last_modified_hdr is None:
316 if not os.path.isfile(encodeFilename(filename)):
318 timestr = last_modified_hdr
321 filetime = timeconvert(timestr)
325 os.utime(filename, (time.time(), filetime))
330 def report_writedescription(self, descfn):
331 """ Report that the description file is being written """
332 self.to_screen(u'[info] Writing video description to: ' + descfn)
334 def report_writesubtitles(self, sub_filename):
335 """ Report that the subtitles file is being written """
336 self.to_screen(u'[info] Writing video subtitles to: ' + sub_filename)
338 def report_writeinfojson(self, infofn):
339 """ Report that the metadata file has been written """
340 self.to_screen(u'[info] Video description metadata as JSON to: ' + infofn)
342 def report_destination(self, filename):
343 """Report destination filename."""
344 self.to_screen(u'[download] Destination: ' + filename)
346 def report_progress(self, percent_str, data_len_str, speed_str, eta_str):
347 """Report download progress."""
348 if self.params.get('noprogress', False):
350 clear_line = (u'\x1b[K' if sys.stderr.isatty() and os.name != 'nt' else u'')
351 if self.params.get('progress_with_newline', False):
352 self.to_screen(u'[download] %s of %s at %s ETA %s' %
353 (percent_str, data_len_str, speed_str, eta_str))
355 self.to_screen(u'\r%s[download] %s of %s at %s ETA %s' %
356 (clear_line, percent_str, data_len_str, speed_str, eta_str), skip_eol=True)
357 self.to_cons_title(u'youtube-dl - %s of %s at %s ETA %s' %
358 (percent_str.strip(), data_len_str.strip(), speed_str.strip(), eta_str.strip()))
360 def report_resuming_byte(self, resume_len):
361 """Report attempt to resume at given byte."""
362 self.to_screen(u'[download] Resuming download at byte %s' % resume_len)
364 def report_retry(self, count, retries):
365 """Report retry in case of HTTP error 5xx"""
366 self.to_screen(u'[download] Got server HTTP error. Retrying (attempt %d of %d)...' % (count, retries))
368 def report_file_already_downloaded(self, file_name):
369 """Report file has already been fully downloaded."""
371 self.to_screen(u'[download] %s has already been downloaded' % file_name)
372 except (UnicodeEncodeError) as err:
373 self.to_screen(u'[download] The file has already been downloaded')
375 def report_unable_to_resume(self):
376 """Report it was impossible to resume download."""
377 self.to_screen(u'[download] Unable to resume')
379 def report_finish(self):
380 """Report download finished."""
381 if self.params.get('noprogress', False):
382 self.to_screen(u'[download] Download completed')
386 def increment_downloads(self):
387 """Increment the ordinal that assigns a number to each file."""
388 self._num_downloads += 1
390 def prepare_filename(self, info_dict):
391 """Generate the output filename."""
393 template_dict = dict(info_dict)
395 template_dict['epoch'] = int(time.time())
396 autonumber_size = self.params.get('autonumber_size')
397 if autonumber_size is None:
399 autonumber_templ = u'%0' + str(autonumber_size) + u'd'
400 template_dict['autonumber'] = autonumber_templ % self._num_downloads
401 if template_dict['playlist_index'] is not None:
402 template_dict['playlist_index'] = u'%05d' % template_dict['playlist_index']
404 sanitize = lambda k,v: sanitize_filename(
405 u'NA' if v is None else compat_str(v),
406 restricted=self.params.get('restrictfilenames'),
408 template_dict = dict((k, sanitize(k, v)) for k,v in template_dict.items())
410 filename = self.params['outtmpl'] % template_dict
412 except KeyError as err:
413 self.report_error(u'Erroneous output template')
415 except ValueError as err:
416 self.report_error(u'Insufficient system charset ' + repr(preferredencoding()))
419 def _match_entry(self, info_dict):
420 """ Returns None iff the file should be downloaded """
422 title = info_dict['title']
423 matchtitle = self.params.get('matchtitle', False)
425 if not re.search(matchtitle, title, re.IGNORECASE):
426 return u'[download] "' + title + '" title did not match pattern "' + matchtitle + '"'
427 rejecttitle = self.params.get('rejecttitle', False)
429 if re.search(rejecttitle, title, re.IGNORECASE):
430 return u'"' + title + '" title matched reject pattern "' + rejecttitle + '"'
431 date = info_dict.get('upload_date', None)
433 dateRange = self.params.get('daterange', DateRange())
434 if date not in dateRange:
435 return u'[download] %s upload date is not in range %s' % (date_from_str(date).isoformat(), dateRange)
438 def extract_info(self, url, download=True, ie_key=None):
440 Returns a list with a dictionary for each video we find.
441 If 'download', also downloads the videos.
445 ie = get_info_extractor(ie_key)()
446 ie.set_downloader(self)
452 if not ie.suitable(url):
456 self.report_warning(u'The program functionality for this site has been marked as broken, '
457 u'and will probably not work.')
460 ie_result = ie.extract(url)
461 if ie_result is None: # Finished already (backwards compatibility; listformats and friends should be moved here)
463 if isinstance(ie_result, list):
464 # Backwards compatibility: old IE result format
466 '_type': 'compat_list',
467 'entries': ie_result,
469 if 'extractor' not in ie_result:
470 ie_result['extractor'] = ie.IE_NAME
471 return self.process_ie_result(ie_result, download=download)
472 except ExtractorError as de: # An error we somewhat expected
473 self.report_error(compat_str(de), de.format_traceback())
475 except Exception as e:
476 if self.params.get('ignoreerrors', False):
477 self.report_error(compat_str(e), tb=compat_str(traceback.format_exc()))
482 self.report_error(u'no suitable InfoExtractor: %s' % url)
484 def process_ie_result(self, ie_result, download=True):
486 Take the result of the ie(may be modified) and resolve all unresolved
487 references (URLs, playlist items).
489 It will also download the videos if 'download'.
490 Returns the resolved ie_result.
493 result_type = ie_result.get('_type', 'video') # If not given we suppose it's a video, support the default old system
494 if result_type == 'video':
495 if 'playlist' not in ie_result:
496 # It isn't part of a playlist
497 ie_result['playlist'] = None
498 ie_result['playlist_index'] = None
500 self.process_info(ie_result)
502 elif result_type == 'url':
503 return self.extract_info(ie_result['url'], download, ie_key=ie_result.get('ie_key'))
504 elif result_type == 'playlist':
505 # We process each entry in the playlist
506 playlist = ie_result.get('title', None) or ie_result.get('id', None)
507 self.to_screen(u'[download] Downloading playlist: %s' % playlist)
509 playlist_results = []
511 n_all_entries = len(ie_result['entries'])
512 playliststart = self.params.get('playliststart', 1) - 1
513 playlistend = self.params.get('playlistend', -1)
515 if playlistend == -1:
516 entries = ie_result['entries'][playliststart:]
518 entries = ie_result['entries'][playliststart:playlistend]
520 n_entries = len(entries)
522 self.to_screen(u"[%s] playlist '%s': Collected %d video ids (downloading %d of them)" %
523 (ie_result['extractor'], playlist, n_all_entries, n_entries))
525 for i,entry in enumerate(entries,1):
526 self.to_screen(u'[download] Downloading video #%s of %s' %(i, n_entries))
527 entry['playlist'] = playlist
528 entry['playlist_index'] = i + playliststart
529 entry_result = self.process_ie_result(entry, download=download)
530 playlist_results.append(entry_result)
531 ie_result['entries'] = playlist_results
533 elif result_type == 'compat_list':
535 r.setdefault('extractor', ie_result['extractor'])
537 ie_result['entries'] = [
538 self.process_ie_result(_fixup(r), download=download)
539 for r in ie_result['entries']
543 raise Exception('Invalid result type: %s' % result_type)
545 def process_info(self, info_dict):
546 """Process a single resolved IE result."""
548 assert info_dict.get('_type', 'video') == 'video'
549 #We increment the download the download count here to match the previous behaviour.
550 self.increment_downloads()
552 info_dict['fulltitle'] = info_dict['title']
553 if len(info_dict['title']) > 200:
554 info_dict['title'] = info_dict['title'][:197] + u'...'
556 # Keep for backwards compatibility
557 info_dict['stitle'] = info_dict['title']
559 if not 'format' in info_dict:
560 info_dict['format'] = info_dict['ext']
562 reason = self._match_entry(info_dict)
563 if reason is not None:
564 self.to_screen(u'[download] ' + reason)
567 max_downloads = self.params.get('max_downloads')
568 if max_downloads is not None:
569 if self._num_downloads > int(max_downloads):
570 raise MaxDownloadsReached()
572 filename = self.prepare_filename(info_dict)
575 if self.params.get('forcetitle', False):
576 compat_print(info_dict['title'])
577 if self.params.get('forceurl', False):
578 compat_print(info_dict['url'])
579 if self.params.get('forcethumbnail', False) and 'thumbnail' in info_dict:
580 compat_print(info_dict['thumbnail'])
581 if self.params.get('forcedescription', False) and 'description' in info_dict:
582 compat_print(info_dict['description'])
583 if self.params.get('forcefilename', False) and filename is not None:
584 compat_print(filename)
585 if self.params.get('forceformat', False):
586 compat_print(info_dict['format'])
588 # Do nothing else if in simulate mode
589 if self.params.get('simulate', False):
596 dn = os.path.dirname(encodeFilename(filename))
597 if dn != '' and not os.path.exists(dn): # dn is already encoded
599 except (OSError, IOError) as err:
600 self.report_error(u'unable to create directory ' + compat_str(err))
603 if self.params.get('writedescription', False):
605 descfn = filename + u'.description'
606 self.report_writedescription(descfn)
607 with io.open(encodeFilename(descfn), 'w', encoding='utf-8') as descfile:
608 descfile.write(info_dict['description'])
609 except (OSError, IOError):
610 self.report_error(u'Cannot write description file ' + descfn)
613 if self.params.get('writesubtitles', False) and 'subtitles' in info_dict and info_dict['subtitles']:
614 # subtitles download errors are already managed as troubles in relevant IE
615 # that way it will silently go on when used with unsupporting IE
616 subtitle = info_dict['subtitles'][0]
617 (sub_error, sub_lang, sub) = subtitle
618 sub_format = self.params.get('subtitlesformat')
620 self.report_warning("Some error while getting the subtitles")
623 sub_filename = filename.rsplit('.', 1)[0] + u'.' + sub_lang + u'.' + sub_format
624 self.report_writesubtitles(sub_filename)
625 with io.open(encodeFilename(sub_filename), 'w', encoding='utf-8') as subfile:
627 except (OSError, IOError):
628 self.report_error(u'Cannot write subtitles file ' + descfn)
630 if self.params.get('onlysubtitles', False):
633 if self.params.get('allsubtitles', False) and 'subtitles' in info_dict and info_dict['subtitles']:
634 subtitles = info_dict['subtitles']
635 sub_format = self.params.get('subtitlesformat')
636 for subtitle in subtitles:
637 (sub_error, sub_lang, sub) = subtitle
639 self.report_warning("Some error while getting the subtitles")
642 sub_filename = filename.rsplit('.', 1)[0] + u'.' + sub_lang + u'.' + sub_format
643 self.report_writesubtitles(sub_filename)
644 with io.open(encodeFilename(sub_filename), 'w', encoding='utf-8') as subfile:
646 except (OSError, IOError):
647 self.report_error(u'Cannot write subtitles file ' + descfn)
649 if self.params.get('onlysubtitles', False):
652 if self.params.get('writeinfojson', False):
653 infofn = filename + u'.info.json'
654 self.report_writeinfojson(infofn)
656 json_info_dict = dict((k, v) for k,v in info_dict.items() if not k in ['urlhandle'])
657 write_json_file(json_info_dict, encodeFilename(infofn))
658 except (OSError, IOError):
659 self.report_error(u'Cannot write metadata to JSON file ' + infofn)
662 if self.params.get('writethumbnail', False):
663 if 'thumbnail' in info_dict:
664 thumb_format = info_dict['thumbnail'].rpartition(u'/')[2].rpartition(u'.')[2]
667 thumb_filename = filename.rpartition('.')[0] + u'.' + thumb_format
668 self.to_screen(u'[%s] %s: Downloading thumbnail ...' %
669 (info_dict['extractor'], info_dict['id']))
670 uf = compat_urllib_request.urlopen(info_dict['thumbnail'])
671 with open(thumb_filename, 'wb') as thumbf:
672 shutil.copyfileobj(uf, thumbf)
673 self.to_screen(u'[%s] %s: Writing thumbnail to: %s' %
674 (info_dict['extractor'], info_dict['id'], thumb_filename))
676 if not self.params.get('skip_download', False):
677 if self.params.get('nooverwrites', False) and os.path.exists(encodeFilename(filename)):
681 success = self._do_download(filename, info_dict)
682 except (OSError, IOError) as err:
683 raise UnavailableVideoError()
684 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
685 self.report_error(u'unable to download video data: %s' % str(err))
687 except (ContentTooShortError, ) as err:
688 self.report_error(u'content too short (expected %s bytes and served %s)' % (err.expected, err.downloaded))
693 self.post_process(filename, info_dict)
694 except (PostProcessingError) as err:
695 self.report_error(u'postprocessing: %s' % str(err))
698 def download(self, url_list):
699 """Download a given list of URLs."""
700 if len(url_list) > 1 and self.fixed_template():
701 raise SameFileError(self.params['outtmpl'])
705 #It also downloads the videos
706 videos = self.extract_info(url)
707 except UnavailableVideoError:
708 self.report_error(u'unable to download video')
709 except MaxDownloadsReached:
710 self.to_screen(u'[info] Maximum number of downloaded files reached.')
713 return self._download_retcode
715 def post_process(self, filename, ie_info):
716 """Run all the postprocessors on the given file."""
718 info['filepath'] = filename
722 keep_video_wish,new_info = pp.run(info)
723 if keep_video_wish is not None:
725 keep_video = keep_video_wish
726 elif keep_video is None:
727 # No clear decision yet, let IE decide
728 keep_video = keep_video_wish
729 except PostProcessingError as e:
730 self.to_stderr(u'ERROR: ' + e.msg)
731 if keep_video is False and not self.params.get('keepvideo', False):
733 self.to_screen(u'Deleting original file %s (pass -k to keep)' % filename)
734 os.remove(encodeFilename(filename))
735 except (IOError, OSError):
736 self.report_warning(u'Unable to remove downloaded video file')
738 def _download_with_rtmpdump(self, filename, url, player_url, page_url, play_path):
739 self.report_destination(filename)
740 tmpfilename = self.temp_name(filename)
742 # Check for rtmpdump first
744 subprocess.call(['rtmpdump', '-h'], stdout=(open(os.path.devnull, 'w')), stderr=subprocess.STDOUT)
745 except (OSError, IOError):
746 self.report_error(u'RTMP download detected but "rtmpdump" could not be run')
749 # Download using rtmpdump. rtmpdump returns exit code 2 when
750 # the connection was interrumpted and resuming appears to be
751 # possible. This is part of rtmpdump's normal usage, AFAIK.
752 basic_args = ['rtmpdump', '-q', '-r', url, '-o', tmpfilename]
753 if player_url is not None:
754 basic_args += ['-W', player_url]
755 if page_url is not None:
756 basic_args += ['--pageUrl', page_url]
757 if play_path is not None:
758 basic_args += ['-y', play_path]
759 args = basic_args + [[], ['-e', '-k', '1']][self.params.get('continuedl', False)]
760 if self.params.get('verbose', False):
763 shell_quote = lambda args: ' '.join(map(pipes.quote, args))
766 self.to_screen(u'[debug] rtmpdump command line: ' + shell_quote(args))
767 retval = subprocess.call(args)
768 while retval == 2 or retval == 1:
769 prevsize = os.path.getsize(encodeFilename(tmpfilename))
770 self.to_screen(u'\r[rtmpdump] %s bytes' % prevsize, skip_eol=True)
771 time.sleep(5.0) # This seems to be needed
772 retval = subprocess.call(basic_args + ['-e'] + [[], ['-k', '1']][retval == 1])
773 cursize = os.path.getsize(encodeFilename(tmpfilename))
774 if prevsize == cursize and retval == 1:
776 # Some rtmp streams seem abort after ~ 99.8%. Don't complain for those
777 if prevsize == cursize and retval == 2 and cursize > 1024:
778 self.to_screen(u'\r[rtmpdump] Could not download the whole video. This can happen for some advertisements.')
782 fsize = os.path.getsize(encodeFilename(tmpfilename))
783 self.to_screen(u'\r[rtmpdump] %s bytes' % fsize)
784 self.try_rename(tmpfilename, filename)
785 self._hook_progress({
786 'downloaded_bytes': fsize,
787 'total_bytes': fsize,
788 'filename': filename,
789 'status': 'finished',
793 self.to_stderr(u"\n")
794 self.report_error(u'rtmpdump exited with code %d' % retval)
797 def _do_download(self, filename, info_dict):
798 url = info_dict['url']
800 # Check file already present
801 if self.params.get('continuedl', False) and os.path.isfile(encodeFilename(filename)) and not self.params.get('nopart', False):
802 self.report_file_already_downloaded(filename)
803 self._hook_progress({
804 'filename': filename,
805 'status': 'finished',
809 # Attempt to download using rtmpdump
810 if url.startswith('rtmp'):
811 return self._download_with_rtmpdump(filename, url,
812 info_dict.get('player_url', None),
813 info_dict.get('page_url', None),
814 info_dict.get('play_path', None))
816 tmpfilename = self.temp_name(filename)
819 # Do not include the Accept-Encoding header
820 headers = {'Youtubedl-no-compression': 'True'}
821 if 'user_agent' in info_dict:
822 headers['Youtubedl-user-agent'] = info_dict['user_agent']
823 basic_request = compat_urllib_request.Request(url, None, headers)
824 request = compat_urllib_request.Request(url, None, headers)
826 if self.params.get('test', False):
827 request.add_header('Range','bytes=0-10240')
829 # Establish possible resume length
830 if os.path.isfile(encodeFilename(tmpfilename)):
831 resume_len = os.path.getsize(encodeFilename(tmpfilename))
837 if self.params.get('continuedl', False):
838 self.report_resuming_byte(resume_len)
839 request.add_header('Range','bytes=%d-' % resume_len)
845 retries = self.params.get('retries', 0)
846 while count <= retries:
847 # Establish connection
849 if count == 0 and 'urlhandle' in info_dict:
850 data = info_dict['urlhandle']
851 data = compat_urllib_request.urlopen(request)
853 except (compat_urllib_error.HTTPError, ) as err:
854 if (err.code < 500 or err.code >= 600) and err.code != 416:
855 # Unexpected HTTP error
857 elif err.code == 416:
858 # Unable to resume (requested range not satisfiable)
860 # Open the connection again without the range header
861 data = compat_urllib_request.urlopen(basic_request)
862 content_length = data.info()['Content-Length']
863 except (compat_urllib_error.HTTPError, ) as err:
864 if err.code < 500 or err.code >= 600:
867 # Examine the reported length
868 if (content_length is not None and
869 (resume_len - 100 < int(content_length) < resume_len + 100)):
870 # The file had already been fully downloaded.
871 # Explanation to the above condition: in issue #175 it was revealed that
872 # YouTube sometimes adds or removes a few bytes from the end of the file,
873 # changing the file size slightly and causing problems for some users. So
874 # I decided to implement a suggested change and consider the file
875 # completely downloaded if the file size differs less than 100 bytes from
876 # the one in the hard drive.
877 self.report_file_already_downloaded(filename)
878 self.try_rename(tmpfilename, filename)
879 self._hook_progress({
880 'filename': filename,
881 'status': 'finished',
885 # The length does not match, we start the download over
886 self.report_unable_to_resume()
892 self.report_retry(count, retries)
895 self.report_error(u'giving up after %s retries' % retries)
898 data_len = data.info().get('Content-length', None)
899 if data_len is not None:
900 data_len = int(data_len) + resume_len
901 min_data_len = self.params.get("min_filesize", None)
902 max_data_len = self.params.get("max_filesize", None)
903 if min_data_len is not None and data_len < min_data_len:
904 self.to_screen(u'\r[download] File is smaller than min-filesize (%s bytes < %s bytes). Aborting.' % (data_len, min_data_len))
906 if max_data_len is not None and data_len > max_data_len:
907 self.to_screen(u'\r[download] File is larger than max-filesize (%s bytes > %s bytes). Aborting.' % (data_len, max_data_len))
910 data_len_str = self.format_bytes(data_len)
911 byte_counter = 0 + resume_len
912 block_size = self.params.get('buffersize', 1024)
917 data_block = data.read(block_size)
919 if len(data_block) == 0:
921 byte_counter += len(data_block)
923 # Open file just in time
926 (stream, tmpfilename) = sanitize_open(tmpfilename, open_mode)
927 assert stream is not None
928 filename = self.undo_temp_name(tmpfilename)
929 self.report_destination(filename)
930 except (OSError, IOError) as err:
931 self.report_error(u'unable to open for writing: %s' % str(err))
934 stream.write(data_block)
935 except (IOError, OSError) as err:
936 self.to_stderr(u"\n")
937 self.report_error(u'unable to write data: %s' % str(err))
939 if not self.params.get('noresizebuffer', False):
940 block_size = self.best_block_size(after - before, len(data_block))
943 speed_str = self.calc_speed(start, time.time(), byte_counter - resume_len)
945 self.report_progress('Unknown %', data_len_str, speed_str, 'Unknown ETA')
947 percent_str = self.calc_percent(byte_counter, data_len)
948 eta_str = self.calc_eta(start, time.time(), data_len - resume_len, byte_counter - resume_len)
949 self.report_progress(percent_str, data_len_str, speed_str, eta_str)
951 self._hook_progress({
952 'downloaded_bytes': byte_counter,
953 'total_bytes': data_len,
954 'tmpfilename': tmpfilename,
955 'filename': filename,
956 'status': 'downloading',
960 self.slow_down(start, byte_counter - resume_len)
963 self.to_stderr(u"\n")
964 self.report_error(u'Did not get any data blocks')
968 if data_len is not None and byte_counter != data_len:
969 raise ContentTooShortError(byte_counter, int(data_len))
970 self.try_rename(tmpfilename, filename)
972 # Update file modification time
973 if self.params.get('updatetime', True):
974 info_dict['filetime'] = self.try_utime(filename, data.info().get('last-modified', None))
976 self._hook_progress({
977 'downloaded_bytes': byte_counter,
978 'total_bytes': byte_counter,
979 'filename': filename,
980 'status': 'finished',
985 def _hook_progress(self, status):
986 for ph in self._progress_hooks:
989 def add_progress_hook(self, ph):
990 """ ph gets called on download progress, with a dictionary with the entries
991 * filename: The final filename
992 * status: One of "downloading" and "finished"
994 It can also have some of the following entries:
996 * downloaded_bytes: Bytes on disks
997 * total_bytes: Total bytes, None if unknown
998 * tmpfilename: The filename we're currently writing to
1000 Hooks are guaranteed to be called at least once (with status "finished")
1001 if the download is successful.
1003 self._progress_hooks.append(ph)