2 # -*- coding: utf-8 -*-
19 class FileDownloader(object):
20 """File Downloader class.
22 File downloader objects are the ones responsible of downloading the
23 actual video file and writing it to disk if the user has requested
24 it, among some other tasks. In most cases there should be one per
25 program. As, given a video URL, the downloader doesn't know how to
26 extract all the needed information, task that InfoExtractors do, it
27 has to pass the URL to one of them.
29 For this, file downloader objects have a method that allows
30 InfoExtractors to be registered in a given order. When it is passed
31 a URL, the file downloader handles it to the first InfoExtractor it
32 finds that reports being able to handle it. The InfoExtractor extracts
33 all the information about the video or videos the URL refers to, and
34 asks the FileDownloader to process the video information, possibly
35 downloading the video.
37 File downloaders accept a lot of parameters. In order not to saturate
38 the object constructor with arguments, it receives a dictionary of
39 options instead. These options are available through the params
40 attribute for the InfoExtractors to use. The FileDownloader also
41 registers itself as the downloader in charge for the InfoExtractors
42 that are added to it, so this is a "mutual registration".
46 username: Username for authentication purposes.
47 password: Password for authentication purposes.
48 usenetrc: Use netrc for authentication instead.
49 quiet: Do not print messages to stdout.
50 forceurl: Force printing final URL.
51 forcetitle: Force printing title.
52 forcethumbnail: Force printing thumbnail URL.
53 forcedescription: Force printing description.
54 forcefilename: Force printing final filename.
55 simulate: Do not download the video files.
56 format: Video format code.
57 format_limit: Highest quality format to try.
58 outtmpl: Template for output names.
59 restrictfilenames: Do not allow "&" and spaces in file names
60 ignoreerrors: Do not stop on download errors.
61 ratelimit: Download speed limit, in bytes/sec.
62 nooverwrites: Prevent overwriting files.
63 retries: Number of times to retry for HTTP error 5xx
64 buffersize: Size of download buffer in bytes.
65 noresizebuffer: Do not automatically resize the download buffer.
66 continuedl: Try to continue downloads if possible.
67 noprogress: Do not print the progress bar.
68 playliststart: Playlist item to start at.
69 playlistend: Playlist item to end at.
70 matchtitle: Download only matching titles.
71 rejecttitle: Reject downloads for matching titles.
72 logtostderr: Log messages to stderr instead of stdout.
73 consoletitle: Display progress in console window's titlebar.
74 nopart: Do not use temporary .part files.
75 updatetime: Use the Last-modified header to set output file timestamps.
76 writedescription: Write the video description to a .description file
77 writeinfojson: Write the video description to a .info.json file
78 writesubtitles: Write the video subtitles to a .srt file
79 subtitleslang: Language of the subtitles to download
85 _download_retcode = None
89 def __init__(self, params):
90 """Create a FileDownloader object with the given options."""
93 self._download_retcode = 0
94 self._num_downloads = 0
95 self._screen_file = [sys.stdout, sys.stderr][params.get('logtostderr', False)]
98 if '%(stitle)s' in self.params['outtmpl']:
99 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.')
102 def format_bytes(bytes):
105 if type(bytes) is str:
110 exponent = int(math.log(bytes, 1024.0))
111 suffix = 'bkMGTPEZY'[exponent]
112 converted = float(bytes) / float(1024 ** exponent)
113 return '%.2f%s' % (converted, suffix)
116 def calc_percent(byte_counter, data_len):
119 return '%6s' % ('%3.1f%%' % (float(byte_counter) / float(data_len) * 100.0))
122 def calc_eta(start, now, total, current):
126 if current == 0 or dif < 0.001: # One millisecond
128 rate = float(current) / dif
129 eta = int((float(total) - float(current)) / rate)
130 (eta_mins, eta_secs) = divmod(eta, 60)
133 return '%02d:%02d' % (eta_mins, eta_secs)
136 def calc_speed(start, now, bytes):
138 if bytes == 0 or dif < 0.001: # One millisecond
139 return '%10s' % '---b/s'
140 return '%10s' % ('%s/s' % FileDownloader.format_bytes(float(bytes) / dif))
143 def best_block_size(elapsed_time, bytes):
144 new_min = max(bytes / 2.0, 1.0)
145 new_max = min(max(bytes * 2.0, 1.0), 4194304) # Do not surpass 4 MB
146 if elapsed_time < 0.001:
148 rate = bytes / elapsed_time
156 def parse_bytes(bytestr):
157 """Parse a string indicating a byte quantity into an integer."""
158 matchobj = re.match(r'(?i)^(\d+(?:\.\d+)?)([kMGTPEZY]?)$', bytestr)
161 number = float(matchobj.group(1))
162 multiplier = 1024.0 ** 'bkmgtpezy'.index(matchobj.group(2).lower())
163 return int(round(number * multiplier))
165 def add_info_extractor(self, ie):
166 """Add an InfoExtractor object to the end of the list."""
168 ie.set_downloader(self)
170 def add_post_processor(self, pp):
171 """Add a PostProcessor object to the end of the chain."""
173 pp.set_downloader(self)
175 def to_screen(self, message, skip_eol=False):
176 """Print message to stdout if not in quiet mode."""
177 assert type(message) == type(u'')
178 if not self.params.get('quiet', False):
179 terminator = [u'\n', u''][skip_eol]
180 output = message + terminator
181 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
182 output = output.encode(preferredencoding(), 'ignore')
183 self._screen_file.write(output)
184 self._screen_file.flush()
186 def to_stderr(self, message):
187 """Print message to stderr."""
188 assert type(message) == type(u'')
189 sys.stderr.write((message + u'\n').encode(preferredencoding()))
191 def to_cons_title(self, message):
192 """Set console/terminal window title to message."""
193 if not self.params.get('consoletitle', False):
195 if os.name == 'nt' and ctypes.windll.kernel32.GetConsoleWindow():
196 # c_wchar_p() might not be necessary if `message` is
197 # already of type unicode()
198 ctypes.windll.kernel32.SetConsoleTitleW(ctypes.c_wchar_p(message))
199 elif 'TERM' in os.environ:
200 sys.stderr.write('\033]0;%s\007' % message.encode(preferredencoding()))
202 def fixed_template(self):
203 """Checks if the output template is fixed."""
204 return (re.search(ur'(?u)%\(.+?\)s', self.params['outtmpl']) is None)
206 def trouble(self, message=None):
207 """Determine action to take when a download problem appears.
209 Depending on if the downloader has been configured to ignore
210 download errors or not, this method may throw an exception or
211 not when errors are found, after printing the message.
213 if message is not None:
214 self.to_stderr(message)
215 if not self.params.get('ignoreerrors', False):
216 raise DownloadError(message)
217 self._download_retcode = 1
219 def slow_down(self, start_time, byte_counter):
220 """Sleep if the download speed is over the rate limit."""
221 rate_limit = self.params.get('ratelimit', None)
222 if rate_limit is None or byte_counter == 0:
225 elapsed = now - start_time
228 speed = float(byte_counter) / elapsed
229 if speed > rate_limit:
230 time.sleep((byte_counter - rate_limit * (now - start_time)) / rate_limit)
232 def temp_name(self, filename):
233 """Returns a temporary filename for the given filename."""
234 if self.params.get('nopart', False) or filename == u'-' or \
235 (os.path.exists(encodeFilename(filename)) and not os.path.isfile(encodeFilename(filename))):
237 return filename + u'.part'
239 def undo_temp_name(self, filename):
240 if filename.endswith(u'.part'):
241 return filename[:-len(u'.part')]
244 def try_rename(self, old_filename, new_filename):
246 if old_filename == new_filename:
248 os.rename(encodeFilename(old_filename), encodeFilename(new_filename))
249 except (IOError, OSError) as err:
250 self.trouble(u'ERROR: unable to rename file')
252 def try_utime(self, filename, last_modified_hdr):
253 """Try to set the last-modified time of the given file."""
254 if last_modified_hdr is None:
256 if not os.path.isfile(encodeFilename(filename)):
258 timestr = last_modified_hdr
261 filetime = timeconvert(timestr)
265 os.utime(filename, (time.time(), filetime))
270 def report_writedescription(self, descfn):
271 """ Report that the description file is being written """
272 self.to_screen(u'[info] Writing video description to: ' + descfn)
274 def report_writesubtitles(self, srtfn):
275 """ Report that the subtitles file is being written """
276 self.to_screen(u'[info] Writing video subtitles to: ' + srtfn)
278 def report_writeinfojson(self, infofn):
279 """ Report that the metadata file has been written """
280 self.to_screen(u'[info] Video description metadata as JSON to: ' + infofn)
282 def report_destination(self, filename):
283 """Report destination filename."""
284 self.to_screen(u'[download] Destination: ' + filename)
286 def report_progress(self, percent_str, data_len_str, speed_str, eta_str):
287 """Report download progress."""
288 if self.params.get('noprogress', False):
290 self.to_screen(u'\r[download] %s of %s at %s ETA %s' %
291 (percent_str, data_len_str, speed_str, eta_str), skip_eol=True)
292 self.to_cons_title(u'youtube-dl - %s of %s at %s ETA %s' %
293 (percent_str.strip(), data_len_str.strip(), speed_str.strip(), eta_str.strip()))
295 def report_resuming_byte(self, resume_len):
296 """Report attempt to resume at given byte."""
297 self.to_screen(u'[download] Resuming download at byte %s' % resume_len)
299 def report_retry(self, count, retries):
300 """Report retry in case of HTTP error 5xx"""
301 self.to_screen(u'[download] Got server HTTP error. Retrying (attempt %d of %d)...' % (count, retries))
303 def report_file_already_downloaded(self, file_name):
304 """Report file has already been fully downloaded."""
306 self.to_screen(u'[download] %s has already been downloaded' % file_name)
307 except (UnicodeEncodeError) as err:
308 self.to_screen(u'[download] The file has already been downloaded')
310 def report_unable_to_resume(self):
311 """Report it was impossible to resume download."""
312 self.to_screen(u'[download] Unable to resume')
314 def report_finish(self):
315 """Report download finished."""
316 if self.params.get('noprogress', False):
317 self.to_screen(u'[download] Download completed')
321 def increment_downloads(self):
322 """Increment the ordinal that assigns a number to each file."""
323 self._num_downloads += 1
325 def prepare_filename(self, info_dict):
326 """Generate the output filename."""
328 template_dict = dict(info_dict)
330 template_dict['epoch'] = int(time.time())
331 template_dict['autonumber'] = u'%05d' % self._num_downloads
333 template_dict = dict((key, u'NA' if val is None else val) for key, val in template_dict.items())
334 template_dict = dict((k, sanitize_filename(compat_str(v), self.params.get('restrictfilenames'))) for k,v in template_dict.items())
336 filename = self.params['outtmpl'] % template_dict
338 except (ValueError, KeyError) as err:
339 self.trouble(u'ERROR: invalid system charset or erroneous output template')
342 def _match_entry(self, info_dict):
343 """ Returns None iff the file should be downloaded """
345 title = info_dict['title']
346 matchtitle = self.params.get('matchtitle', False)
348 matchtitle = matchtitle.decode('utf8')
349 if not re.search(matchtitle, title, re.IGNORECASE):
350 return u'[download] "' + title + '" title did not match pattern "' + matchtitle + '"'
351 rejecttitle = self.params.get('rejecttitle', False)
353 rejecttitle = rejecttitle.decode('utf8')
354 if re.search(rejecttitle, title, re.IGNORECASE):
355 return u'"' + title + '" title matched reject pattern "' + rejecttitle + '"'
358 def process_info(self, info_dict):
359 """Process a single dictionary returned by an InfoExtractor."""
361 # Keep for backwards compatibility
362 info_dict['stitle'] = info_dict['title']
364 if not 'format' in info_dict:
365 info_dict['format'] = info_dict['ext']
367 reason = self._match_entry(info_dict)
368 if reason is not None:
369 self.to_screen(u'[download] ' + reason)
372 max_downloads = self.params.get('max_downloads')
373 if max_downloads is not None:
374 if self._num_downloads > int(max_downloads):
375 raise MaxDownloadsReached()
377 filename = self.prepare_filename(info_dict)
380 if self.params.get('forcetitle', False):
381 print(info_dict['title'].encode(preferredencoding(), 'xmlcharrefreplace'))
382 if self.params.get('forceurl', False):
383 print(info_dict['url'].encode(preferredencoding(), 'xmlcharrefreplace'))
384 if self.params.get('forcethumbnail', False) and 'thumbnail' in info_dict:
385 print(info_dict['thumbnail'].encode(preferredencoding(), 'xmlcharrefreplace'))
386 if self.params.get('forcedescription', False) and 'description' in info_dict:
387 print(info_dict['description'].encode(preferredencoding(), 'xmlcharrefreplace'))
388 if self.params.get('forcefilename', False) and filename is not None:
389 print(filename.encode(preferredencoding(), 'xmlcharrefreplace'))
390 if self.params.get('forceformat', False):
391 print(info_dict['format'].encode(preferredencoding(), 'xmlcharrefreplace'))
393 # Do nothing else if in simulate mode
394 if self.params.get('simulate', False):
401 dn = os.path.dirname(encodeFilename(filename))
402 if dn != '' and not os.path.exists(dn): # dn is already encoded
404 except (OSError, IOError) as err:
405 self.trouble(u'ERROR: unable to create directory ' + compat_str(err))
408 if self.params.get('writedescription', False):
410 descfn = filename + u'.description'
411 self.report_writedescription(descfn)
412 descfile = open(encodeFilename(descfn), 'wb')
414 descfile.write(info_dict['description'].encode('utf-8'))
417 except (OSError, IOError):
418 self.trouble(u'ERROR: Cannot write description file ' + descfn)
421 if self.params.get('writesubtitles', False) and 'subtitles' in info_dict and info_dict['subtitles']:
422 # subtitles download errors are already managed as troubles in relevant IE
423 # that way it will silently go on when used with unsupporting IE
425 srtfn = filename.rsplit('.', 1)[0] + u'.srt'
426 self.report_writesubtitles(srtfn)
427 srtfile = open(encodeFilename(srtfn), 'wb')
429 srtfile.write(info_dict['subtitles'].encode('utf-8'))
432 except (OSError, IOError):
433 self.trouble(u'ERROR: Cannot write subtitles file ' + descfn)
436 if self.params.get('writeinfojson', False):
437 infofn = filename + u'.info.json'
438 self.report_writeinfojson(infofn)
441 except (NameError,AttributeError):
442 self.trouble(u'ERROR: No JSON encoder found. Update to Python 2.6+, setup a json module, or leave out --write-info-json.')
445 infof = open(encodeFilename(infofn), 'wb')
447 json_info_dict = dict((k,v) for k,v in info_dict.iteritems() if not k in ('urlhandle',))
448 json.dump(json_info_dict, infof)
451 except (OSError, IOError):
452 self.trouble(u'ERROR: Cannot write metadata to JSON file ' + infofn)
455 if not self.params.get('skip_download', False):
456 if self.params.get('nooverwrites', False) and os.path.exists(encodeFilename(filename)):
460 success = self._do_download(filename, info_dict)
461 except (OSError, IOError) as err:
462 raise UnavailableVideoError
463 except (compat_urllib_error.URLError, httplib.HTTPException, socket.error) as err:
464 self.trouble(u'ERROR: unable to download video data: %s' % str(err))
466 except (ContentTooShortError, ) as err:
467 self.trouble(u'ERROR: content too short (expected %s bytes and served %s)' % (err.expected, err.downloaded))
472 self.post_process(filename, info_dict)
473 except (PostProcessingError) as err:
474 self.trouble(u'ERROR: postprocessing: %s' % str(err))
477 def download(self, url_list):
478 """Download a given list of URLs."""
479 if len(url_list) > 1 and self.fixed_template():
480 raise SameFileError(self.params['outtmpl'])
483 suitable_found = False
485 # Go to next InfoExtractor if not suitable
486 if not ie.suitable(url):
489 # Warn if the _WORKING attribute is False
491 self.trouble(u'WARNING: the program functionality for this site has been marked as broken, '
492 u'and will probably not work. If you want to go on, use the -i option.')
494 # Suitable InfoExtractor found
495 suitable_found = True
497 # Extract information from URL and process it
498 videos = ie.extract(url)
499 for video in videos or []:
500 video['extractor'] = ie.IE_NAME
502 self.increment_downloads()
503 self.process_info(video)
504 except UnavailableVideoError:
505 self.trouble(u'\nERROR: unable to download video')
507 # Suitable InfoExtractor had been found; go to next URL
510 if not suitable_found:
511 self.trouble(u'ERROR: no suitable InfoExtractor: %s' % url)
513 return self._download_retcode
515 def post_process(self, filename, ie_info):
516 """Run the postprocessing chain on the given file."""
518 info['filepath'] = filename
524 def _download_with_rtmpdump(self, filename, url, player_url):
525 self.report_destination(filename)
526 tmpfilename = self.temp_name(filename)
528 # Check for rtmpdump first
530 subprocess.call(['rtmpdump', '-h'], stdout=(file(os.path.devnull, 'w')), stderr=subprocess.STDOUT)
531 except (OSError, IOError):
532 self.trouble(u'ERROR: RTMP download detected but "rtmpdump" could not be run')
535 # Download using rtmpdump. rtmpdump returns exit code 2 when
536 # the connection was interrumpted and resuming appears to be
537 # possible. This is part of rtmpdump's normal usage, AFAIK.
538 basic_args = ['rtmpdump', '-q'] + [[], ['-W', player_url]][player_url is not None] + ['-r', url, '-o', tmpfilename]
539 args = basic_args + [[], ['-e', '-k', '1']][self.params.get('continuedl', False)]
540 if self.params.get('verbose', False):
543 shell_quote = lambda args: ' '.join(map(pipes.quote, args))
546 self.to_screen(u'[debug] rtmpdump command line: ' + shell_quote(args))
547 retval = subprocess.call(args)
548 while retval == 2 or retval == 1:
549 prevsize = os.path.getsize(encodeFilename(tmpfilename))
550 self.to_screen(u'\r[rtmpdump] %s bytes' % prevsize, skip_eol=True)
551 time.sleep(5.0) # This seems to be needed
552 retval = subprocess.call(basic_args + ['-e'] + [[], ['-k', '1']][retval == 1])
553 cursize = os.path.getsize(encodeFilename(tmpfilename))
554 if prevsize == cursize and retval == 1:
556 # Some rtmp streams seem abort after ~ 99.8%. Don't complain for those
557 if prevsize == cursize and retval == 2 and cursize > 1024:
558 self.to_screen(u'\r[rtmpdump] Could not download the whole video. This can happen for some advertisements.')
562 self.to_screen(u'\r[rtmpdump] %s bytes' % os.path.getsize(encodeFilename(tmpfilename)))
563 self.try_rename(tmpfilename, filename)
566 self.trouble(u'\nERROR: rtmpdump exited with code %d' % retval)
569 def _do_download(self, filename, info_dict):
570 url = info_dict['url']
571 player_url = info_dict.get('player_url', None)
573 # Check file already present
574 if self.params.get('continuedl', False) and os.path.isfile(encodeFilename(filename)) and not self.params.get('nopart', False):
575 self.report_file_already_downloaded(filename)
578 # Attempt to download using rtmpdump
579 if url.startswith('rtmp'):
580 return self._download_with_rtmpdump(filename, url, player_url)
582 tmpfilename = self.temp_name(filename)
585 # Do not include the Accept-Encoding header
586 headers = {'Youtubedl-no-compression': 'True'}
587 basic_request = compat_urllib_request.Request(url, None, headers)
588 request = compat_urllib_request.Request(url, None, headers)
590 # Establish possible resume length
591 if os.path.isfile(encodeFilename(tmpfilename)):
592 resume_len = os.path.getsize(encodeFilename(tmpfilename))
598 if self.params.get('continuedl', False):
599 self.report_resuming_byte(resume_len)
600 request.add_header('Range','bytes=%d-' % resume_len)
606 retries = self.params.get('retries', 0)
607 while count <= retries:
608 # Establish connection
610 if count == 0 and 'urlhandle' in info_dict:
611 data = info_dict['urlhandle']
612 data = compat_urllib_request.urlopen(request)
614 except (compat_urllib_error.HTTPError, ) as err:
615 if (err.code < 500 or err.code >= 600) and err.code != 416:
616 # Unexpected HTTP error
618 elif err.code == 416:
619 # Unable to resume (requested range not satisfiable)
621 # Open the connection again without the range header
622 data = compat_urllib_request.urlopen(basic_request)
623 content_length = data.info()['Content-Length']
624 except (compat_urllib_error.HTTPError, ) as err:
625 if err.code < 500 or err.code >= 600:
628 # Examine the reported length
629 if (content_length is not None and
630 (resume_len - 100 < int(content_length) < resume_len + 100)):
631 # The file had already been fully downloaded.
632 # Explanation to the above condition: in issue #175 it was revealed that
633 # YouTube sometimes adds or removes a few bytes from the end of the file,
634 # changing the file size slightly and causing problems for some users. So
635 # I decided to implement a suggested change and consider the file
636 # completely downloaded if the file size differs less than 100 bytes from
637 # the one in the hard drive.
638 self.report_file_already_downloaded(filename)
639 self.try_rename(tmpfilename, filename)
642 # The length does not match, we start the download over
643 self.report_unable_to_resume()
649 self.report_retry(count, retries)
652 self.trouble(u'ERROR: giving up after %s retries' % retries)
655 data_len = data.info().get('Content-length', None)
656 if data_len is not None:
657 data_len = int(data_len) + resume_len
658 data_len_str = self.format_bytes(data_len)
659 byte_counter = 0 + resume_len
660 block_size = self.params.get('buffersize', 1024)
665 data_block = data.read(block_size)
667 if len(data_block) == 0:
669 byte_counter += len(data_block)
671 # Open file just in time
674 (stream, tmpfilename) = sanitize_open(tmpfilename, open_mode)
675 assert stream is not None
676 filename = self.undo_temp_name(tmpfilename)
677 self.report_destination(filename)
678 except (OSError, IOError) as err:
679 self.trouble(u'ERROR: unable to open for writing: %s' % str(err))
682 stream.write(data_block)
683 except (IOError, OSError) as err:
684 self.trouble(u'\nERROR: unable to write data: %s' % str(err))
686 if not self.params.get('noresizebuffer', False):
687 block_size = self.best_block_size(after - before, len(data_block))
690 speed_str = self.calc_speed(start, time.time(), byte_counter - resume_len)
692 self.report_progress('Unknown %', data_len_str, speed_str, 'Unknown ETA')
694 percent_str = self.calc_percent(byte_counter, data_len)
695 eta_str = self.calc_eta(start, time.time(), data_len - resume_len, byte_counter - resume_len)
696 self.report_progress(percent_str, data_len_str, speed_str, eta_str)
699 self.slow_down(start, byte_counter - resume_len)
702 self.trouble(u'\nERROR: Did not get any data blocks')
706 if data_len is not None and byte_counter != data_len:
707 raise ContentTooShortError(byte_counter, int(data_len))
708 self.try_rename(tmpfilename, filename)
710 # Update file modification time
711 if self.params.get('updatetime', True):
712 info_dict['filetime'] = self.try_utime(filename, data.info().get('last-modified', None))