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):
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 if message is not None:
221 self.to_stderr(message)
222 if self.params.get('verbose'):
223 self.to_stderr(u''.join(traceback.format_list(traceback.extract_stack())))
224 if not self.params.get('ignoreerrors', False):
225 raise DownloadError(message)
226 self._download_retcode = 1
228 def slow_down(self, start_time, byte_counter):
229 """Sleep if the download speed is over the rate limit."""
230 rate_limit = self.params.get('ratelimit', None)
231 if rate_limit is None or byte_counter == 0:
234 elapsed = now - start_time
237 speed = float(byte_counter) / elapsed
238 if speed > rate_limit:
239 time.sleep((byte_counter - rate_limit * (now - start_time)) / rate_limit)
241 def temp_name(self, filename):
242 """Returns a temporary filename for the given filename."""
243 if self.params.get('nopart', False) or filename == u'-' or \
244 (os.path.exists(encodeFilename(filename)) and not os.path.isfile(encodeFilename(filename))):
246 return filename + u'.part'
248 def undo_temp_name(self, filename):
249 if filename.endswith(u'.part'):
250 return filename[:-len(u'.part')]
253 def try_rename(self, old_filename, new_filename):
255 if old_filename == new_filename:
257 os.rename(encodeFilename(old_filename), encodeFilename(new_filename))
258 except (IOError, OSError) as err:
259 self.trouble(u'ERROR: unable to rename file')
261 def try_utime(self, filename, last_modified_hdr):
262 """Try to set the last-modified time of the given file."""
263 if last_modified_hdr is None:
265 if not os.path.isfile(encodeFilename(filename)):
267 timestr = last_modified_hdr
270 filetime = timeconvert(timestr)
274 os.utime(filename, (time.time(), filetime))
279 def report_writedescription(self, descfn):
280 """ Report that the description file is being written """
281 self.to_screen(u'[info] Writing video description to: ' + descfn)
283 def report_writesubtitles(self, srtfn):
284 """ Report that the subtitles file is being written """
285 self.to_screen(u'[info] Writing video subtitles to: ' + srtfn)
287 def report_writeinfojson(self, infofn):
288 """ Report that the metadata file has been written """
289 self.to_screen(u'[info] Video description metadata as JSON to: ' + infofn)
291 def report_destination(self, filename):
292 """Report destination filename."""
293 self.to_screen(u'[download] Destination: ' + filename)
295 def report_progress(self, percent_str, data_len_str, speed_str, eta_str):
296 """Report download progress."""
297 if self.params.get('noprogress', False):
299 self.to_screen(u'\r[download] %s of %s at %s ETA %s' %
300 (percent_str, data_len_str, speed_str, eta_str), skip_eol=True)
301 self.to_cons_title(u'youtube-dl - %s of %s at %s ETA %s' %
302 (percent_str.strip(), data_len_str.strip(), speed_str.strip(), eta_str.strip()))
304 def report_resuming_byte(self, resume_len):
305 """Report attempt to resume at given byte."""
306 self.to_screen(u'[download] Resuming download at byte %s' % resume_len)
308 def report_retry(self, count, retries):
309 """Report retry in case of HTTP error 5xx"""
310 self.to_screen(u'[download] Got server HTTP error. Retrying (attempt %d of %d)...' % (count, retries))
312 def report_file_already_downloaded(self, file_name):
313 """Report file has already been fully downloaded."""
315 self.to_screen(u'[download] %s has already been downloaded' % file_name)
316 except (UnicodeEncodeError) as err:
317 self.to_screen(u'[download] The file has already been downloaded')
319 def report_unable_to_resume(self):
320 """Report it was impossible to resume download."""
321 self.to_screen(u'[download] Unable to resume')
323 def report_finish(self):
324 """Report download finished."""
325 if self.params.get('noprogress', False):
326 self.to_screen(u'[download] Download completed')
330 def increment_downloads(self):
331 """Increment the ordinal that assigns a number to each file."""
332 self._num_downloads += 1
334 def prepare_filename(self, info_dict):
335 """Generate the output filename."""
337 template_dict = dict(info_dict)
339 template_dict['epoch'] = int(time.time())
340 template_dict['autonumber'] = u'%05d' % self._num_downloads
342 sanitize = lambda k,v: sanitize_filename(
343 u'NA' if v is None else compat_str(v),
344 restricted=self.params.get('restrictfilenames'),
346 template_dict = dict((k, sanitize(k, v)) for k,v in template_dict.items())
348 filename = self.params['outtmpl'] % template_dict
350 except (ValueError, KeyError) as err:
351 self.trouble(u'ERROR: invalid system charset or erroneous output template')
354 def _match_entry(self, info_dict):
355 """ Returns None iff the file should be downloaded """
357 title = info_dict['title']
358 matchtitle = self.params.get('matchtitle', False)
360 matchtitle = matchtitle.decode('utf8')
361 if not re.search(matchtitle, title, re.IGNORECASE):
362 return u'[download] "' + title + '" title did not match pattern "' + matchtitle + '"'
363 rejecttitle = self.params.get('rejecttitle', False)
365 rejecttitle = rejecttitle.decode('utf8')
366 if re.search(rejecttitle, title, re.IGNORECASE):
367 return u'"' + title + '" title matched reject pattern "' + rejecttitle + '"'
370 def process_info(self, info_dict):
371 """Process a single dictionary returned by an InfoExtractor."""
373 # Keep for backwards compatibility
374 info_dict['stitle'] = info_dict['title']
376 if not 'format' in info_dict:
377 info_dict['format'] = info_dict['ext']
379 reason = self._match_entry(info_dict)
380 if reason is not None:
381 self.to_screen(u'[download] ' + reason)
384 max_downloads = self.params.get('max_downloads')
385 if max_downloads is not None:
386 if self._num_downloads > int(max_downloads):
387 raise MaxDownloadsReached()
389 filename = self.prepare_filename(info_dict)
392 if self.params.get('forcetitle', False):
393 compat_print(info_dict['title'])
394 if self.params.get('forceurl', False):
395 compat_print(info_dict['url'])
396 if self.params.get('forcethumbnail', False) and 'thumbnail' in info_dict:
397 compat_print(info_dict['thumbnail'])
398 if self.params.get('forcedescription', False) and 'description' in info_dict:
399 compat_print(info_dict['description'])
400 if self.params.get('forcefilename', False) and filename is not None:
401 compat_print(filename)
402 if self.params.get('forceformat', False):
403 compat_print(info_dict['format'])
405 # Do nothing else if in simulate mode
406 if self.params.get('simulate', False):
413 dn = os.path.dirname(encodeFilename(filename))
414 if dn != '' and not os.path.exists(dn): # dn is already encoded
416 except (OSError, IOError) as err:
417 self.trouble(u'ERROR: unable to create directory ' + compat_str(err))
420 if self.params.get('writedescription', False):
422 descfn = filename + u'.description'
423 self.report_writedescription(descfn)
424 descfile = open(encodeFilename(descfn), 'wb')
426 descfile.write(info_dict['description'].encode('utf-8'))
429 except (OSError, IOError):
430 self.trouble(u'ERROR: Cannot write description file ' + descfn)
433 if self.params.get('writesubtitles', False) and 'subtitles' in info_dict and info_dict['subtitles']:
434 # subtitles download errors are already managed as troubles in relevant IE
435 # that way it will silently go on when used with unsupporting IE
437 srtfn = filename.rsplit('.', 1)[0] + u'.srt'
438 self.report_writesubtitles(srtfn)
439 srtfile = open(encodeFilename(srtfn), 'wb')
441 srtfile.write(info_dict['subtitles'].encode('utf-8'))
444 except (OSError, IOError):
445 self.trouble(u'ERROR: Cannot write subtitles file ' + descfn)
448 if self.params.get('writeinfojson', False):
449 infofn = filename + u'.info.json'
450 self.report_writeinfojson(infofn)
453 except (NameError,AttributeError):
454 self.trouble(u'ERROR: No JSON encoder found. Update to Python 2.6+, setup a json module, or leave out --write-info-json.')
457 with io.open(encodeFilename(infofn), 'w', 'utf-8') as infof:
458 json_info_dict = dict((k, v) for k,v in info_dict.items() if not k in ['urlhandle'])
459 json.dump(json_info_dict, infof)
460 except (OSError, IOError):
461 self.trouble(u'ERROR: Cannot write metadata to JSON file ' + infofn)
464 if not self.params.get('skip_download', False):
465 if self.params.get('nooverwrites', False) and os.path.exists(encodeFilename(filename)):
469 success = self._do_download(filename, info_dict)
470 except (OSError, IOError) as err:
471 raise UnavailableVideoError()
472 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
473 self.trouble(u'ERROR: unable to download video data: %s' % str(err))
475 except (ContentTooShortError, ) as err:
476 self.trouble(u'ERROR: content too short (expected %s bytes and served %s)' % (err.expected, err.downloaded))
481 self.post_process(filename, info_dict)
482 except (PostProcessingError) as err:
483 self.trouble(u'ERROR: postprocessing: %s' % str(err))
486 def download(self, url_list):
487 """Download a given list of URLs."""
488 if len(url_list) > 1 and self.fixed_template():
489 raise SameFileError(self.params['outtmpl'])
492 suitable_found = False
494 # Go to next InfoExtractor if not suitable
495 if not ie.suitable(url):
498 # Warn if the _WORKING attribute is False
500 self.trouble(u'WARNING: the program functionality for this site has been marked as broken, '
501 u'and will probably not work. If you want to go on, use the -i option.')
503 # Suitable InfoExtractor found
504 suitable_found = True
506 # Extract information from URL and process it
507 videos = ie.extract(url)
508 for video in videos or []:
509 video['extractor'] = ie.IE_NAME
511 self.increment_downloads()
512 self.process_info(video)
513 except UnavailableVideoError:
514 self.trouble(u'\nERROR: unable to download video')
516 # Suitable InfoExtractor had been found; go to next URL
519 if not suitable_found:
520 self.trouble(u'ERROR: no suitable InfoExtractor: %s' % url)
522 return self._download_retcode
524 def post_process(self, filename, ie_info):
525 """Run the postprocessing chain on the given file."""
527 info['filepath'] = filename
533 def _download_with_rtmpdump(self, filename, url, player_url):
534 self.report_destination(filename)
535 tmpfilename = self.temp_name(filename)
537 # Check for rtmpdump first
539 subprocess.call(['rtmpdump', '-h'], stdout=(file(os.path.devnull, 'w')), stderr=subprocess.STDOUT)
540 except (OSError, IOError):
541 self.trouble(u'ERROR: RTMP download detected but "rtmpdump" could not be run')
544 # Download using rtmpdump. rtmpdump returns exit code 2 when
545 # the connection was interrumpted and resuming appears to be
546 # possible. This is part of rtmpdump's normal usage, AFAIK.
547 basic_args = ['rtmpdump', '-q'] + [[], ['-W', player_url]][player_url is not None] + ['-r', url, '-o', tmpfilename]
548 args = basic_args + [[], ['-e', '-k', '1']][self.params.get('continuedl', False)]
549 if self.params.get('verbose', False):
552 shell_quote = lambda args: ' '.join(map(pipes.quote, args))
555 self.to_screen(u'[debug] rtmpdump command line: ' + shell_quote(args))
556 retval = subprocess.call(args)
557 while retval == 2 or retval == 1:
558 prevsize = os.path.getsize(encodeFilename(tmpfilename))
559 self.to_screen(u'\r[rtmpdump] %s bytes' % prevsize, skip_eol=True)
560 time.sleep(5.0) # This seems to be needed
561 retval = subprocess.call(basic_args + ['-e'] + [[], ['-k', '1']][retval == 1])
562 cursize = os.path.getsize(encodeFilename(tmpfilename))
563 if prevsize == cursize and retval == 1:
565 # Some rtmp streams seem abort after ~ 99.8%. Don't complain for those
566 if prevsize == cursize and retval == 2 and cursize > 1024:
567 self.to_screen(u'\r[rtmpdump] Could not download the whole video. This can happen for some advertisements.')
571 self.to_screen(u'\r[rtmpdump] %s bytes' % os.path.getsize(encodeFilename(tmpfilename)))
572 self.try_rename(tmpfilename, filename)
575 self.trouble(u'\nERROR: rtmpdump exited with code %d' % retval)
578 def _do_download(self, filename, info_dict):
579 url = info_dict['url']
580 player_url = info_dict.get('player_url', None)
582 # Check file already present
583 if self.params.get('continuedl', False) and os.path.isfile(encodeFilename(filename)) and not self.params.get('nopart', False):
584 self.report_file_already_downloaded(filename)
587 # Attempt to download using rtmpdump
588 if url.startswith('rtmp'):
589 return self._download_with_rtmpdump(filename, url, player_url)
591 tmpfilename = self.temp_name(filename)
594 # Do not include the Accept-Encoding header
595 headers = {'Youtubedl-no-compression': 'True'}
596 basic_request = compat_urllib_request.Request(url, None, headers)
597 request = compat_urllib_request.Request(url, None, headers)
599 if self.params.get('test', False):
600 request.add_header('Range','bytes=0-10240')
602 # Establish possible resume length
603 if os.path.isfile(encodeFilename(tmpfilename)):
604 resume_len = os.path.getsize(encodeFilename(tmpfilename))
610 if self.params.get('continuedl', False):
611 self.report_resuming_byte(resume_len)
612 request.add_header('Range','bytes=%d-' % resume_len)
618 retries = self.params.get('retries', 0)
619 while count <= retries:
620 # Establish connection
622 if count == 0 and 'urlhandle' in info_dict:
623 data = info_dict['urlhandle']
624 data = compat_urllib_request.urlopen(request)
626 except (compat_urllib_error.HTTPError, ) as err:
627 if (err.code < 500 or err.code >= 600) and err.code != 416:
628 # Unexpected HTTP error
630 elif err.code == 416:
631 # Unable to resume (requested range not satisfiable)
633 # Open the connection again without the range header
634 data = compat_urllib_request.urlopen(basic_request)
635 content_length = data.info()['Content-Length']
636 except (compat_urllib_error.HTTPError, ) as err:
637 if err.code < 500 or err.code >= 600:
640 # Examine the reported length
641 if (content_length is not None and
642 (resume_len - 100 < int(content_length) < resume_len + 100)):
643 # The file had already been fully downloaded.
644 # Explanation to the above condition: in issue #175 it was revealed that
645 # YouTube sometimes adds or removes a few bytes from the end of the file,
646 # changing the file size slightly and causing problems for some users. So
647 # I decided to implement a suggested change and consider the file
648 # completely downloaded if the file size differs less than 100 bytes from
649 # the one in the hard drive.
650 self.report_file_already_downloaded(filename)
651 self.try_rename(tmpfilename, filename)
654 # The length does not match, we start the download over
655 self.report_unable_to_resume()
661 self.report_retry(count, retries)
664 self.trouble(u'ERROR: giving up after %s retries' % retries)
667 data_len = data.info().get('Content-length', None)
668 if data_len is not None:
669 data_len = int(data_len) + resume_len
670 data_len_str = self.format_bytes(data_len)
671 byte_counter = 0 + resume_len
672 block_size = self.params.get('buffersize', 1024)
677 data_block = data.read(block_size)
679 if len(data_block) == 0:
681 byte_counter += len(data_block)
683 # Open file just in time
686 (stream, tmpfilename) = sanitize_open(tmpfilename, open_mode)
687 assert stream is not None
688 filename = self.undo_temp_name(tmpfilename)
689 self.report_destination(filename)
690 except (OSError, IOError) as err:
691 self.trouble(u'ERROR: unable to open for writing: %s' % str(err))
694 stream.write(data_block)
695 except (IOError, OSError) as err:
696 self.trouble(u'\nERROR: unable to write data: %s' % str(err))
698 if not self.params.get('noresizebuffer', False):
699 block_size = self.best_block_size(after - before, len(data_block))
702 speed_str = self.calc_speed(start, time.time(), byte_counter - resume_len)
704 self.report_progress('Unknown %', data_len_str, speed_str, 'Unknown ETA')
706 percent_str = self.calc_percent(byte_counter, data_len)
707 eta_str = self.calc_eta(start, time.time(), data_len - resume_len, byte_counter - resume_len)
708 self.report_progress(percent_str, data_len_str, speed_str, eta_str)
711 self.slow_down(start, byte_counter - resume_len)
714 self.trouble(u'\nERROR: Did not get any data blocks')
718 if data_len is not None and byte_counter != data_len:
719 raise ContentTooShortError(byte_counter, int(data_len))
720 self.try_rename(tmpfilename, filename)
722 # Update file modification time
723 if self.params.get('updatetime', True):
724 info_dict['filetime'] = self.try_utime(filename, data.info().get('last-modified', None))