2 # -*- coding: utf-8 -*-
18 class FileDownloader(object):
19 """File Downloader class.
21 File downloader objects are the ones responsible of downloading the
22 actual video file and writing it to disk if the user has requested
23 it, among some other tasks. In most cases there should be one per
24 program. As, given a video URL, the downloader doesn't know how to
25 extract all the needed information, task that InfoExtractors do, it
26 has to pass the URL to one of them.
28 For this, file downloader objects have a method that allows
29 InfoExtractors to be registered in a given order. When it is passed
30 a URL, the file downloader handles it to the first InfoExtractor it
31 finds that reports being able to handle it. The InfoExtractor extracts
32 all the information about the video or videos the URL refers to, and
33 asks the FileDownloader to process the video information, possibly
34 downloading the video.
36 File downloaders accept a lot of parameters. In order not to saturate
37 the object constructor with arguments, it receives a dictionary of
38 options instead. These options are available through the params
39 attribute for the InfoExtractors to use. The FileDownloader also
40 registers itself as the downloader in charge for the InfoExtractors
41 that are added to it, so this is a "mutual registration".
45 username: Username for authentication purposes.
46 password: Password for authentication purposes.
47 usenetrc: Use netrc for authentication instead.
48 quiet: Do not print messages to stdout.
49 forceurl: Force printing final URL.
50 forcetitle: Force printing title.
51 forcethumbnail: Force printing thumbnail URL.
52 forcedescription: Force printing description.
53 forcefilename: Force printing final filename.
54 simulate: Do not download the video files.
55 format: Video format code.
56 format_limit: Highest quality format to try.
57 outtmpl: Template for output names.
58 restrictfilenames: Do not allow "&" and spaces in file names
59 ignoreerrors: Do not stop on download errors.
60 ratelimit: Download speed limit, in bytes/sec.
61 nooverwrites: Prevent overwriting files.
62 retries: Number of times to retry for HTTP error 5xx
63 buffersize: Size of download buffer in bytes.
64 noresizebuffer: Do not automatically resize the download buffer.
65 continuedl: Try to continue downloads if possible.
66 noprogress: Do not print the progress bar.
67 playliststart: Playlist item to start at.
68 playlistend: Playlist item to end at.
69 matchtitle: Download only matching titles.
70 rejecttitle: Reject downloads for matching titles.
71 logtostderr: Log messages to stderr instead of stdout.
72 consoletitle: Display progress in console window's titlebar.
73 nopart: Do not use temporary .part files.
74 updatetime: Use the Last-modified header to set output file timestamps.
75 writedescription: Write the video description to a .description file
76 writeinfojson: Write the video description to a .info.json file
77 writesubtitles: Write the video subtitles to a .srt file
78 subtitleslang: Language of the subtitles to download
84 _download_retcode = None
88 def __init__(self, params):
89 """Create a FileDownloader object with the given options."""
92 self._download_retcode = 0
93 self._num_downloads = 0
94 self._screen_file = [sys.stdout, sys.stderr][params.get('logtostderr', False)]
97 if '%(stitle)s' in self.params['outtmpl']:
98 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.')
101 def format_bytes(bytes):
104 if type(bytes) is str:
109 exponent = int(math.log(bytes, 1024.0))
110 suffix = 'bkMGTPEZY'[exponent]
111 converted = float(bytes) / float(1024 ** exponent)
112 return '%.2f%s' % (converted, suffix)
115 def calc_percent(byte_counter, data_len):
118 return '%6s' % ('%3.1f%%' % (float(byte_counter) / float(data_len) * 100.0))
121 def calc_eta(start, now, total, current):
125 if current == 0 or dif < 0.001: # One millisecond
127 rate = float(current) / dif
128 eta = int((float(total) - float(current)) / rate)
129 (eta_mins, eta_secs) = divmod(eta, 60)
132 return '%02d:%02d' % (eta_mins, eta_secs)
135 def calc_speed(start, now, bytes):
137 if bytes == 0 or dif < 0.001: # One millisecond
138 return '%10s' % '---b/s'
139 return '%10s' % ('%s/s' % FileDownloader.format_bytes(float(bytes) / dif))
142 def best_block_size(elapsed_time, bytes):
143 new_min = max(bytes / 2.0, 1.0)
144 new_max = min(max(bytes * 2.0, 1.0), 4194304) # Do not surpass 4 MB
145 if elapsed_time < 0.001:
147 rate = bytes / elapsed_time
155 def parse_bytes(bytestr):
156 """Parse a string indicating a byte quantity into an integer."""
157 matchobj = re.match(r'(?i)^(\d+(?:\.\d+)?)([kMGTPEZY]?)$', bytestr)
160 number = float(matchobj.group(1))
161 multiplier = 1024.0 ** 'bkmgtpezy'.index(matchobj.group(2).lower())
162 return int(round(number * multiplier))
164 def add_info_extractor(self, ie):
165 """Add an InfoExtractor object to the end of the list."""
167 ie.set_downloader(self)
169 def add_post_processor(self, pp):
170 """Add a PostProcessor object to the end of the chain."""
172 pp.set_downloader(self)
174 def to_screen(self, message, skip_eol=False):
175 """Print message to stdout if not in quiet mode."""
176 assert type(message) == type(u'')
177 if not self.params.get('quiet', False):
178 terminator = [u'\n', u''][skip_eol]
179 output = message + terminator
180 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
181 output = output.encode(preferredencoding(), 'ignore')
182 self._screen_file.write(output)
183 self._screen_file.flush()
185 def to_stderr(self, message):
186 """Print message to stderr."""
187 assert type(message) == type(u'')
188 output = message + u'\n'
189 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
190 output = output.encode(preferredencoding())
191 sys.stderr.write(output)
193 def to_cons_title(self, message):
194 """Set console/terminal window title to message."""
195 if not self.params.get('consoletitle', False):
197 if os.name == 'nt' and ctypes.windll.kernel32.GetConsoleWindow():
198 # c_wchar_p() might not be necessary if `message` is
199 # already of type unicode()
200 ctypes.windll.kernel32.SetConsoleTitleW(ctypes.c_wchar_p(message))
201 elif 'TERM' in os.environ:
202 sys.stderr.write('\033]0;%s\007' % message.encode(preferredencoding()))
204 def fixed_template(self):
205 """Checks if the output template is fixed."""
206 return (re.search(u'(?u)%\\(.+?\\)s', self.params['outtmpl']) is None)
208 def trouble(self, message=None):
209 """Determine action to take when a download problem appears.
211 Depending on if the downloader has been configured to ignore
212 download errors or not, this method may throw an exception or
213 not when errors are found, after printing the message.
215 if message is not None:
216 self.to_stderr(message)
217 if not self.params.get('ignoreerrors', False):
218 raise DownloadError(message)
219 self._download_retcode = 1
221 def slow_down(self, start_time, byte_counter):
222 """Sleep if the download speed is over the rate limit."""
223 rate_limit = self.params.get('ratelimit', None)
224 if rate_limit is None or byte_counter == 0:
227 elapsed = now - start_time
230 speed = float(byte_counter) / elapsed
231 if speed > rate_limit:
232 time.sleep((byte_counter - rate_limit * (now - start_time)) / rate_limit)
234 def temp_name(self, filename):
235 """Returns a temporary filename for the given filename."""
236 if self.params.get('nopart', False) or filename == u'-' or \
237 (os.path.exists(encodeFilename(filename)) and not os.path.isfile(encodeFilename(filename))):
239 return filename + u'.part'
241 def undo_temp_name(self, filename):
242 if filename.endswith(u'.part'):
243 return filename[:-len(u'.part')]
246 def try_rename(self, old_filename, new_filename):
248 if old_filename == new_filename:
250 os.rename(encodeFilename(old_filename), encodeFilename(new_filename))
251 except (IOError, OSError) as err:
252 self.trouble(u'ERROR: unable to rename file')
254 def try_utime(self, filename, last_modified_hdr):
255 """Try to set the last-modified time of the given file."""
256 if last_modified_hdr is None:
258 if not os.path.isfile(encodeFilename(filename)):
260 timestr = last_modified_hdr
263 filetime = timeconvert(timestr)
267 os.utime(filename, (time.time(), filetime))
272 def report_writedescription(self, descfn):
273 """ Report that the description file is being written """
274 self.to_screen(u'[info] Writing video description to: ' + descfn)
276 def report_writesubtitles(self, srtfn):
277 """ Report that the subtitles file is being written """
278 self.to_screen(u'[info] Writing video subtitles to: ' + srtfn)
280 def report_writeinfojson(self, infofn):
281 """ Report that the metadata file has been written """
282 self.to_screen(u'[info] Video description metadata as JSON to: ' + infofn)
284 def report_destination(self, filename):
285 """Report destination filename."""
286 self.to_screen(u'[download] Destination: ' + filename)
288 def report_progress(self, percent_str, data_len_str, speed_str, eta_str):
289 """Report download progress."""
290 if self.params.get('noprogress', False):
292 self.to_screen(u'\r[download] %s of %s at %s ETA %s' %
293 (percent_str, data_len_str, speed_str, eta_str), skip_eol=True)
294 self.to_cons_title(u'youtube-dl - %s of %s at %s ETA %s' %
295 (percent_str.strip(), data_len_str.strip(), speed_str.strip(), eta_str.strip()))
297 def report_resuming_byte(self, resume_len):
298 """Report attempt to resume at given byte."""
299 self.to_screen(u'[download] Resuming download at byte %s' % resume_len)
301 def report_retry(self, count, retries):
302 """Report retry in case of HTTP error 5xx"""
303 self.to_screen(u'[download] Got server HTTP error. Retrying (attempt %d of %d)...' % (count, retries))
305 def report_file_already_downloaded(self, file_name):
306 """Report file has already been fully downloaded."""
308 self.to_screen(u'[download] %s has already been downloaded' % file_name)
309 except (UnicodeEncodeError) as err:
310 self.to_screen(u'[download] The file has already been downloaded')
312 def report_unable_to_resume(self):
313 """Report it was impossible to resume download."""
314 self.to_screen(u'[download] Unable to resume')
316 def report_finish(self):
317 """Report download finished."""
318 if self.params.get('noprogress', False):
319 self.to_screen(u'[download] Download completed')
323 def increment_downloads(self):
324 """Increment the ordinal that assigns a number to each file."""
325 self._num_downloads += 1
327 def prepare_filename(self, info_dict):
328 """Generate the output filename."""
330 template_dict = dict(info_dict)
332 template_dict['epoch'] = int(time.time())
333 template_dict['autonumber'] = u'%05d' % self._num_downloads
335 template_dict = dict((key, u'NA' if val is None else val) for key, val in template_dict.items())
336 template_dict = dict((k, sanitize_filename(compat_str(v), self.params.get('restrictfilenames'))) for k,v in template_dict.items())
338 filename = self.params['outtmpl'] % template_dict
340 except (ValueError, KeyError) as err:
341 self.trouble(u'ERROR: invalid system charset or erroneous output template')
344 def _match_entry(self, info_dict):
345 """ Returns None iff the file should be downloaded """
347 title = info_dict['title']
348 matchtitle = self.params.get('matchtitle', False)
350 matchtitle = matchtitle.decode('utf8')
351 if not re.search(matchtitle, title, re.IGNORECASE):
352 return u'[download] "' + title + '" title did not match pattern "' + matchtitle + '"'
353 rejecttitle = self.params.get('rejecttitle', False)
355 rejecttitle = rejecttitle.decode('utf8')
356 if re.search(rejecttitle, title, re.IGNORECASE):
357 return u'"' + title + '" title matched reject pattern "' + rejecttitle + '"'
360 def process_info(self, info_dict):
361 """Process a single dictionary returned by an InfoExtractor."""
363 # Keep for backwards compatibility
364 info_dict['stitle'] = info_dict['title']
366 if not 'format' in info_dict:
367 info_dict['format'] = info_dict['ext']
369 reason = self._match_entry(info_dict)
370 if reason is not None:
371 self.to_screen(u'[download] ' + reason)
374 max_downloads = self.params.get('max_downloads')
375 if max_downloads is not None:
376 if self._num_downloads > int(max_downloads):
377 raise MaxDownloadsReached()
379 filename = self.prepare_filename(info_dict)
382 if self.params.get('forcetitle', False):
383 compat_print(info_dict['title'])
384 if self.params.get('forceurl', False):
385 compat_print(info_dict['url'])
386 if self.params.get('forcethumbnail', False) and 'thumbnail' in info_dict:
387 compat_print(info_dict['thumbnail'])
388 if self.params.get('forcedescription', False) and 'description' in info_dict:
389 compat_print(info_dict['description'])
390 if self.params.get('forcefilename', False) and filename is not None:
391 compat_print(filename)
392 if self.params.get('forceformat', False):
393 compat_print(info_dict['format'])
395 # Do nothing else if in simulate mode
396 if self.params.get('simulate', False):
403 dn = os.path.dirname(encodeFilename(filename))
404 if dn != '' and not os.path.exists(dn): # dn is already encoded
406 except (OSError, IOError) as err:
407 self.trouble(u'ERROR: unable to create directory ' + compat_str(err))
410 if self.params.get('writedescription', False):
412 descfn = filename + u'.description'
413 self.report_writedescription(descfn)
414 descfile = open(encodeFilename(descfn), 'wb')
416 descfile.write(info_dict['description'].encode('utf-8'))
419 except (OSError, IOError):
420 self.trouble(u'ERROR: Cannot write description file ' + descfn)
423 if self.params.get('writesubtitles', False) and 'subtitles' in info_dict and info_dict['subtitles']:
424 # subtitles download errors are already managed as troubles in relevant IE
425 # that way it will silently go on when used with unsupporting IE
427 srtfn = filename.rsplit('.', 1)[0] + u'.srt'
428 self.report_writesubtitles(srtfn)
429 srtfile = open(encodeFilename(srtfn), 'wb')
431 srtfile.write(info_dict['subtitles'].encode('utf-8'))
434 except (OSError, IOError):
435 self.trouble(u'ERROR: Cannot write subtitles file ' + descfn)
438 if self.params.get('writeinfojson', False):
439 infofn = filename + u'.info.json'
440 self.report_writeinfojson(infofn)
443 except (NameError,AttributeError):
444 self.trouble(u'ERROR: No JSON encoder found. Update to Python 2.6+, setup a json module, or leave out --write-info-json.')
447 infof = open(encodeFilename(infofn), 'wb')
449 json_info_dict = dict((k,v) for k,v in info_dict.iteritems() if not k in ('urlhandle',))
450 json.dump(json_info_dict, infof)
453 except (OSError, IOError):
454 self.trouble(u'ERROR: Cannot write metadata to JSON file ' + infofn)
457 if not self.params.get('skip_download', False):
458 if self.params.get('nooverwrites', False) and os.path.exists(encodeFilename(filename)):
462 success = self._do_download(filename, info_dict)
463 except (OSError, IOError) as err:
464 raise UnavailableVideoError
465 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
466 self.trouble(u'ERROR: unable to download video data: %s' % str(err))
468 except (ContentTooShortError, ) as err:
469 self.trouble(u'ERROR: content too short (expected %s bytes and served %s)' % (err.expected, err.downloaded))
474 self.post_process(filename, info_dict)
475 except (PostProcessingError) as err:
476 self.trouble(u'ERROR: postprocessing: %s' % str(err))
479 def download(self, url_list):
480 """Download a given list of URLs."""
481 if len(url_list) > 1 and self.fixed_template():
482 raise SameFileError(self.params['outtmpl'])
485 suitable_found = False
487 # Go to next InfoExtractor if not suitable
488 if not ie.suitable(url):
491 # Warn if the _WORKING attribute is False
493 self.trouble(u'WARNING: the program functionality for this site has been marked as broken, '
494 u'and will probably not work. If you want to go on, use the -i option.')
496 # Suitable InfoExtractor found
497 suitable_found = True
499 # Extract information from URL and process it
500 videos = ie.extract(url)
501 for video in videos or []:
502 video['extractor'] = ie.IE_NAME
504 self.increment_downloads()
505 self.process_info(video)
506 except UnavailableVideoError:
507 self.trouble(u'\nERROR: unable to download video')
509 # Suitable InfoExtractor had been found; go to next URL
512 if not suitable_found:
513 self.trouble(u'ERROR: no suitable InfoExtractor: %s' % url)
515 return self._download_retcode
517 def post_process(self, filename, ie_info):
518 """Run the postprocessing chain on the given file."""
520 info['filepath'] = filename
526 def _download_with_rtmpdump(self, filename, url, player_url):
527 self.report_destination(filename)
528 tmpfilename = self.temp_name(filename)
530 # Check for rtmpdump first
532 subprocess.call(['rtmpdump', '-h'], stdout=(file(os.path.devnull, 'w')), stderr=subprocess.STDOUT)
533 except (OSError, IOError):
534 self.trouble(u'ERROR: RTMP download detected but "rtmpdump" could not be run')
537 # Download using rtmpdump. rtmpdump returns exit code 2 when
538 # the connection was interrumpted and resuming appears to be
539 # possible. This is part of rtmpdump's normal usage, AFAIK.
540 basic_args = ['rtmpdump', '-q'] + [[], ['-W', player_url]][player_url is not None] + ['-r', url, '-o', tmpfilename]
541 args = basic_args + [[], ['-e', '-k', '1']][self.params.get('continuedl', False)]
542 if self.params.get('verbose', False):
545 shell_quote = lambda args: ' '.join(map(pipes.quote, args))
548 self.to_screen(u'[debug] rtmpdump command line: ' + shell_quote(args))
549 retval = subprocess.call(args)
550 while retval == 2 or retval == 1:
551 prevsize = os.path.getsize(encodeFilename(tmpfilename))
552 self.to_screen(u'\r[rtmpdump] %s bytes' % prevsize, skip_eol=True)
553 time.sleep(5.0) # This seems to be needed
554 retval = subprocess.call(basic_args + ['-e'] + [[], ['-k', '1']][retval == 1])
555 cursize = os.path.getsize(encodeFilename(tmpfilename))
556 if prevsize == cursize and retval == 1:
558 # Some rtmp streams seem abort after ~ 99.8%. Don't complain for those
559 if prevsize == cursize and retval == 2 and cursize > 1024:
560 self.to_screen(u'\r[rtmpdump] Could not download the whole video. This can happen for some advertisements.')
564 self.to_screen(u'\r[rtmpdump] %s bytes' % os.path.getsize(encodeFilename(tmpfilename)))
565 self.try_rename(tmpfilename, filename)
568 self.trouble(u'\nERROR: rtmpdump exited with code %d' % retval)
571 def _do_download(self, filename, info_dict):
572 url = info_dict['url']
573 player_url = info_dict.get('player_url', None)
575 # Check file already present
576 if self.params.get('continuedl', False) and os.path.isfile(encodeFilename(filename)) and not self.params.get('nopart', False):
577 self.report_file_already_downloaded(filename)
580 # Attempt to download using rtmpdump
581 if url.startswith('rtmp'):
582 return self._download_with_rtmpdump(filename, url, player_url)
584 tmpfilename = self.temp_name(filename)
587 # Do not include the Accept-Encoding header
588 headers = {'Youtubedl-no-compression': 'True'}
589 basic_request = compat_urllib_request.Request(url, None, headers)
590 request = compat_urllib_request.Request(url, None, headers)
592 # Establish possible resume length
593 if os.path.isfile(encodeFilename(tmpfilename)):
594 resume_len = os.path.getsize(encodeFilename(tmpfilename))
600 if self.params.get('continuedl', False):
601 self.report_resuming_byte(resume_len)
602 request.add_header('Range','bytes=%d-' % resume_len)
608 retries = self.params.get('retries', 0)
609 while count <= retries:
610 # Establish connection
612 if count == 0 and 'urlhandle' in info_dict:
613 data = info_dict['urlhandle']
614 data = compat_urllib_request.urlopen(request)
616 except (compat_urllib_error.HTTPError, ) as err:
617 if (err.code < 500 or err.code >= 600) and err.code != 416:
618 # Unexpected HTTP error
620 elif err.code == 416:
621 # Unable to resume (requested range not satisfiable)
623 # Open the connection again without the range header
624 data = compat_urllib_request.urlopen(basic_request)
625 content_length = data.info()['Content-Length']
626 except (compat_urllib_error.HTTPError, ) as err:
627 if err.code < 500 or err.code >= 600:
630 # Examine the reported length
631 if (content_length is not None and
632 (resume_len - 100 < int(content_length) < resume_len + 100)):
633 # The file had already been fully downloaded.
634 # Explanation to the above condition: in issue #175 it was revealed that
635 # YouTube sometimes adds or removes a few bytes from the end of the file,
636 # changing the file size slightly and causing problems for some users. So
637 # I decided to implement a suggested change and consider the file
638 # completely downloaded if the file size differs less than 100 bytes from
639 # the one in the hard drive.
640 self.report_file_already_downloaded(filename)
641 self.try_rename(tmpfilename, filename)
644 # The length does not match, we start the download over
645 self.report_unable_to_resume()
651 self.report_retry(count, retries)
654 self.trouble(u'ERROR: giving up after %s retries' % retries)
657 data_len = data.info().get('Content-length', None)
658 if data_len is not None:
659 data_len = int(data_len) + resume_len
660 data_len_str = self.format_bytes(data_len)
661 byte_counter = 0 + resume_len
662 block_size = self.params.get('buffersize', 1024)
667 data_block = data.read(block_size)
669 if len(data_block) == 0:
671 byte_counter += len(data_block)
673 # Open file just in time
676 (stream, tmpfilename) = sanitize_open(tmpfilename, open_mode)
677 assert stream is not None
678 filename = self.undo_temp_name(tmpfilename)
679 self.report_destination(filename)
680 except (OSError, IOError) as err:
681 self.trouble(u'ERROR: unable to open for writing: %s' % str(err))
684 stream.write(data_block)
685 except (IOError, OSError) as err:
686 self.trouble(u'\nERROR: unable to write data: %s' % str(err))
688 if not self.params.get('noresizebuffer', False):
689 block_size = self.best_block_size(after - before, len(data_block))
692 speed_str = self.calc_speed(start, time.time(), byte_counter - resume_len)
694 self.report_progress('Unknown %', data_len_str, speed_str, 'Unknown ETA')
696 percent_str = self.calc_percent(byte_counter, data_len)
697 eta_str = self.calc_eta(start, time.time(), data_len - resume_len, byte_counter - resume_len)
698 self.report_progress(percent_str, data_len_str, speed_str, eta_str)
701 self.slow_down(start, byte_counter - resume_len)
704 self.trouble(u'\nERROR: Did not get any data blocks')
708 if data_len is not None and byte_counter != data_len:
709 raise ContentTooShortError(byte_counter, int(data_len))
710 self.try_rename(tmpfilename, filename)
712 # Update file modification time
713 if self.params.get('updatetime', True):
714 info_dict['filetime'] = self.try_utime(filename, data.info().get('last-modified', None))