2 # -*- coding: utf-8 -*-
4 from __future__ import absolute_import
20 class FileDownloader(object):
21 """File Downloader class.
23 File downloader objects are the ones responsible of downloading the
24 actual video file and writing it to disk if the user has requested
25 it, among some other tasks. In most cases there should be one per
26 program. As, given a video URL, the downloader doesn't know how to
27 extract all the needed information, task that InfoExtractors do, it
28 has to pass the URL to one of them.
30 For this, file downloader objects have a method that allows
31 InfoExtractors to be registered in a given order. When it is passed
32 a URL, the file downloader handles it to the first InfoExtractor it
33 finds that reports being able to handle it. The InfoExtractor extracts
34 all the information about the video or videos the URL refers to, and
35 asks the FileDownloader to process the video information, possibly
36 downloading the video.
38 File downloaders accept a lot of parameters. In order not to saturate
39 the object constructor with arguments, it receives a dictionary of
40 options instead. These options are available through the params
41 attribute for the InfoExtractors to use. The FileDownloader also
42 registers itself as the downloader in charge for the InfoExtractors
43 that are added to it, so this is a "mutual registration".
47 username: Username for authentication purposes.
48 password: Password for authentication purposes.
49 usenetrc: Use netrc for authentication instead.
50 quiet: Do not print messages to stdout.
51 forceurl: Force printing final URL.
52 forcetitle: Force printing title.
53 forcethumbnail: Force printing thumbnail URL.
54 forcedescription: Force printing description.
55 forcefilename: Force printing final filename.
56 simulate: Do not download the video files.
57 format: Video format code.
58 format_limit: Highest quality format to try.
59 outtmpl: Template for output names.
60 restrictfilenames: Do not allow "&" and spaces in file names
61 ignoreerrors: Do not stop on download errors.
62 ratelimit: Download speed limit, in bytes/sec.
63 nooverwrites: Prevent overwriting files.
64 retries: Number of times to retry for HTTP error 5xx
65 buffersize: Size of download buffer in bytes.
66 noresizebuffer: Do not automatically resize the download buffer.
67 continuedl: Try to continue downloads if possible.
68 noprogress: Do not print the progress bar.
69 playliststart: Playlist item to start at.
70 playlistend: Playlist item to end at.
71 matchtitle: Download only matching titles.
72 rejecttitle: Reject downloads for matching titles.
73 logtostderr: Log messages to stderr instead of stdout.
74 consoletitle: Display progress in console window's titlebar.
75 nopart: Do not use temporary .part files.
76 updatetime: Use the Last-modified header to set output file timestamps.
77 writedescription: Write the video description to a .description file
78 writeinfojson: Write the video description to a .info.json file
79 writesubtitles: Write the video subtitles to a .srt file
80 subtitleslang: Language of the subtitles to download
86 _download_retcode = None
90 def __init__(self, params):
91 """Create a FileDownloader object with the given options."""
94 self._download_retcode = 0
95 self._num_downloads = 0
96 self._screen_file = [sys.stdout, sys.stderr][params.get('logtostderr', False)]
99 if '%(stitle)s' in self.params['outtmpl']:
100 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.')
103 def format_bytes(bytes):
106 if type(bytes) is str:
111 exponent = int(math.log(bytes, 1024.0))
112 suffix = 'bkMGTPEZY'[exponent]
113 converted = float(bytes) / float(1024 ** exponent)
114 return '%.2f%s' % (converted, suffix)
117 def calc_percent(byte_counter, data_len):
120 return '%6s' % ('%3.1f%%' % (float(byte_counter) / float(data_len) * 100.0))
123 def calc_eta(start, now, total, current):
127 if current == 0 or dif < 0.001: # One millisecond
129 rate = float(current) / dif
130 eta = int((float(total) - float(current)) / rate)
131 (eta_mins, eta_secs) = divmod(eta, 60)
134 return '%02d:%02d' % (eta_mins, eta_secs)
137 def calc_speed(start, now, bytes):
139 if bytes == 0 or dif < 0.001: # One millisecond
140 return '%10s' % '---b/s'
141 return '%10s' % ('%s/s' % FileDownloader.format_bytes(float(bytes) / dif))
144 def best_block_size(elapsed_time, bytes):
145 new_min = max(bytes / 2.0, 1.0)
146 new_max = min(max(bytes * 2.0, 1.0), 4194304) # Do not surpass 4 MB
147 if elapsed_time < 0.001:
149 rate = bytes / elapsed_time
157 def parse_bytes(bytestr):
158 """Parse a string indicating a byte quantity into an integer."""
159 matchobj = re.match(r'(?i)^(\d+(?:\.\d+)?)([kMGTPEZY]?)$', bytestr)
162 number = float(matchobj.group(1))
163 multiplier = 1024.0 ** 'bkmgtpezy'.index(matchobj.group(2).lower())
164 return int(round(number * multiplier))
166 def add_info_extractor(self, ie):
167 """Add an InfoExtractor object to the end of the list."""
169 ie.set_downloader(self)
171 def add_post_processor(self, pp):
172 """Add a PostProcessor object to the end of the chain."""
174 pp.set_downloader(self)
176 def to_screen(self, message, skip_eol=False):
177 """Print message to stdout if not in quiet mode."""
178 assert type(message) == type(u'')
179 if not self.params.get('quiet', False):
180 terminator = [u'\n', u''][skip_eol]
181 output = message + terminator
182 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
183 output = output.encode(preferredencoding(), 'ignore')
184 self._screen_file.write(output)
185 self._screen_file.flush()
187 def to_stderr(self, message):
188 """Print message to stderr."""
189 assert type(message) == type(u'')
190 output = message + u'\n'
191 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
192 output = output.encode(preferredencoding())
193 sys.stderr.write(output)
195 def to_cons_title(self, message):
196 """Set console/terminal window title to message."""
197 if not self.params.get('consoletitle', False):
199 if os.name == 'nt' and ctypes.windll.kernel32.GetConsoleWindow():
200 # c_wchar_p() might not be necessary if `message` is
201 # already of type unicode()
202 ctypes.windll.kernel32.SetConsoleTitleW(ctypes.c_wchar_p(message))
203 elif 'TERM' in os.environ:
204 sys.stderr.write('\033]0;%s\007' % message.encode(preferredencoding()))
206 def fixed_template(self):
207 """Checks if the output template is fixed."""
208 return (re.search(u'(?u)%\\(.+?\\)s', self.params['outtmpl']) is None)
210 def trouble(self, message=None):
211 """Determine action to take when a download problem appears.
213 Depending on if the downloader has been configured to ignore
214 download errors or not, this method may throw an exception or
215 not when errors are found, after printing the message.
217 if message is not None:
218 self.to_stderr(message)
219 if not self.params.get('ignoreerrors', False):
220 raise DownloadError(message)
221 self._download_retcode = 1
223 def slow_down(self, start_time, byte_counter):
224 """Sleep if the download speed is over the rate limit."""
225 rate_limit = self.params.get('ratelimit', None)
226 if rate_limit is None or byte_counter == 0:
229 elapsed = now - start_time
232 speed = float(byte_counter) / elapsed
233 if speed > rate_limit:
234 time.sleep((byte_counter - rate_limit * (now - start_time)) / rate_limit)
236 def temp_name(self, filename):
237 """Returns a temporary filename for the given filename."""
238 if self.params.get('nopart', False) or filename == u'-' or \
239 (os.path.exists(encodeFilename(filename)) and not os.path.isfile(encodeFilename(filename))):
241 return filename + u'.part'
243 def undo_temp_name(self, filename):
244 if filename.endswith(u'.part'):
245 return filename[:-len(u'.part')]
248 def try_rename(self, old_filename, new_filename):
250 if old_filename == new_filename:
252 os.rename(encodeFilename(old_filename), encodeFilename(new_filename))
253 except (IOError, OSError) as err:
254 self.trouble(u'ERROR: unable to rename file')
256 def try_utime(self, filename, last_modified_hdr):
257 """Try to set the last-modified time of the given file."""
258 if last_modified_hdr is None:
260 if not os.path.isfile(encodeFilename(filename)):
262 timestr = last_modified_hdr
265 filetime = timeconvert(timestr)
269 os.utime(filename, (time.time(), filetime))
274 def report_writedescription(self, descfn):
275 """ Report that the description file is being written """
276 self.to_screen(u'[info] Writing video description to: ' + descfn)
278 def report_writesubtitles(self, srtfn):
279 """ Report that the subtitles file is being written """
280 self.to_screen(u'[info] Writing video subtitles to: ' + srtfn)
282 def report_writeinfojson(self, infofn):
283 """ Report that the metadata file has been written """
284 self.to_screen(u'[info] Video description metadata as JSON to: ' + infofn)
286 def report_destination(self, filename):
287 """Report destination filename."""
288 self.to_screen(u'[download] Destination: ' + filename)
290 def report_progress(self, percent_str, data_len_str, speed_str, eta_str):
291 """Report download progress."""
292 if self.params.get('noprogress', False):
294 self.to_screen(u'\r[download] %s of %s at %s ETA %s' %
295 (percent_str, data_len_str, speed_str, eta_str), skip_eol=True)
296 self.to_cons_title(u'youtube-dl - %s of %s at %s ETA %s' %
297 (percent_str.strip(), data_len_str.strip(), speed_str.strip(), eta_str.strip()))
299 def report_resuming_byte(self, resume_len):
300 """Report attempt to resume at given byte."""
301 self.to_screen(u'[download] Resuming download at byte %s' % resume_len)
303 def report_retry(self, count, retries):
304 """Report retry in case of HTTP error 5xx"""
305 self.to_screen(u'[download] Got server HTTP error. Retrying (attempt %d of %d)...' % (count, retries))
307 def report_file_already_downloaded(self, file_name):
308 """Report file has already been fully downloaded."""
310 self.to_screen(u'[download] %s has already been downloaded' % file_name)
311 except (UnicodeEncodeError) as err:
312 self.to_screen(u'[download] The file has already been downloaded')
314 def report_unable_to_resume(self):
315 """Report it was impossible to resume download."""
316 self.to_screen(u'[download] Unable to resume')
318 def report_finish(self):
319 """Report download finished."""
320 if self.params.get('noprogress', False):
321 self.to_screen(u'[download] Download completed')
325 def increment_downloads(self):
326 """Increment the ordinal that assigns a number to each file."""
327 self._num_downloads += 1
329 def prepare_filename(self, info_dict):
330 """Generate the output filename."""
332 template_dict = dict(info_dict)
334 template_dict['epoch'] = int(time.time())
335 template_dict['autonumber'] = u'%05d' % self._num_downloads
337 template_dict = dict((key, u'NA' if val is None else val) for key, val in template_dict.items())
338 template_dict = dict((k, sanitize_filename(compat_str(v), self.params.get('restrictfilenames'))) for k,v in template_dict.items())
340 filename = self.params['outtmpl'] % template_dict
342 except (ValueError, KeyError) as err:
343 self.trouble(u'ERROR: invalid system charset or erroneous output template')
346 def _match_entry(self, info_dict):
347 """ Returns None iff the file should be downloaded """
349 title = info_dict['title']
350 matchtitle = self.params.get('matchtitle', False)
352 matchtitle = matchtitle.decode('utf8')
353 if not re.search(matchtitle, title, re.IGNORECASE):
354 return u'[download] "' + title + '" title did not match pattern "' + matchtitle + '"'
355 rejecttitle = self.params.get('rejecttitle', False)
357 rejecttitle = rejecttitle.decode('utf8')
358 if re.search(rejecttitle, title, re.IGNORECASE):
359 return u'"' + title + '" title matched reject pattern "' + rejecttitle + '"'
362 def process_info(self, info_dict):
363 """Process a single dictionary returned by an InfoExtractor."""
365 # Keep for backwards compatibility
366 info_dict['stitle'] = info_dict['title']
368 if not 'format' in info_dict:
369 info_dict['format'] = info_dict['ext']
371 reason = self._match_entry(info_dict)
372 if reason is not None:
373 self.to_screen(u'[download] ' + reason)
376 max_downloads = self.params.get('max_downloads')
377 if max_downloads is not None:
378 if self._num_downloads > int(max_downloads):
379 raise MaxDownloadsReached()
381 filename = self.prepare_filename(info_dict)
384 if self.params.get('forcetitle', False):
385 compat_print(info_dict['title'])
386 if self.params.get('forceurl', False):
387 compat_print(info_dict['url'])
388 if self.params.get('forcethumbnail', False) and 'thumbnail' in info_dict:
389 compat_print(info_dict['thumbnail'])
390 if self.params.get('forcedescription', False) and 'description' in info_dict:
391 compat_print(info_dict['description'])
392 if self.params.get('forcefilename', False) and filename is not None:
393 compat_print(filename)
394 if self.params.get('forceformat', False):
395 compat_print(info_dict['format'])
397 # Do nothing else if in simulate mode
398 if self.params.get('simulate', False):
405 dn = os.path.dirname(encodeFilename(filename))
406 if dn != '' and not os.path.exists(dn): # dn is already encoded
408 except (OSError, IOError) as err:
409 self.trouble(u'ERROR: unable to create directory ' + compat_str(err))
412 if self.params.get('writedescription', False):
414 descfn = filename + u'.description'
415 self.report_writedescription(descfn)
416 descfile = open(encodeFilename(descfn), 'wb')
418 descfile.write(info_dict['description'].encode('utf-8'))
421 except (OSError, IOError):
422 self.trouble(u'ERROR: Cannot write description file ' + descfn)
425 if self.params.get('writesubtitles', False) and 'subtitles' in info_dict and info_dict['subtitles']:
426 # subtitles download errors are already managed as troubles in relevant IE
427 # that way it will silently go on when used with unsupporting IE
429 srtfn = filename.rsplit('.', 1)[0] + u'.srt'
430 self.report_writesubtitles(srtfn)
431 srtfile = open(encodeFilename(srtfn), 'wb')
433 srtfile.write(info_dict['subtitles'].encode('utf-8'))
436 except (OSError, IOError):
437 self.trouble(u'ERROR: Cannot write subtitles file ' + descfn)
440 if self.params.get('writeinfojson', False):
441 infofn = filename + u'.info.json'
442 self.report_writeinfojson(infofn)
445 except (NameError,AttributeError):
446 self.trouble(u'ERROR: No JSON encoder found. Update to Python 2.6+, setup a json module, or leave out --write-info-json.')
449 infof = open(encodeFilename(infofn), 'wb')
451 json_info_dict = dict((k,v) for k,v in info_dict.iteritems() if not k in ('urlhandle',))
452 json.dump(json_info_dict, infof)
455 except (OSError, IOError):
456 self.trouble(u'ERROR: Cannot write metadata to JSON file ' + infofn)
459 if not self.params.get('skip_download', False):
460 if self.params.get('nooverwrites', False) and os.path.exists(encodeFilename(filename)):
464 success = self._do_download(filename, info_dict)
465 except (OSError, IOError) as err:
466 raise UnavailableVideoError()
467 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
468 self.trouble(u'ERROR: unable to download video data: %s' % str(err))
470 except (ContentTooShortError, ) as err:
471 self.trouble(u'ERROR: content too short (expected %s bytes and served %s)' % (err.expected, err.downloaded))
476 self.post_process(filename, info_dict)
477 except (PostProcessingError) as err:
478 self.trouble(u'ERROR: postprocessing: %s' % str(err))
481 def download(self, url_list):
482 """Download a given list of URLs."""
483 if len(url_list) > 1 and self.fixed_template():
484 raise SameFileError(self.params['outtmpl'])
487 suitable_found = False
489 # Go to next InfoExtractor if not suitable
490 if not ie.suitable(url):
493 # Warn if the _WORKING attribute is False
495 self.trouble(u'WARNING: the program functionality for this site has been marked as broken, '
496 u'and will probably not work. If you want to go on, use the -i option.')
498 # Suitable InfoExtractor found
499 suitable_found = True
501 # Extract information from URL and process it
502 videos = ie.extract(url)
503 for video in videos or []:
504 video['extractor'] = ie.IE_NAME
506 self.increment_downloads()
507 self.process_info(video)
508 except UnavailableVideoError:
509 self.trouble(u'\nERROR: unable to download video')
511 # Suitable InfoExtractor had been found; go to next URL
514 if not suitable_found:
515 self.trouble(u'ERROR: no suitable InfoExtractor: %s' % url)
517 return self._download_retcode
519 def post_process(self, filename, ie_info):
520 """Run the postprocessing chain on the given file."""
522 info['filepath'] = filename
528 def _download_with_rtmpdump(self, filename, url, player_url):
529 self.report_destination(filename)
530 tmpfilename = self.temp_name(filename)
532 # Check for rtmpdump first
534 subprocess.call(['rtmpdump', '-h'], stdout=(file(os.path.devnull, 'w')), stderr=subprocess.STDOUT)
535 except (OSError, IOError):
536 self.trouble(u'ERROR: RTMP download detected but "rtmpdump" could not be run')
539 # Download using rtmpdump. rtmpdump returns exit code 2 when
540 # the connection was interrumpted and resuming appears to be
541 # possible. This is part of rtmpdump's normal usage, AFAIK.
542 basic_args = ['rtmpdump', '-q'] + [[], ['-W', player_url]][player_url is not None] + ['-r', url, '-o', tmpfilename]
543 args = basic_args + [[], ['-e', '-k', '1']][self.params.get('continuedl', False)]
544 if self.params.get('verbose', False):
547 shell_quote = lambda args: ' '.join(map(pipes.quote, args))
550 self.to_screen(u'[debug] rtmpdump command line: ' + shell_quote(args))
551 retval = subprocess.call(args)
552 while retval == 2 or retval == 1:
553 prevsize = os.path.getsize(encodeFilename(tmpfilename))
554 self.to_screen(u'\r[rtmpdump] %s bytes' % prevsize, skip_eol=True)
555 time.sleep(5.0) # This seems to be needed
556 retval = subprocess.call(basic_args + ['-e'] + [[], ['-k', '1']][retval == 1])
557 cursize = os.path.getsize(encodeFilename(tmpfilename))
558 if prevsize == cursize and retval == 1:
560 # Some rtmp streams seem abort after ~ 99.8%. Don't complain for those
561 if prevsize == cursize and retval == 2 and cursize > 1024:
562 self.to_screen(u'\r[rtmpdump] Could not download the whole video. This can happen for some advertisements.')
566 self.to_screen(u'\r[rtmpdump] %s bytes' % os.path.getsize(encodeFilename(tmpfilename)))
567 self.try_rename(tmpfilename, filename)
570 self.trouble(u'\nERROR: rtmpdump exited with code %d' % retval)
573 def _do_download(self, filename, info_dict):
574 url = info_dict['url']
575 player_url = info_dict.get('player_url', None)
577 # Check file already present
578 if self.params.get('continuedl', False) and os.path.isfile(encodeFilename(filename)) and not self.params.get('nopart', False):
579 self.report_file_already_downloaded(filename)
582 # Attempt to download using rtmpdump
583 if url.startswith('rtmp'):
584 return self._download_with_rtmpdump(filename, url, player_url)
586 tmpfilename = self.temp_name(filename)
589 # Do not include the Accept-Encoding header
590 headers = {'Youtubedl-no-compression': 'True'}
591 basic_request = compat_urllib_request.Request(url, None, headers)
592 request = compat_urllib_request.Request(url, None, headers)
594 # Establish possible resume length
595 if os.path.isfile(encodeFilename(tmpfilename)):
596 resume_len = os.path.getsize(encodeFilename(tmpfilename))
602 if self.params.get('continuedl', False):
603 self.report_resuming_byte(resume_len)
604 request.add_header('Range','bytes=%d-' % resume_len)
610 retries = self.params.get('retries', 0)
611 while count <= retries:
612 # Establish connection
614 if count == 0 and 'urlhandle' in info_dict:
615 data = info_dict['urlhandle']
616 data = compat_urllib_request.urlopen(request)
618 except (compat_urllib_error.HTTPError, ) as err:
619 if (err.code < 500 or err.code >= 600) and err.code != 416:
620 # Unexpected HTTP error
622 elif err.code == 416:
623 # Unable to resume (requested range not satisfiable)
625 # Open the connection again without the range header
626 data = compat_urllib_request.urlopen(basic_request)
627 content_length = data.info()['Content-Length']
628 except (compat_urllib_error.HTTPError, ) as err:
629 if err.code < 500 or err.code >= 600:
632 # Examine the reported length
633 if (content_length is not None and
634 (resume_len - 100 < int(content_length) < resume_len + 100)):
635 # The file had already been fully downloaded.
636 # Explanation to the above condition: in issue #175 it was revealed that
637 # YouTube sometimes adds or removes a few bytes from the end of the file,
638 # changing the file size slightly and causing problems for some users. So
639 # I decided to implement a suggested change and consider the file
640 # completely downloaded if the file size differs less than 100 bytes from
641 # the one in the hard drive.
642 self.report_file_already_downloaded(filename)
643 self.try_rename(tmpfilename, filename)
646 # The length does not match, we start the download over
647 self.report_unable_to_resume()
653 self.report_retry(count, retries)
656 self.trouble(u'ERROR: giving up after %s retries' % retries)
659 data_len = data.info().get('Content-length', None)
660 if data_len is not None:
661 data_len = int(data_len) + resume_len
662 data_len_str = self.format_bytes(data_len)
663 byte_counter = 0 + resume_len
664 block_size = self.params.get('buffersize', 1024)
669 data_block = data.read(block_size)
671 if len(data_block) == 0:
673 byte_counter += len(data_block)
675 # Open file just in time
678 (stream, tmpfilename) = sanitize_open(tmpfilename, open_mode)
679 assert stream is not None
680 filename = self.undo_temp_name(tmpfilename)
681 self.report_destination(filename)
682 except (OSError, IOError) as err:
683 self.trouble(u'ERROR: unable to open for writing: %s' % str(err))
686 stream.write(data_block)
687 except (IOError, OSError) as err:
688 self.trouble(u'\nERROR: unable to write data: %s' % str(err))
690 if not self.params.get('noresizebuffer', False):
691 block_size = self.best_block_size(after - before, len(data_block))
694 speed_str = self.calc_speed(start, time.time(), byte_counter - resume_len)
696 self.report_progress('Unknown %', data_len_str, speed_str, 'Unknown ETA')
698 percent_str = self.calc_percent(byte_counter, data_len)
699 eta_str = self.calc_eta(start, time.time(), data_len - resume_len, byte_counter - resume_len)
700 self.report_progress(percent_str, data_len_str, speed_str, eta_str)
703 self.slow_down(start, byte_counter - resume_len)
706 self.trouble(u'\nERROR: Did not get any data blocks')
710 if data_len is not None and byte_counter != data_len:
711 raise ContentTooShortError(byte_counter, int(data_len))
712 self.try_rename(tmpfilename, filename)
714 # Update file modification time
715 if self.params.get('updatetime', True):
716 info_dict['filetime'] = self.try_utime(filename, data.info().get('last-modified', None))