2 # -*- coding: utf-8 -*-
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 = long(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 = long((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' not in 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 sys.stderr.write((message + u'\n').encode(preferredencoding()))
192 def to_cons_title(self, message):
193 """Set console/terminal window title to message."""
194 if not self.params.get('consoletitle', False):
196 if os.name == 'nt' and ctypes.windll.kernel32.GetConsoleWindow():
197 # c_wchar_p() might not be necessary if `message` is
198 # already of type unicode()
199 ctypes.windll.kernel32.SetConsoleTitleW(ctypes.c_wchar_p(message))
200 elif 'TERM' in os.environ:
201 sys.stderr.write('\033]0;%s\007' % message.encode(preferredencoding()))
203 def fixed_template(self):
204 """Checks if the output template is fixed."""
205 return (re.search(ur'(?u)%\(.+?\)s', self.params['outtmpl']) is None)
207 def trouble(self, message=None):
208 """Determine action to take when a download problem appears.
210 Depending on if the downloader has been configured to ignore
211 download errors or not, this method may throw an exception or
212 not when errors are found, after printing the message.
214 if message is not None:
215 self.to_stderr(message)
216 if not self.params.get('ignoreerrors', False):
217 raise DownloadError(message)
218 self._download_retcode = 1
220 def slow_down(self, start_time, byte_counter):
221 """Sleep if the download speed is over the rate limit."""
222 rate_limit = self.params.get('ratelimit', None)
223 if rate_limit is None or byte_counter == 0:
226 elapsed = now - start_time
229 speed = float(byte_counter) / elapsed
230 if speed > rate_limit:
231 time.sleep((byte_counter - rate_limit * (now - start_time)) / rate_limit)
233 def temp_name(self, filename):
234 """Returns a temporary filename for the given filename."""
235 if self.params.get('nopart', False) or filename == u'-' or \
236 (os.path.exists(encodeFilename(filename)) and not os.path.isfile(encodeFilename(filename))):
238 return filename + u'.part'
240 def undo_temp_name(self, filename):
241 if filename.endswith(u'.part'):
242 return filename[:-len(u'.part')]
245 def try_rename(self, old_filename, new_filename):
247 if old_filename == new_filename:
249 os.rename(encodeFilename(old_filename), encodeFilename(new_filename))
250 except (IOError, OSError), err:
251 self.trouble(u'ERROR: unable to rename file')
253 def try_utime(self, filename, last_modified_hdr):
254 """Try to set the last-modified time of the given file."""
255 if last_modified_hdr is None:
257 if not os.path.isfile(encodeFilename(filename)):
259 timestr = last_modified_hdr
262 filetime = timeconvert(timestr)
266 os.utime(filename, (time.time(), filetime))
271 def report_writedescription(self, descfn):
272 """ Report that the description file is being written """
273 self.to_screen(u'[info] Writing video description to: ' + descfn)
275 def report_writesubtitles(self, srtfn):
276 """ Report that the subtitles file is being written """
277 self.to_screen(u'[info] Writing video subtitles to: ' + srtfn)
279 def report_writeinfojson(self, infofn):
280 """ Report that the metadata file has been written """
281 self.to_screen(u'[info] Video description metadata as JSON to: ' + infofn)
283 def report_destination(self, filename):
284 """Report destination filename."""
285 self.to_screen(u'[download] Destination: ' + filename)
287 def report_progress(self, percent_str, data_len_str, speed_str, eta_str):
288 """Report download progress."""
289 if self.params.get('noprogress', False):
291 self.to_screen(u'\r[download] %s of %s at %s ETA %s' %
292 (percent_str, data_len_str, speed_str, eta_str), skip_eol=True)
293 self.to_cons_title(u'youtube-dl - %s of %s at %s ETA %s' %
294 (percent_str.strip(), data_len_str.strip(), speed_str.strip(), eta_str.strip()))
296 def report_resuming_byte(self, resume_len):
297 """Report attempt to resume at given byte."""
298 self.to_screen(u'[download] Resuming download at byte %s' % resume_len)
300 def report_retry(self, count, retries):
301 """Report retry in case of HTTP error 5xx"""
302 self.to_screen(u'[download] Got server HTTP error. Retrying (attempt %d of %d)...' % (count, retries))
304 def report_file_already_downloaded(self, file_name):
305 """Report file has already been fully downloaded."""
307 self.to_screen(u'[download] %s has already been downloaded' % file_name)
308 except (UnicodeEncodeError), err:
309 self.to_screen(u'[download] The file has already been downloaded')
311 def report_unable_to_resume(self):
312 """Report it was impossible to resume download."""
313 self.to_screen(u'[download] Unable to resume')
315 def report_finish(self):
316 """Report download finished."""
317 if self.params.get('noprogress', False):
318 self.to_screen(u'[download] Download completed')
322 def increment_downloads(self):
323 """Increment the ordinal that assigns a number to each file."""
324 self._num_downloads += 1
326 def prepare_filename(self, info_dict):
327 """Generate the output filename."""
329 template_dict = dict(info_dict)
330 template_dict['epoch'] = unicode(int(time.time()))
331 template_dict['autonumber'] = unicode('%05d' % self._num_downloads)
332 filename = self.params['outtmpl'] % template_dict
334 except (ValueError, KeyError), err:
335 self.trouble(u'ERROR: invalid system charset or erroneous output template')
338 def _match_entry(self, info_dict):
339 """ Returns None iff the file should be downloaded """
341 title = info_dict['title']
342 matchtitle = self.params.get('matchtitle', False)
344 matchtitle = matchtitle.decode('utf8')
345 if not re.search(matchtitle, title, re.IGNORECASE):
346 return u'[download] "' + title + '" title did not match pattern "' + matchtitle + '"'
347 rejecttitle = self.params.get('rejecttitle', False)
349 rejecttitle = rejecttitle.decode('utf8')
350 if re.search(rejecttitle, title, re.IGNORECASE):
351 return u'"' + title + '" title matched reject pattern "' + rejecttitle + '"'
354 def process_info(self, info_dict):
355 """Process a single dictionary returned by an InfoExtractor."""
357 # Keep for backwards compatibility
358 info_dict['stitle'] = info_dict['title']
360 reason = self._match_entry(info_dict)
361 if reason is not None:
362 self.to_screen(u'[download] ' + reason)
365 max_downloads = self.params.get('max_downloads')
366 if max_downloads is not None:
367 if self._num_downloads > int(max_downloads):
368 raise MaxDownloadsReached()
370 filename = self.prepare_filename(info_dict)
371 filename = sanitize_filename(filename, self.params.get('restrictfilenames'))
374 if self.params.get('forcetitle', False):
375 print(info_dict['title'].encode(preferredencoding(), 'xmlcharrefreplace'))
376 if self.params.get('forceurl', False):
377 print(info_dict['url'].encode(preferredencoding(), 'xmlcharrefreplace'))
378 if self.params.get('forcethumbnail', False) and 'thumbnail' in info_dict:
379 print(info_dict['thumbnail'].encode(preferredencoding(), 'xmlcharrefreplace'))
380 if self.params.get('forcedescription', False) and 'description' in info_dict:
381 print(info_dict['description'].encode(preferredencoding(), 'xmlcharrefreplace'))
382 if self.params.get('forcefilename', False) and filename is not None:
383 print(filename.encode(preferredencoding(), 'xmlcharrefreplace'))
384 if self.params.get('forceformat', False):
385 print(info_dict['format'].encode(preferredencoding(), 'xmlcharrefreplace'))
387 # Do nothing else if in simulate mode
388 if self.params.get('simulate', False):
395 dn = os.path.dirname(encodeFilename(filename))
396 if dn != '' and not os.path.exists(dn): # dn is already encoded
398 except (OSError, IOError), err:
399 self.trouble(u'ERROR: unable to create directory ' + unicode(err))
402 if self.params.get('writedescription', False):
404 descfn = filename + u'.description'
405 self.report_writedescription(descfn)
406 descfile = open(encodeFilename(descfn), 'wb')
408 descfile.write(info_dict['description'].encode('utf-8'))
411 except (OSError, IOError):
412 self.trouble(u'ERROR: Cannot write description file ' + descfn)
415 if self.params.get('writesubtitles', False) and 'subtitles' in info_dict and info_dict['subtitles']:
416 # subtitles download errors are already managed as troubles in relevant IE
417 # that way it will silently go on when used with unsupporting IE
419 srtfn = filename.rsplit('.', 1)[0] + u'.srt'
420 self.report_writesubtitles(srtfn)
421 srtfile = open(encodeFilename(srtfn), 'wb')
423 srtfile.write(info_dict['subtitles'].encode('utf-8'))
426 except (OSError, IOError):
427 self.trouble(u'ERROR: Cannot write subtitles file ' + descfn)
430 if self.params.get('writeinfojson', False):
431 infofn = filename + u'.info.json'
432 self.report_writeinfojson(infofn)
435 except (NameError,AttributeError):
436 self.trouble(u'ERROR: No JSON encoder found. Update to Python 2.6+, setup a json module, or leave out --write-info-json.')
439 infof = open(encodeFilename(infofn), 'wb')
441 json_info_dict = dict((k,v) for k,v in info_dict.iteritems() if not k in ('urlhandle',))
442 json.dump(json_info_dict, infof)
445 except (OSError, IOError):
446 self.trouble(u'ERROR: Cannot write metadata to JSON file ' + infofn)
449 if not self.params.get('skip_download', False):
450 if self.params.get('nooverwrites', False) and os.path.exists(encodeFilename(filename)):
454 success = self._do_download(filename, info_dict)
455 except (OSError, IOError), err:
456 raise UnavailableVideoError
457 except (urllib2.URLError, httplib.HTTPException, socket.error), err:
458 self.trouble(u'ERROR: unable to download video data: %s' % str(err))
460 except (ContentTooShortError, ), err:
461 self.trouble(u'ERROR: content too short (expected %s bytes and served %s)' % (err.expected, err.downloaded))
466 self.post_process(filename, info_dict)
467 except (PostProcessingError), err:
468 self.trouble(u'ERROR: postprocessing: %s' % str(err))
471 def download(self, url_list):
472 """Download a given list of URLs."""
473 if len(url_list) > 1 and self.fixed_template():
474 raise SameFileError(self.params['outtmpl'])
477 suitable_found = False
479 # Go to next InfoExtractor if not suitable
480 if not ie.suitable(url):
483 # Suitable InfoExtractor found
484 suitable_found = True
486 # Extract information from URL and process it
487 videos = ie.extract(url)
488 for video in videos or []:
489 video['extractor'] = ie.IE_NAME
491 self.increment_downloads()
492 self.process_info(video)
493 except UnavailableVideoError:
494 self.trouble(u'\nERROR: unable to download video')
496 # Suitable InfoExtractor had been found; go to next URL
499 if not suitable_found:
500 self.trouble(u'ERROR: no suitable InfoExtractor: %s' % url)
502 return self._download_retcode
504 def post_process(self, filename, ie_info):
505 """Run the postprocessing chain on the given file."""
507 info['filepath'] = filename
513 def _download_with_rtmpdump(self, filename, url, player_url):
514 self.report_destination(filename)
515 tmpfilename = self.temp_name(filename)
517 # Check for rtmpdump first
519 subprocess.call(['rtmpdump', '-h'], stdout=(file(os.path.devnull, 'w')), stderr=subprocess.STDOUT)
520 except (OSError, IOError):
521 self.trouble(u'ERROR: RTMP download detected but "rtmpdump" could not be run')
524 # Download using rtmpdump. rtmpdump returns exit code 2 when
525 # the connection was interrumpted and resuming appears to be
526 # possible. This is part of rtmpdump's normal usage, AFAIK.
527 basic_args = ['rtmpdump', '-q'] + [[], ['-W', player_url]][player_url is not None] + ['-r', url, '-o', tmpfilename]
528 args = basic_args + [[], ['-e', '-k', '1']][self.params.get('continuedl', False)]
529 if self.params.get('verbose', False):
532 shell_quote = lambda args: ' '.join(map(pipes.quote, args))
535 self.to_screen(u'[debug] rtmpdump command line: ' + shell_quote(args))
536 retval = subprocess.call(args)
537 while retval == 2 or retval == 1:
538 prevsize = os.path.getsize(encodeFilename(tmpfilename))
539 self.to_screen(u'\r[rtmpdump] %s bytes' % prevsize, skip_eol=True)
540 time.sleep(5.0) # This seems to be needed
541 retval = subprocess.call(basic_args + ['-e'] + [[], ['-k', '1']][retval == 1])
542 cursize = os.path.getsize(encodeFilename(tmpfilename))
543 if prevsize == cursize and retval == 1:
545 # Some rtmp streams seem abort after ~ 99.8%. Don't complain for those
546 if prevsize == cursize and retval == 2 and cursize > 1024:
547 self.to_screen(u'\r[rtmpdump] Could not download the whole video. This can happen for some advertisements.')
551 self.to_screen(u'\r[rtmpdump] %s bytes' % os.path.getsize(encodeFilename(tmpfilename)))
552 self.try_rename(tmpfilename, filename)
555 self.trouble(u'\nERROR: rtmpdump exited with code %d' % retval)
558 def _do_download(self, filename, info_dict):
559 url = info_dict['url']
560 player_url = info_dict.get('player_url', None)
562 # Check file already present
563 if self.params.get('continuedl', False) and os.path.isfile(encodeFilename(filename)) and not self.params.get('nopart', False):
564 self.report_file_already_downloaded(filename)
567 # Attempt to download using rtmpdump
568 if url.startswith('rtmp'):
569 return self._download_with_rtmpdump(filename, url, player_url)
571 tmpfilename = self.temp_name(filename)
574 # Do not include the Accept-Encoding header
575 headers = {'Youtubedl-no-compression': 'True'}
576 basic_request = urllib2.Request(url, None, headers)
577 request = urllib2.Request(url, None, headers)
579 # Establish possible resume length
580 if os.path.isfile(encodeFilename(tmpfilename)):
581 resume_len = os.path.getsize(encodeFilename(tmpfilename))
587 if self.params.get('continuedl', False):
588 self.report_resuming_byte(resume_len)
589 request.add_header('Range','bytes=%d-' % resume_len)
595 retries = self.params.get('retries', 0)
596 while count <= retries:
597 # Establish connection
599 if count == 0 and 'urlhandle' in info_dict:
600 data = info_dict['urlhandle']
601 data = urllib2.urlopen(request)
603 except (urllib2.HTTPError, ), err:
604 if (err.code < 500 or err.code >= 600) and err.code != 416:
605 # Unexpected HTTP error
607 elif err.code == 416:
608 # Unable to resume (requested range not satisfiable)
610 # Open the connection again without the range header
611 data = urllib2.urlopen(basic_request)
612 content_length = data.info()['Content-Length']
613 except (urllib2.HTTPError, ), err:
614 if err.code < 500 or err.code >= 600:
617 # Examine the reported length
618 if (content_length is not None and
619 (resume_len - 100 < long(content_length) < resume_len + 100)):
620 # The file had already been fully downloaded.
621 # Explanation to the above condition: in issue #175 it was revealed that
622 # YouTube sometimes adds or removes a few bytes from the end of the file,
623 # changing the file size slightly and causing problems for some users. So
624 # I decided to implement a suggested change and consider the file
625 # completely downloaded if the file size differs less than 100 bytes from
626 # the one in the hard drive.
627 self.report_file_already_downloaded(filename)
628 self.try_rename(tmpfilename, filename)
631 # The length does not match, we start the download over
632 self.report_unable_to_resume()
638 self.report_retry(count, retries)
641 self.trouble(u'ERROR: giving up after %s retries' % retries)
644 data_len = data.info().get('Content-length', None)
645 if data_len is not None:
646 data_len = long(data_len) + resume_len
647 data_len_str = self.format_bytes(data_len)
648 byte_counter = 0 + resume_len
649 block_size = self.params.get('buffersize', 1024)
654 data_block = data.read(block_size)
656 if len(data_block) == 0:
658 byte_counter += len(data_block)
660 # Open file just in time
663 (stream, tmpfilename) = sanitize_open(tmpfilename, open_mode)
664 assert stream is not None
665 filename = self.undo_temp_name(tmpfilename)
666 self.report_destination(filename)
667 except (OSError, IOError), err:
668 self.trouble(u'ERROR: unable to open for writing: %s' % str(err))
671 stream.write(data_block)
672 except (IOError, OSError), err:
673 self.trouble(u'\nERROR: unable to write data: %s' % str(err))
675 if not self.params.get('noresizebuffer', False):
676 block_size = self.best_block_size(after - before, len(data_block))
679 speed_str = self.calc_speed(start, time.time(), byte_counter - resume_len)
681 self.report_progress('Unknown %', data_len_str, speed_str, 'Unknown ETA')
683 percent_str = self.calc_percent(byte_counter, data_len)
684 eta_str = self.calc_eta(start, time.time(), data_len - resume_len, byte_counter - resume_len)
685 self.report_progress(percent_str, data_len_str, speed_str, eta_str)
688 self.slow_down(start, byte_counter - resume_len)
691 self.trouble(u'\nERROR: Did not get any data blocks')
695 if data_len is not None and byte_counter != data_len:
696 raise ContentTooShortError(byte_counter, long(data_len))
697 self.try_rename(tmpfilename, filename)
699 # Update file modification time
700 if self.params.get('updatetime', True):
701 info_dict['filetime'] = self.try_utime(filename, data.info().get('last-modified', None))