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 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)]
98 def format_bytes(bytes):
101 if type(bytes) is str:
106 exponent = long(math.log(bytes, 1024.0))
107 suffix = 'bkMGTPEZY'[exponent]
108 converted = float(bytes) / float(1024 ** exponent)
109 return '%.2f%s' % (converted, suffix)
112 def calc_percent(byte_counter, data_len):
115 return '%6s' % ('%3.1f%%' % (float(byte_counter) / float(data_len) * 100.0))
118 def calc_eta(start, now, total, current):
122 if current == 0 or dif < 0.001: # One millisecond
124 rate = float(current) / dif
125 eta = long((float(total) - float(current)) / rate)
126 (eta_mins, eta_secs) = divmod(eta, 60)
129 return '%02d:%02d' % (eta_mins, eta_secs)
132 def calc_speed(start, now, bytes):
134 if bytes == 0 or dif < 0.001: # One millisecond
135 return '%10s' % '---b/s'
136 return '%10s' % ('%s/s' % FileDownloader.format_bytes(float(bytes) / dif))
139 def best_block_size(elapsed_time, bytes):
140 new_min = max(bytes / 2.0, 1.0)
141 new_max = min(max(bytes * 2.0, 1.0), 4194304) # Do not surpass 4 MB
142 if elapsed_time < 0.001:
144 rate = bytes / elapsed_time
152 def parse_bytes(bytestr):
153 """Parse a string indicating a byte quantity into an integer."""
154 matchobj = re.match(r'(?i)^(\d+(?:\.\d+)?)([kMGTPEZY]?)$', bytestr)
157 number = float(matchobj.group(1))
158 multiplier = 1024.0 ** 'bkmgtpezy'.index(matchobj.group(2).lower())
159 return int(round(number * multiplier))
161 def add_info_extractor(self, ie):
162 """Add an InfoExtractor object to the end of the list."""
164 ie.set_downloader(self)
166 def add_post_processor(self, pp):
167 """Add a PostProcessor object to the end of the chain."""
169 pp.set_downloader(self)
171 def to_screen(self, message, skip_eol=False):
172 """Print message to stdout if not in quiet mode."""
173 assert type(message) == type(u'')
174 if not self.params.get('quiet', False):
175 terminator = [u'\n', u''][skip_eol]
176 output = message + terminator
177 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
178 output = output.encode(preferredencoding(), 'ignore')
179 self._screen_file.write(output)
180 self._screen_file.flush()
182 def to_stderr(self, message):
183 """Print message to stderr."""
184 assert type(message) == type(u'')
185 sys.stderr.write((message + u'\n').encode(preferredencoding()))
187 def to_cons_title(self, message):
188 """Set console/terminal window title to message."""
189 if not self.params.get('consoletitle', False):
191 if os.name == 'nt' and ctypes.windll.kernel32.GetConsoleWindow():
192 # c_wchar_p() might not be necessary if `message` is
193 # already of type unicode()
194 ctypes.windll.kernel32.SetConsoleTitleW(ctypes.c_wchar_p(message))
195 elif 'TERM' in os.environ:
196 sys.stderr.write('\033]0;%s\007' % message.encode(preferredencoding()))
198 def fixed_template(self):
199 """Checks if the output template is fixed."""
200 return (re.search(ur'(?u)%\(.+?\)s', self.params['outtmpl']) is None)
202 def trouble(self, message=None):
203 """Determine action to take when a download problem appears.
205 Depending on if the downloader has been configured to ignore
206 download errors or not, this method may throw an exception or
207 not when errors are found, after printing the message.
209 if message is not None:
210 self.to_stderr(message)
211 if not self.params.get('ignoreerrors', False):
212 raise DownloadError(message)
213 self._download_retcode = 1
215 def slow_down(self, start_time, byte_counter):
216 """Sleep if the download speed is over the rate limit."""
217 rate_limit = self.params.get('ratelimit', None)
218 if rate_limit is None or byte_counter == 0:
221 elapsed = now - start_time
224 speed = float(byte_counter) / elapsed
225 if speed > rate_limit:
226 time.sleep((byte_counter - rate_limit * (now - start_time)) / rate_limit)
228 def temp_name(self, filename):
229 """Returns a temporary filename for the given filename."""
230 if self.params.get('nopart', False) or filename == u'-' or \
231 (os.path.exists(encodeFilename(filename)) and not os.path.isfile(encodeFilename(filename))):
233 return filename + u'.part'
235 def undo_temp_name(self, filename):
236 if filename.endswith(u'.part'):
237 return filename[:-len(u'.part')]
240 def try_rename(self, old_filename, new_filename):
242 if old_filename == new_filename:
244 os.rename(encodeFilename(old_filename), encodeFilename(new_filename))
245 except (IOError, OSError), err:
246 self.trouble(u'ERROR: unable to rename file')
248 def try_utime(self, filename, last_modified_hdr):
249 """Try to set the last-modified time of the given file."""
250 if last_modified_hdr is None:
252 if not os.path.isfile(encodeFilename(filename)):
254 timestr = last_modified_hdr
257 filetime = timeconvert(timestr)
261 os.utime(filename, (time.time(), filetime))
266 def report_writedescription(self, descfn):
267 """ Report that the description file is being written """
268 self.to_screen(u'[info] Writing video description to: ' + descfn)
270 def report_writesubtitles(self, srtfn):
271 """ Report that the subtitles file is being written """
272 self.to_screen(u'[info] Writing video subtitles to: ' + srtfn)
274 def report_writeinfojson(self, infofn):
275 """ Report that the metadata file has been written """
276 self.to_screen(u'[info] Video description metadata as JSON to: ' + infofn)
278 def report_destination(self, filename):
279 """Report destination filename."""
280 self.to_screen(u'[download] Destination: ' + filename)
282 def report_progress(self, percent_str, data_len_str, speed_str, eta_str):
283 """Report download progress."""
284 if self.params.get('noprogress', False):
286 self.to_screen(u'\r[download] %s of %s at %s ETA %s' %
287 (percent_str, data_len_str, speed_str, eta_str), skip_eol=True)
288 self.to_cons_title(u'youtube-dl - %s of %s at %s ETA %s' %
289 (percent_str.strip(), data_len_str.strip(), speed_str.strip(), eta_str.strip()))
291 def report_resuming_byte(self, resume_len):
292 """Report attempt to resume at given byte."""
293 self.to_screen(u'[download] Resuming download at byte %s' % resume_len)
295 def report_retry(self, count, retries):
296 """Report retry in case of HTTP error 5xx"""
297 self.to_screen(u'[download] Got server HTTP error. Retrying (attempt %d of %d)...' % (count, retries))
299 def report_file_already_downloaded(self, file_name):
300 """Report file has already been fully downloaded."""
302 self.to_screen(u'[download] %s has already been downloaded' % file_name)
303 except (UnicodeEncodeError), err:
304 self.to_screen(u'[download] The file has already been downloaded')
306 def report_unable_to_resume(self):
307 """Report it was impossible to resume download."""
308 self.to_screen(u'[download] Unable to resume')
310 def report_finish(self):
311 """Report download finished."""
312 if self.params.get('noprogress', False):
313 self.to_screen(u'[download] Download completed')
317 def increment_downloads(self):
318 """Increment the ordinal that assigns a number to each file."""
319 self._num_downloads += 1
321 def prepare_filename(self, info_dict):
322 """Generate the output filename."""
324 template_dict = dict(info_dict)
325 template_dict['epoch'] = unicode(long(time.time()))
326 template_dict['autonumber'] = unicode('%05d' % self._num_downloads)
327 template_dict['title'] = template_dict['stitle'] # Keep both for backwards compatibility
328 filename = self.params['outtmpl'] % template_dict
330 except (ValueError, KeyError), err:
331 self.trouble(u'ERROR: invalid system charset or erroneous output template')
334 def _match_entry(self, info_dict):
335 """ Returns None iff the file should be downloaded """
337 title = info_dict['title']
338 matchtitle = self.params.get('matchtitle', False)
340 matchtitle = matchtitle.decode('utf8')
341 if not re.search(matchtitle, title, re.IGNORECASE):
342 return u'[download] "' + title + '" title did not match pattern "' + matchtitle + '"'
343 rejecttitle = self.params.get('rejecttitle', False)
345 rejecttitle = rejecttitle.decode('utf8')
346 if re.search(rejecttitle, title, re.IGNORECASE):
347 return u'"' + title + '" title matched reject pattern "' + rejecttitle + '"'
350 def process_info(self, info_dict):
351 """Process a single dictionary returned by an InfoExtractor."""
353 info_dict['stitle'] = sanitize_filename(info_dict['title'], self.params.get('restrictfilenames'))
355 reason = self._match_entry(info_dict)
356 if reason is not None:
357 self.to_screen(u'[download] ' + reason)
360 max_downloads = self.params.get('max_downloads')
361 if max_downloads is not None:
362 if self._num_downloads > int(max_downloads):
363 raise MaxDownloadsReached()
365 filename = self.prepare_filename(info_dict)
368 if self.params.get('forcetitle', False):
369 print(info_dict['title'].encode(preferredencoding(), 'xmlcharrefreplace'))
370 if self.params.get('forceurl', False):
371 print(info_dict['url'].encode(preferredencoding(), 'xmlcharrefreplace'))
372 if self.params.get('forcethumbnail', False) and 'thumbnail' in info_dict:
373 print(info_dict['thumbnail'].encode(preferredencoding(), 'xmlcharrefreplace'))
374 if self.params.get('forcedescription', False) and 'description' in info_dict:
375 print(info_dict['description'].encode(preferredencoding(), 'xmlcharrefreplace'))
376 if self.params.get('forcefilename', False) and filename is not None:
377 print(filename.encode(preferredencoding(), 'xmlcharrefreplace'))
378 if self.params.get('forceformat', False):
379 print(info_dict['format'].encode(preferredencoding(), 'xmlcharrefreplace'))
381 # Do nothing else if in simulate mode
382 if self.params.get('simulate', False):
389 dn = os.path.dirname(encodeFilename(filename))
390 if dn != '' and not os.path.exists(dn): # dn is already encoded
392 except (OSError, IOError), err:
393 self.trouble(u'ERROR: unable to create directory ' + unicode(err))
396 if self.params.get('writedescription', False):
398 descfn = filename + u'.description'
399 self.report_writedescription(descfn)
400 descfile = open(encodeFilename(descfn), 'wb')
402 descfile.write(info_dict['description'].encode('utf-8'))
405 except (OSError, IOError):
406 self.trouble(u'ERROR: Cannot write description file ' + descfn)
409 if self.params.get('writesubtitles', False) and 'subtitles' in info_dict and info_dict['subtitles']:
410 # subtitles download errors are already managed as troubles in relevant IE
411 # that way it will silently go on when used with unsupporting IE
413 srtfn = filename.rsplit('.', 1)[0] + u'.srt'
414 self.report_writesubtitles(srtfn)
415 srtfile = open(encodeFilename(srtfn), 'wb')
417 srtfile.write(info_dict['subtitles'].encode('utf-8'))
420 except (OSError, IOError):
421 self.trouble(u'ERROR: Cannot write subtitles file ' + descfn)
424 if self.params.get('writeinfojson', False):
425 infofn = filename + u'.info.json'
426 self.report_writeinfojson(infofn)
429 except (NameError,AttributeError):
430 self.trouble(u'ERROR: No JSON encoder found. Update to Python 2.6+, setup a json module, or leave out --write-info-json.')
433 infof = open(encodeFilename(infofn), 'wb')
435 json_info_dict = dict((k,v) for k,v in info_dict.iteritems() if not k in ('urlhandle',))
436 json.dump(json_info_dict, infof)
439 except (OSError, IOError):
440 self.trouble(u'ERROR: Cannot write metadata to JSON file ' + infofn)
443 if not self.params.get('skip_download', False):
444 if self.params.get('nooverwrites', False) and os.path.exists(encodeFilename(filename)):
448 success = self._do_download(filename, info_dict)
449 except (OSError, IOError), err:
450 raise UnavailableVideoError
451 except (urllib2.URLError, httplib.HTTPException, socket.error), err:
452 self.trouble(u'ERROR: unable to download video data: %s' % str(err))
454 except (ContentTooShortError, ), err:
455 self.trouble(u'ERROR: content too short (expected %s bytes and served %s)' % (err.expected, err.downloaded))
460 self.post_process(filename, info_dict)
461 except (PostProcessingError), err:
462 self.trouble(u'ERROR: postprocessing: %s' % str(err))
465 def download(self, url_list):
466 """Download a given list of URLs."""
467 if len(url_list) > 1 and self.fixed_template():
468 raise SameFileError(self.params['outtmpl'])
471 suitable_found = False
473 # Go to next InfoExtractor if not suitable
474 if not ie.suitable(url):
477 # Suitable InfoExtractor found
478 suitable_found = True
480 # Extract information from URL and process it
481 videos = ie.extract(url)
482 for video in videos or []:
483 video['extractor'] = ie.IE_NAME
485 self.increment_downloads()
486 self.process_info(video)
487 except UnavailableVideoError:
488 self.trouble(u'\nERROR: unable to download video')
490 # Suitable InfoExtractor had been found; go to next URL
493 if not suitable_found:
494 self.trouble(u'ERROR: no suitable InfoExtractor: %s' % url)
496 return self._download_retcode
498 def post_process(self, filename, ie_info):
499 """Run the postprocessing chain on the given file."""
501 info['filepath'] = filename
507 def _download_with_rtmpdump(self, filename, url, player_url):
508 self.report_destination(filename)
509 tmpfilename = self.temp_name(filename)
511 # Check for rtmpdump first
513 subprocess.call(['rtmpdump', '-h'], stdout=(file(os.path.devnull, 'w')), stderr=subprocess.STDOUT)
514 except (OSError, IOError):
515 self.trouble(u'ERROR: RTMP download detected but "rtmpdump" could not be run')
518 # Download using rtmpdump. rtmpdump returns exit code 2 when
519 # the connection was interrumpted and resuming appears to be
520 # possible. This is part of rtmpdump's normal usage, AFAIK.
521 basic_args = ['rtmpdump', '-q'] + [[], ['-W', player_url]][player_url is not None] + ['-r', url, '-o', tmpfilename]
522 args = basic_args + [[], ['-e', '-k', '1']][self.params.get('continuedl', False)]
523 if self.params.get('verbose', False):
526 shell_quote = lambda args: ' '.join(map(pipes.quote, args))
529 self.to_screen(u'[debug] rtmpdump command line: ' + shell_quote(args))
530 retval = subprocess.call(args)
531 while retval == 2 or retval == 1:
532 prevsize = os.path.getsize(encodeFilename(tmpfilename))
533 self.to_screen(u'\r[rtmpdump] %s bytes' % prevsize, skip_eol=True)
534 time.sleep(5.0) # This seems to be needed
535 retval = subprocess.call(basic_args + ['-e'] + [[], ['-k', '1']][retval == 1])
536 cursize = os.path.getsize(encodeFilename(tmpfilename))
537 if prevsize == cursize and retval == 1:
539 # Some rtmp streams seem abort after ~ 99.8%. Don't complain for those
540 if prevsize == cursize and retval == 2 and cursize > 1024:
541 self.to_screen(u'\r[rtmpdump] Could not download the whole video. This can happen for some advertisements.')
545 self.to_screen(u'\r[rtmpdump] %s bytes' % os.path.getsize(encodeFilename(tmpfilename)))
546 self.try_rename(tmpfilename, filename)
549 self.trouble(u'\nERROR: rtmpdump exited with code %d' % retval)
552 def _do_download(self, filename, info_dict):
553 url = info_dict['url']
554 player_url = info_dict.get('player_url', None)
556 # Check file already present
557 if self.params.get('continuedl', False) and os.path.isfile(encodeFilename(filename)) and not self.params.get('nopart', False):
558 self.report_file_already_downloaded(filename)
561 # Attempt to download using rtmpdump
562 if url.startswith('rtmp'):
563 return self._download_with_rtmpdump(filename, url, player_url)
565 tmpfilename = self.temp_name(filename)
568 # Do not include the Accept-Encoding header
569 headers = {'Youtubedl-no-compression': 'True'}
570 basic_request = urllib2.Request(url, None, headers)
571 request = urllib2.Request(url, None, headers)
573 # Establish possible resume length
574 if os.path.isfile(encodeFilename(tmpfilename)):
575 resume_len = os.path.getsize(encodeFilename(tmpfilename))
581 if self.params.get('continuedl', False):
582 self.report_resuming_byte(resume_len)
583 request.add_header('Range','bytes=%d-' % resume_len)
589 retries = self.params.get('retries', 0)
590 while count <= retries:
591 # Establish connection
593 if count == 0 and 'urlhandle' in info_dict:
594 data = info_dict['urlhandle']
595 data = urllib2.urlopen(request)
597 except (urllib2.HTTPError, ), err:
598 if (err.code < 500 or err.code >= 600) and err.code != 416:
599 # Unexpected HTTP error
601 elif err.code == 416:
602 # Unable to resume (requested range not satisfiable)
604 # Open the connection again without the range header
605 data = urllib2.urlopen(basic_request)
606 content_length = data.info()['Content-Length']
607 except (urllib2.HTTPError, ), err:
608 if err.code < 500 or err.code >= 600:
611 # Examine the reported length
612 if (content_length is not None and
613 (resume_len - 100 < long(content_length) < resume_len + 100)):
614 # The file had already been fully downloaded.
615 # Explanation to the above condition: in issue #175 it was revealed that
616 # YouTube sometimes adds or removes a few bytes from the end of the file,
617 # changing the file size slightly and causing problems for some users. So
618 # I decided to implement a suggested change and consider the file
619 # completely downloaded if the file size differs less than 100 bytes from
620 # the one in the hard drive.
621 self.report_file_already_downloaded(filename)
622 self.try_rename(tmpfilename, filename)
625 # The length does not match, we start the download over
626 self.report_unable_to_resume()
632 self.report_retry(count, retries)
635 self.trouble(u'ERROR: giving up after %s retries' % retries)
638 data_len = data.info().get('Content-length', None)
639 if data_len is not None:
640 data_len = long(data_len) + resume_len
641 data_len_str = self.format_bytes(data_len)
642 byte_counter = 0 + resume_len
648 data_block = data.read(block_size)
650 if len(data_block) == 0:
652 byte_counter += len(data_block)
654 # Open file just in time
657 (stream, tmpfilename) = sanitize_open(tmpfilename, open_mode)
658 assert stream is not None
659 filename = self.undo_temp_name(tmpfilename)
660 self.report_destination(filename)
661 except (OSError, IOError), err:
662 self.trouble(u'ERROR: unable to open for writing: %s' % str(err))
665 stream.write(data_block)
666 except (IOError, OSError), err:
667 self.trouble(u'\nERROR: unable to write data: %s' % str(err))
669 block_size = self.best_block_size(after - before, len(data_block))
672 speed_str = self.calc_speed(start, time.time(), byte_counter - resume_len)
674 self.report_progress('Unknown %', data_len_str, speed_str, 'Unknown ETA')
676 percent_str = self.calc_percent(byte_counter, data_len)
677 eta_str = self.calc_eta(start, time.time(), data_len - resume_len, byte_counter - resume_len)
678 self.report_progress(percent_str, data_len_str, speed_str, eta_str)
681 self.slow_down(start, byte_counter - resume_len)
684 self.trouble(u'\nERROR: Did not get any data blocks')
688 if data_len is not None and byte_counter != data_len:
689 raise ContentTooShortError(byte_counter, long(data_len))
690 self.try_rename(tmpfilename, filename)
692 # Update file modification time
693 if self.params.get('updatetime', True):
694 info_dict['filetime'] = self.try_utime(filename, data.info().get('last-modified', None))