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 ignoreerrors: Do not stop on download errors.
61 ratelimit: Download speed limit, in bytes/sec.
62 nooverwrites: Prevent overwriting files.
63 retries: Number of times to retry for HTTP error 5xx
64 continuedl: Try to continue downloads if possible.
65 noprogress: Do not print the progress bar.
66 playliststart: Playlist item to start at.
67 playlistend: Playlist item to end at.
68 matchtitle: Download only matching titles.
69 rejecttitle: Reject downloads for matching titles.
70 logtostderr: Log messages to stderr instead of stdout.
71 consoletitle: Display progress in console window's titlebar.
72 nopart: Do not use temporary .part files.
73 updatetime: Use the Last-modified header to set output file timestamps.
74 writedescription: Write the video description to a .description file
75 writeinfojson: Write the video description to a .info.json file
76 writesubtitles: Write the video subtitles to a .srt file
77 subtitleslang: Language of the subtitles to download
83 _download_retcode = None
87 def __init__(self, params):
88 """Create a FileDownloader object with the given options."""
91 self._download_retcode = 0
92 self._num_downloads = 0
93 self._screen_file = [sys.stdout, sys.stderr][params.get('logtostderr', False)]
97 def format_bytes(bytes):
100 if type(bytes) is str:
105 exponent = long(math.log(bytes, 1024.0))
106 suffix = 'bkMGTPEZY'[exponent]
107 converted = float(bytes) / float(1024 ** exponent)
108 return '%.2f%s' % (converted, suffix)
111 def calc_percent(byte_counter, data_len):
114 return '%6s' % ('%3.1f%%' % (float(byte_counter) / float(data_len) * 100.0))
117 def calc_eta(start, now, total, current):
121 if current == 0 or dif < 0.001: # One millisecond
123 rate = float(current) / dif
124 eta = long((float(total) - float(current)) / rate)
125 (eta_mins, eta_secs) = divmod(eta, 60)
128 return '%02d:%02d' % (eta_mins, eta_secs)
131 def calc_speed(start, now, bytes):
133 if bytes == 0 or dif < 0.001: # One millisecond
134 return '%10s' % '---b/s'
135 return '%10s' % ('%s/s' % FileDownloader.format_bytes(float(bytes) / dif))
138 def best_block_size(elapsed_time, bytes):
139 new_min = max(bytes / 2.0, 1.0)
140 new_max = min(max(bytes * 2.0, 1.0), 4194304) # Do not surpass 4 MB
141 if elapsed_time < 0.001:
143 rate = bytes / elapsed_time
151 def parse_bytes(bytestr):
152 """Parse a string indicating a byte quantity into an integer."""
153 matchobj = re.match(r'(?i)^(\d+(?:\.\d+)?)([kMGTPEZY]?)$', bytestr)
156 number = float(matchobj.group(1))
157 multiplier = 1024.0 ** 'bkmgtpezy'.index(matchobj.group(2).lower())
158 return int(round(number * multiplier))
160 def add_info_extractor(self, ie):
161 """Add an InfoExtractor object to the end of the list."""
163 ie.set_downloader(self)
165 def add_post_processor(self, pp):
166 """Add a PostProcessor object to the end of the chain."""
168 pp.set_downloader(self)
170 def to_screen(self, message, skip_eol=False):
171 """Print message to stdout if not in quiet mode."""
172 assert type(message) == type(u'')
173 if not self.params.get('quiet', False):
174 terminator = [u'\n', u''][skip_eol]
175 output = message + terminator
176 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
177 output = output.encode(preferredencoding(), 'ignore')
178 self._screen_file.write(output)
179 self._screen_file.flush()
181 def to_stderr(self, message):
182 """Print message to stderr."""
183 assert type(message) == type(u'')
184 sys.stderr.write((message + u'\n').encode(preferredencoding()))
186 def to_cons_title(self, message):
187 """Set console/terminal window title to message."""
188 if not self.params.get('consoletitle', False):
190 if os.name == 'nt' and ctypes.windll.kernel32.GetConsoleWindow():
191 # c_wchar_p() might not be necessary if `message` is
192 # already of type unicode()
193 ctypes.windll.kernel32.SetConsoleTitleW(ctypes.c_wchar_p(message))
194 elif 'TERM' in os.environ:
195 sys.stderr.write('\033]0;%s\007' % message.encode(preferredencoding()))
197 def fixed_template(self):
198 """Checks if the output template is fixed."""
199 return (re.search(ur'(?u)%\(.+?\)s', self.params['outtmpl']) is None)
201 def trouble(self, message=None):
202 """Determine action to take when a download problem appears.
204 Depending on if the downloader has been configured to ignore
205 download errors or not, this method may throw an exception or
206 not when errors are found, after printing the message.
208 if message is not None:
209 self.to_stderr(message)
210 if not self.params.get('ignoreerrors', False):
211 raise DownloadError(message)
212 self._download_retcode = 1
214 def slow_down(self, start_time, byte_counter):
215 """Sleep if the download speed is over the rate limit."""
216 rate_limit = self.params.get('ratelimit', None)
217 if rate_limit is None or byte_counter == 0:
220 elapsed = now - start_time
223 speed = float(byte_counter) / elapsed
224 if speed > rate_limit:
225 time.sleep((byte_counter - rate_limit * (now - start_time)) / rate_limit)
227 def temp_name(self, filename):
228 """Returns a temporary filename for the given filename."""
229 if self.params.get('nopart', False) or filename == u'-' or \
230 (os.path.exists(encodeFilename(filename)) and not os.path.isfile(encodeFilename(filename))):
232 return filename + u'.part'
234 def undo_temp_name(self, filename):
235 if filename.endswith(u'.part'):
236 return filename[:-len(u'.part')]
239 def try_rename(self, old_filename, new_filename):
241 if old_filename == new_filename:
243 os.rename(encodeFilename(old_filename), encodeFilename(new_filename))
244 except (IOError, OSError), err:
245 self.trouble(u'ERROR: unable to rename file')
247 def try_utime(self, filename, last_modified_hdr):
248 """Try to set the last-modified time of the given file."""
249 if last_modified_hdr is None:
251 if not os.path.isfile(encodeFilename(filename)):
253 timestr = last_modified_hdr
256 filetime = timeconvert(timestr)
260 os.utime(filename, (time.time(), filetime))
265 def report_writedescription(self, descfn):
266 """ Report that the description file is being written """
267 self.to_screen(u'[info] Writing video description to: ' + descfn)
269 def report_writesubtitles(self, srtfn):
270 """ Report that the subtitles file is being written """
271 self.to_screen(u'[info] Writing video subtitles to: ' + srtfn)
273 def report_writeinfojson(self, infofn):
274 """ Report that the metadata file has been written """
275 self.to_screen(u'[info] Video description metadata as JSON to: ' + infofn)
277 def report_destination(self, filename):
278 """Report destination filename."""
279 self.to_screen(u'[download] Destination: ' + filename)
281 def report_progress(self, percent_str, data_len_str, speed_str, eta_str):
282 """Report download progress."""
283 if self.params.get('noprogress', False):
285 self.to_screen(u'\r[download] %s of %s at %s ETA %s' %
286 (percent_str, data_len_str, speed_str, eta_str), skip_eol=True)
287 self.to_cons_title(u'youtube-dl - %s of %s at %s ETA %s' %
288 (percent_str.strip(), data_len_str.strip(), speed_str.strip(), eta_str.strip()))
290 def report_resuming_byte(self, resume_len):
291 """Report attempt to resume at given byte."""
292 self.to_screen(u'[download] Resuming download at byte %s' % resume_len)
294 def report_retry(self, count, retries):
295 """Report retry in case of HTTP error 5xx"""
296 self.to_screen(u'[download] Got server HTTP error. Retrying (attempt %d of %d)...' % (count, retries))
298 def report_file_already_downloaded(self, file_name):
299 """Report file has already been fully downloaded."""
301 self.to_screen(u'[download] %s has already been downloaded' % file_name)
302 except (UnicodeEncodeError), err:
303 self.to_screen(u'[download] The file has already been downloaded')
305 def report_unable_to_resume(self):
306 """Report it was impossible to resume download."""
307 self.to_screen(u'[download] Unable to resume')
309 def report_finish(self):
310 """Report download finished."""
311 if self.params.get('noprogress', False):
312 self.to_screen(u'[download] Download completed')
316 def increment_downloads(self):
317 """Increment the ordinal that assigns a number to each file."""
318 self._num_downloads += 1
320 def prepare_filename(self, info_dict):
321 """Generate the output filename."""
323 template_dict = dict(info_dict)
324 template_dict['epoch'] = unicode(long(time.time()))
325 template_dict['autonumber'] = unicode('%05d' % self._num_downloads)
326 template_dict['title'] = template_dict['stitle'] # Keep both for backwards compatibility
327 filename = self.params['outtmpl'] % template_dict
329 except (ValueError, KeyError), err:
330 self.trouble(u'ERROR: invalid system charset or erroneous output template')
333 def _match_entry(self, info_dict):
334 """ Returns None iff the file should be downloaded """
336 title = info_dict['title']
337 matchtitle = self.params.get('matchtitle', False)
339 matchtitle = matchtitle.decode('utf8')
340 if not re.search(matchtitle, title, re.IGNORECASE):
341 return u'[download] "' + title + '" title did not match pattern "' + matchtitle + '"'
342 rejecttitle = self.params.get('rejecttitle', False)
344 rejecttitle = rejecttitle.decode('utf8')
345 if re.search(rejecttitle, title, re.IGNORECASE):
346 return u'"' + title + '" title matched reject pattern "' + rejecttitle + '"'
349 def process_info(self, info_dict):
350 """Process a single dictionary returned by an InfoExtractor."""
352 info_dict['stitle'] = sanitize_filename(info_dict['title'])
354 reason = self._match_entry(info_dict)
355 if reason is not None:
356 self.to_screen(u'[download] ' + reason)
359 max_downloads = self.params.get('max_downloads')
360 if max_downloads is not None:
361 if self._num_downloads > int(max_downloads):
362 raise MaxDownloadsReached()
364 filename = self.prepare_filename(info_dict)
367 if self.params.get('forcetitle', False):
368 print(info_dict['title'].encode(preferredencoding(), 'xmlcharrefreplace'))
369 if self.params.get('forceurl', False):
370 print(info_dict['url'].encode(preferredencoding(), 'xmlcharrefreplace'))
371 if self.params.get('forcethumbnail', False) and 'thumbnail' in info_dict:
372 print(info_dict['thumbnail'].encode(preferredencoding(), 'xmlcharrefreplace'))
373 if self.params.get('forcedescription', False) and 'description' in info_dict:
374 print(info_dict['description'].encode(preferredencoding(), 'xmlcharrefreplace'))
375 if self.params.get('forcefilename', False) and filename is not None:
376 print(filename.encode(preferredencoding(), 'xmlcharrefreplace'))
377 if self.params.get('forceformat', False):
378 print(info_dict['format'].encode(preferredencoding(), 'xmlcharrefreplace'))
380 # Do nothing else if in simulate mode
381 if self.params.get('simulate', False):
388 dn = os.path.dirname(encodeFilename(filename))
389 if dn != '' and not os.path.exists(dn): # dn is already encoded
391 except (OSError, IOError), err:
392 self.trouble(u'ERROR: unable to create directory ' + unicode(err))
395 if self.params.get('writedescription', False):
397 descfn = filename + u'.description'
398 self.report_writedescription(descfn)
399 descfile = open(encodeFilename(descfn), 'wb')
401 descfile.write(info_dict['description'].encode('utf-8'))
404 except (OSError, IOError):
405 self.trouble(u'ERROR: Cannot write description file ' + descfn)
408 if self.params.get('writesubtitles', False) and 'subtitles' in info_dict and info_dict['subtitles']:
409 # subtitles download errors are already managed as troubles in relevant IE
410 # that way it will silently go on when used with unsupporting IE
412 srtfn = filename.rsplit('.', 1)[0] + u'.srt'
413 self.report_writesubtitles(srtfn)
414 srtfile = open(encodeFilename(srtfn), 'wb')
416 srtfile.write(info_dict['subtitles'].encode('utf-8'))
419 except (OSError, IOError):
420 self.trouble(u'ERROR: Cannot write subtitles file ' + descfn)
423 if self.params.get('writeinfojson', False):
424 infofn = filename + u'.info.json'
425 self.report_writeinfojson(infofn)
428 except (NameError,AttributeError):
429 self.trouble(u'ERROR: No JSON encoder found. Update to Python 2.6+, setup a json module, or leave out --write-info-json.')
432 infof = open(encodeFilename(infofn), 'wb')
434 json_info_dict = dict((k,v) for k,v in info_dict.iteritems() if not k in ('urlhandle',))
435 json.dump(json_info_dict, infof)
438 except (OSError, IOError):
439 self.trouble(u'ERROR: Cannot write metadata to JSON file ' + infofn)
442 if not self.params.get('skip_download', False):
443 if self.params.get('nooverwrites', False) and os.path.exists(encodeFilename(filename)):
447 success = self._do_download(filename, info_dict)
448 except (OSError, IOError), err:
449 raise UnavailableVideoError
450 except (urllib2.URLError, httplib.HTTPException, socket.error), err:
451 self.trouble(u'ERROR: unable to download video data: %s' % str(err))
453 except (ContentTooShortError, ), err:
454 self.trouble(u'ERROR: content too short (expected %s bytes and served %s)' % (err.expected, err.downloaded))
459 self.post_process(filename, info_dict)
460 except (PostProcessingError), err:
461 self.trouble(u'ERROR: postprocessing: %s' % str(err))
464 def download(self, url_list):
465 """Download a given list of URLs."""
466 if len(url_list) > 1 and self.fixed_template():
467 raise SameFileError(self.params['outtmpl'])
470 suitable_found = False
472 # Go to next InfoExtractor if not suitable
473 if not ie.suitable(url):
476 # Suitable InfoExtractor found
477 suitable_found = True
479 # Extract information from URL and process it
480 videos = ie.extract(url)
481 for video in videos or []:
482 video['extractor'] = ie.IE_NAME
484 self.increment_downloads()
485 self.process_info(video)
486 except UnavailableVideoError:
487 self.trouble(u'\nERROR: unable to download video')
489 # Suitable InfoExtractor had been found; go to next URL
492 if not suitable_found:
493 self.trouble(u'ERROR: no suitable InfoExtractor: %s' % url)
495 return self._download_retcode
497 def post_process(self, filename, ie_info):
498 """Run the postprocessing chain on the given file."""
500 info['filepath'] = filename
506 def _download_with_rtmpdump(self, filename, url, player_url):
507 self.report_destination(filename)
508 tmpfilename = self.temp_name(filename)
510 # Check for rtmpdump first
512 subprocess.call(['rtmpdump', '-h'], stdout=(file(os.path.devnull, 'w')), stderr=subprocess.STDOUT)
513 except (OSError, IOError):
514 self.trouble(u'ERROR: RTMP download detected but "rtmpdump" could not be run')
517 # Download using rtmpdump. rtmpdump returns exit code 2 when
518 # the connection was interrumpted and resuming appears to be
519 # possible. This is part of rtmpdump's normal usage, AFAIK.
520 basic_args = ['rtmpdump', '-q'] + [[], ['-W', player_url]][player_url is not None] + ['-r', url, '-o', tmpfilename]
521 args = basic_args + [[], ['-e', '-k', '1']][self.params.get('continuedl', False)]
522 if self.params.get('verbose', False):
525 shell_quote = lambda args: ' '.join(map(pipes.quote, args))
528 self.to_screen(u'[debug] rtmpdump command line: ' + shell_quote(args))
529 retval = subprocess.call(args)
530 while retval == 2 or retval == 1:
531 prevsize = os.path.getsize(encodeFilename(tmpfilename))
532 self.to_screen(u'\r[rtmpdump] %s bytes' % prevsize, skip_eol=True)
533 time.sleep(5.0) # This seems to be needed
534 retval = subprocess.call(basic_args + ['-e'] + [[], ['-k', '1']][retval == 1])
535 cursize = os.path.getsize(encodeFilename(tmpfilename))
536 if prevsize == cursize and retval == 1:
538 # Some rtmp streams seem abort after ~ 99.8%. Don't complain for those
539 if prevsize == cursize and retval == 2 and cursize > 1024:
540 self.to_screen(u'\r[rtmpdump] Could not download the whole video. This can happen for some advertisements.')
544 self.to_screen(u'\r[rtmpdump] %s bytes' % os.path.getsize(encodeFilename(tmpfilename)))
545 self.try_rename(tmpfilename, filename)
548 self.trouble(u'\nERROR: rtmpdump exited with code %d' % retval)
551 def _do_download(self, filename, info_dict):
552 url = info_dict['url']
553 player_url = info_dict.get('player_url', None)
555 # Check file already present
556 if self.params.get('continuedl', False) and os.path.isfile(encodeFilename(filename)) and not self.params.get('nopart', False):
557 self.report_file_already_downloaded(filename)
560 # Attempt to download using rtmpdump
561 if url.startswith('rtmp'):
562 return self._download_with_rtmpdump(filename, url, player_url)
564 tmpfilename = self.temp_name(filename)
567 # Do not include the Accept-Encoding header
568 headers = {'Youtubedl-no-compression': 'True'}
569 basic_request = urllib2.Request(url, None, headers)
570 request = urllib2.Request(url, None, headers)
572 # Establish possible resume length
573 if os.path.isfile(encodeFilename(tmpfilename)):
574 resume_len = os.path.getsize(encodeFilename(tmpfilename))
580 if self.params.get('continuedl', False):
581 self.report_resuming_byte(resume_len)
582 request.add_header('Range','bytes=%d-' % resume_len)
588 retries = self.params.get('retries', 0)
589 while count <= retries:
590 # Establish connection
592 if count == 0 and 'urlhandle' in info_dict:
593 data = info_dict['urlhandle']
594 data = urllib2.urlopen(request)
596 except (urllib2.HTTPError, ), err:
597 if (err.code < 500 or err.code >= 600) and err.code != 416:
598 # Unexpected HTTP error
600 elif err.code == 416:
601 # Unable to resume (requested range not satisfiable)
603 # Open the connection again without the range header
604 data = urllib2.urlopen(basic_request)
605 content_length = data.info()['Content-Length']
606 except (urllib2.HTTPError, ), err:
607 if err.code < 500 or err.code >= 600:
610 # Examine the reported length
611 if (content_length is not None and
612 (resume_len - 100 < long(content_length) < resume_len + 100)):
613 # The file had already been fully downloaded.
614 # Explanation to the above condition: in issue #175 it was revealed that
615 # YouTube sometimes adds or removes a few bytes from the end of the file,
616 # changing the file size slightly and causing problems for some users. So
617 # I decided to implement a suggested change and consider the file
618 # completely downloaded if the file size differs less than 100 bytes from
619 # the one in the hard drive.
620 self.report_file_already_downloaded(filename)
621 self.try_rename(tmpfilename, filename)
624 # The length does not match, we start the download over
625 self.report_unable_to_resume()
631 self.report_retry(count, retries)
634 self.trouble(u'ERROR: giving up after %s retries' % retries)
637 data_len = data.info().get('Content-length', None)
638 if data_len is not None:
639 data_len = long(data_len) + resume_len
640 data_len_str = self.format_bytes(data_len)
641 byte_counter = 0 + resume_len
647 data_block = data.read(block_size)
649 if len(data_block) == 0:
651 byte_counter += len(data_block)
653 # Open file just in time
656 (stream, tmpfilename) = sanitize_open(tmpfilename, open_mode)
657 assert stream is not None
658 filename = self.undo_temp_name(tmpfilename)
659 self.report_destination(filename)
660 except (OSError, IOError), err:
661 self.trouble(u'ERROR: unable to open for writing: %s' % str(err))
664 stream.write(data_block)
665 except (IOError, OSError), err:
666 self.trouble(u'\nERROR: unable to write data: %s' % str(err))
668 block_size = self.best_block_size(after - before, len(data_block))
671 speed_str = self.calc_speed(start, time.time(), byte_counter - resume_len)
673 self.report_progress('Unknown %', data_len_str, speed_str, 'Unknown ETA')
675 percent_str = self.calc_percent(byte_counter, data_len)
676 eta_str = self.calc_eta(start, time.time(), data_len - resume_len, byte_counter - resume_len)
677 self.report_progress(percent_str, data_len_str, speed_str, eta_str)
680 self.slow_down(start, byte_counter - resume_len)
683 self.trouble(u'\nERROR: Did not get any data blocks')
687 if data_len is not None and byte_counter != data_len:
688 raise ContentTooShortError(byte_counter, long(data_len))
689 self.try_rename(tmpfilename, filename)
691 # Update file modification time
692 if self.params.get('updatetime', True):
693 info_dict['filetime'] = self.try_utime(filename, data.info().get('last-modified', None))