]>
Raphaƫl G. Git Repositories - youtubedl/blob - youtube_dl/FileDownloader.py
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 a long 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 long(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
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 print >>sys
.stderr
, message
.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 filename
= self
.params
['outtmpl'] % template_dict
328 except (ValueError, KeyError), err
:
329 self
.trouble(u
'ERROR: invalid system charset or erroneous output template')
332 def _match_entry(self
, info_dict
):
333 """ Returns None iff the file should be downloaded """
335 title
= info_dict
['title']
336 matchtitle
= self
.params
.get('matchtitle', False)
337 if matchtitle
and not re
.search(matchtitle
, title
, re
.IGNORECASE
):
338 return u
'[download] "' + title
+ '" title did not match pattern "' + matchtitle
+ '"'
339 rejecttitle
= self
.params
.get('rejecttitle', False)
340 if rejecttitle
and re
.search(rejecttitle
, title
, re
.IGNORECASE
):
341 return u
'"' + title
+ '" title matched reject pattern "' + rejecttitle
+ '"'
344 def process_info(self
, info_dict
):
345 """Process a single dictionary returned by an InfoExtractor."""
347 info_dict
['stitle'] = sanitize_filename(info_dict
['title'])
349 reason
= self
._match
_entry
(info_dict
)
350 if reason
is not None:
351 self
.to_screen(u
'[download] ' + reason
)
354 max_downloads
= self
.params
.get('max_downloads')
355 if max_downloads
is not None:
356 if self
._num
_downloads
> int(max_downloads
):
357 raise MaxDownloadsReached()
359 filename
= self
.prepare_filename(info_dict
)
362 if self
.params
.get('forcetitle', False):
363 print info_dict
['title'].encode(preferredencoding(), 'xmlcharrefreplace')
364 if self
.params
.get('forceurl', False):
365 print info_dict
['url'].encode(preferredencoding(), 'xmlcharrefreplace')
366 if self
.params
.get('forcethumbnail', False) and 'thumbnail' in info_dict
:
367 print info_dict
['thumbnail'].encode(preferredencoding(), 'xmlcharrefreplace')
368 if self
.params
.get('forcedescription', False) and 'description' in info_dict
:
369 print info_dict
['description'].encode(preferredencoding(), 'xmlcharrefreplace')
370 if self
.params
.get('forcefilename', False) and filename
is not None:
371 print filename
.encode(preferredencoding(), 'xmlcharrefreplace')
372 if self
.params
.get('forceformat', False):
373 print info_dict
['format'].encode(preferredencoding(), 'xmlcharrefreplace')
375 # Do nothing else if in simulate mode
376 if self
.params
.get('simulate', False):
383 dn
= os
.path
.dirname(encodeFilename(filename
))
384 if dn
!= '' and not os
.path
.exists(dn
): # dn is already encoded
386 except (OSError, IOError), err
:
387 self
.trouble(u
'ERROR: unable to create directory ' + unicode(err
))
390 if self
.params
.get('writedescription', False):
392 descfn
= filename
+ u
'.description'
393 self
.report_writedescription(descfn
)
394 descfile
= open(encodeFilename(descfn
), 'wb')
396 descfile
.write(info_dict
['description'].encode('utf-8'))
399 except (OSError, IOError):
400 self
.trouble(u
'ERROR: Cannot write description file ' + descfn
)
403 if self
.params
.get('writesubtitles', False) and 'subtitles' in info_dict
and info_dict
['subtitles']:
404 # subtitles download errors are already managed as troubles in relevant IE
405 # that way it will silently go on when used with unsupporting IE
407 srtfn
= filename
.rsplit('.', 1)[0] + u
'.srt'
408 self
.report_writesubtitles(srtfn
)
409 srtfile
= open(encodeFilename(srtfn
), 'wb')
411 srtfile
.write(info_dict
['subtitles'].encode('utf-8'))
414 except (OSError, IOError):
415 self
.trouble(u
'ERROR: Cannot write subtitles file ' + descfn
)
418 if self
.params
.get('writeinfojson', False):
419 infofn
= filename
+ u
'.info.json'
420 self
.report_writeinfojson(infofn
)
423 except (NameError,AttributeError):
424 self
.trouble(u
'ERROR: No JSON encoder found. Update to Python 2.6+, setup a json module, or leave out --write-info-json.')
427 infof
= open(encodeFilename(infofn
), 'wb')
429 json_info_dict
= dict((k
,v
) for k
,v
in info_dict
.iteritems() if not k
in ('urlhandle',))
430 json
.dump(json_info_dict
, infof
)
433 except (OSError, IOError):
434 self
.trouble(u
'ERROR: Cannot write metadata to JSON file ' + infofn
)
437 if not self
.params
.get('skip_download', False):
438 if self
.params
.get('nooverwrites', False) and os
.path
.exists(encodeFilename(filename
)):
442 success
= self
._do
_download
(filename
, info_dict
)
443 except (OSError, IOError), err
:
444 raise UnavailableVideoError
445 except (urllib2
.URLError
, httplib
.HTTPException
, socket
.error
), err
:
446 self
.trouble(u
'ERROR: unable to download video data: %s' % str(err
))
448 except (ContentTooShortError
, ), err
:
449 self
.trouble(u
'ERROR: content too short (expected %s bytes and served %s)' % (err
.expected
, err
.downloaded
))
454 self
.post_process(filename
, info_dict
)
455 except (PostProcessingError
), err
:
456 self
.trouble(u
'ERROR: postprocessing: %s' % str(err
))
459 def download(self
, url_list
):
460 """Download a given list of URLs."""
461 if len(url_list
) > 1 and self
.fixed_template():
462 raise SameFileError(self
.params
['outtmpl'])
465 suitable_found
= False
467 # Go to next InfoExtractor if not suitable
468 if not ie
.suitable(url
):
471 # Suitable InfoExtractor found
472 suitable_found
= True
474 # Extract information from URL and process it
475 videos
= ie
.extract(url
)
476 for video
in videos
or []:
478 self
.increment_downloads()
479 self
.process_info(video
)
480 except UnavailableVideoError
:
481 self
.trouble(u
'\nERROR: unable to download video')
483 # Suitable InfoExtractor had been found; go to next URL
486 if not suitable_found
:
487 self
.trouble(u
'ERROR: no suitable InfoExtractor: %s' % url
)
489 return self
._download
_retcode
491 def post_process(self
, filename
, ie_info
):
492 """Run the postprocessing chain on the given file."""
494 info
['filepath'] = filename
500 def _download_with_rtmpdump(self
, filename
, url
, player_url
):
501 self
.report_destination(filename
)
502 tmpfilename
= self
.temp_name(filename
)
504 # Check for rtmpdump first
506 subprocess
.call(['rtmpdump', '-h'], stdout
=(file(os
.path
.devnull
, 'w')), stderr
=subprocess
.STDOUT
)
507 except (OSError, IOError):
508 self
.trouble(u
'ERROR: RTMP download detected but "rtmpdump" could not be run')
511 # Download using rtmpdump. rtmpdump returns exit code 2 when
512 # the connection was interrumpted and resuming appears to be
513 # possible. This is part of rtmpdump's normal usage, AFAIK.
514 basic_args
= ['rtmpdump', '-q'] + [[], ['-W', player_url
]][player_url
is not None] + ['-r', url
, '-o', tmpfilename
]
515 args
= basic_args
+ [[], ['-e', '-k', '1']][self
.params
.get('continuedl', False)]
516 if self
.params
.get('verbose', False):
519 shell_quote
= lambda args
: ' '.join(map(pipes
.quote
, args
))
522 self
.to_screen(u
'[debug] rtmpdump command line: ' + shell_quote(args
))
523 retval
= subprocess
.call(args
)
524 while retval
== 2 or retval
== 1:
525 prevsize
= os
.path
.getsize(encodeFilename(tmpfilename
))
526 self
.to_screen(u
'\r[rtmpdump] %s bytes' % prevsize
, skip_eol
=True)
527 time
.sleep(5.0) # This seems to be needed
528 retval
= subprocess
.call(basic_args
+ ['-e'] + [[], ['-k', '1']][retval
== 1])
529 cursize
= os
.path
.getsize(encodeFilename(tmpfilename
))
530 if prevsize
== cursize
and retval
== 1:
532 # Some rtmp streams seem abort after ~ 99.8%. Don't complain for those
533 if prevsize
== cursize
and retval
== 2 and cursize
> 1024:
534 self
.to_screen(u
'\r[rtmpdump] Could not download the whole video. This can happen for some advertisements.')
538 self
.to_screen(u
'\r[rtmpdump] %s bytes' % os
.path
.getsize(encodeFilename(tmpfilename
)))
539 self
.try_rename(tmpfilename
, filename
)
542 self
.trouble(u
'\nERROR: rtmpdump exited with code %d' % retval
)
545 def _do_download(self
, filename
, info_dict
):
546 url
= info_dict
['url']
547 player_url
= info_dict
.get('player_url', None)
549 # Check file already present
550 if self
.params
.get('continuedl', False) and os
.path
.isfile(encodeFilename(filename
)) and not self
.params
.get('nopart', False):
551 self
.report_file_already_downloaded(filename
)
554 # Attempt to download using rtmpdump
555 if url
.startswith('rtmp'):
556 return self
._download
_with
_rtmpdump
(filename
, url
, player_url
)
558 tmpfilename
= self
.temp_name(filename
)
561 # Do not include the Accept-Encoding header
562 headers
= {'Youtubedl-no-compression': 'True'}
563 basic_request
= urllib2
.Request(url
, None, headers
)
564 request
= urllib2
.Request(url
, None, headers
)
566 # Establish possible resume length
567 if os
.path
.isfile(encodeFilename(tmpfilename
)):
568 resume_len
= os
.path
.getsize(encodeFilename(tmpfilename
))
574 if self
.params
.get('continuedl', False):
575 self
.report_resuming_byte(resume_len
)
576 request
.add_header('Range','bytes=%d-' % resume_len
)
582 retries
= self
.params
.get('retries', 0)
583 while count
<= retries
:
584 # Establish connection
586 if count
== 0 and 'urlhandle' in info_dict
:
587 data
= info_dict
['urlhandle']
588 data
= urllib2
.urlopen(request
)
590 except (urllib2
.HTTPError
, ), err
:
591 if (err
.code
< 500 or err
.code
>= 600) and err
.code
!= 416:
592 # Unexpected HTTP error
594 elif err
.code
== 416:
595 # Unable to resume (requested range not satisfiable)
597 # Open the connection again without the range header
598 data
= urllib2
.urlopen(basic_request
)
599 content_length
= data
.info()['Content-Length']
600 except (urllib2
.HTTPError
, ), err
:
601 if err
.code
< 500 or err
.code
>= 600:
604 # Examine the reported length
605 if (content_length
is not None and
606 (resume_len
- 100 < long(content_length
) < resume_len
+ 100)):
607 # The file had already been fully downloaded.
608 # Explanation to the above condition: in issue #175 it was revealed that
609 # YouTube sometimes adds or removes a few bytes from the end of the file,
610 # changing the file size slightly and causing problems for some users. So
611 # I decided to implement a suggested change and consider the file
612 # completely downloaded if the file size differs less than 100 bytes from
613 # the one in the hard drive.
614 self
.report_file_already_downloaded(filename
)
615 self
.try_rename(tmpfilename
, filename
)
618 # The length does not match, we start the download over
619 self
.report_unable_to_resume()
625 self
.report_retry(count
, retries
)
628 self
.trouble(u
'ERROR: giving up after %s retries' % retries
)
631 data_len
= data
.info().get('Content-length', None)
632 if data_len
is not None:
633 data_len
= long(data_len
) + resume_len
634 data_len_str
= self
.format_bytes(data_len
)
635 byte_counter
= 0 + resume_len
641 data_block
= data
.read(block_size
)
643 if len(data_block
) == 0:
645 byte_counter
+= len(data_block
)
647 # Open file just in time
650 (stream
, tmpfilename
) = sanitize_open(tmpfilename
, open_mode
)
651 assert stream
is not None
652 filename
= self
.undo_temp_name(tmpfilename
)
653 self
.report_destination(filename
)
654 except (OSError, IOError), err
:
655 self
.trouble(u
'ERROR: unable to open for writing: %s' % str(err
))
658 stream
.write(data_block
)
659 except (IOError, OSError), err
:
660 self
.trouble(u
'\nERROR: unable to write data: %s' % str(err
))
662 block_size
= self
.best_block_size(after
- before
, len(data_block
))
665 speed_str
= self
.calc_speed(start
, time
.time(), byte_counter
- resume_len
)
667 self
.report_progress('Unknown %', data_len_str
, speed_str
, 'Unknown ETA')
669 percent_str
= self
.calc_percent(byte_counter
, data_len
)
670 eta_str
= self
.calc_eta(start
, time
.time(), data_len
- resume_len
, byte_counter
- resume_len
)
671 self
.report_progress(percent_str
, data_len_str
, speed_str
, eta_str
)
674 self
.slow_down(start
, byte_counter
- resume_len
)
677 self
.trouble(u
'\nERROR: Did not get any data blocks')
681 if data_len
is not None and byte_counter
!= data_len
:
682 raise ContentTooShortError(byte_counter
, long(data_len
))
683 self
.try_rename(tmpfilename
, filename
)
685 # Update file modification time
686 if self
.params
.get('updatetime', True):
687 info_dict
['filetime'] = self
.try_utime(filename
, data
.info().get('last-modified', None))