]> Raphaƫl G. Git Repositories - youtubedl/blob - youtube_dl/utils.py
Prepare for upload.
[youtubedl] / youtube_dl / utils.py
1 #!/usr/bin/env python
2 # -*- coding: utf-8 -*-
3
4 from __future__ import unicode_literals
5
6 import calendar
7 import codecs
8 import contextlib
9 import ctypes
10 import datetime
11 import email.utils
12 import errno
13 import functools
14 import gzip
15 import itertools
16 import io
17 import json
18 import locale
19 import math
20 import operator
21 import os
22 import pipes
23 import platform
24 import re
25 import ssl
26 import socket
27 import struct
28 import subprocess
29 import sys
30 import tempfile
31 import traceback
32 import xml.etree.ElementTree
33 import zlib
34
35 from .compat import (
36 compat_basestring,
37 compat_chr,
38 compat_html_entities,
39 compat_http_client,
40 compat_kwargs,
41 compat_parse_qs,
42 compat_socket_create_connection,
43 compat_str,
44 compat_urllib_error,
45 compat_urllib_parse,
46 compat_urllib_parse_urlparse,
47 compat_urllib_request,
48 compat_urlparse,
49 shlex_quote,
50 )
51
52
53 # This is not clearly defined otherwise
54 compiled_regex_type = type(re.compile(''))
55
56 std_headers = {
57 'User-Agent': 'Mozilla/5.0 (X11; Linux x86_64; rv:10.0) Gecko/20150101 Firefox/20.0 (Chrome)',
58 'Accept-Charset': 'ISO-8859-1,utf-8;q=0.7,*;q=0.7',
59 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
60 'Accept-Encoding': 'gzip, deflate',
61 'Accept-Language': 'en-us,en;q=0.5',
62 }
63
64
65 ENGLISH_MONTH_NAMES = [
66 'January', 'February', 'March', 'April', 'May', 'June',
67 'July', 'August', 'September', 'October', 'November', 'December']
68
69
70 def preferredencoding():
71 """Get preferred encoding.
72
73 Returns the best encoding scheme for the system, based on
74 locale.getpreferredencoding() and some further tweaks.
75 """
76 try:
77 pref = locale.getpreferredencoding()
78 'TEST'.encode(pref)
79 except Exception:
80 pref = 'UTF-8'
81
82 return pref
83
84
85 def write_json_file(obj, fn):
86 """ Encode obj as JSON and write it to fn, atomically if possible """
87
88 fn = encodeFilename(fn)
89 if sys.version_info < (3, 0) and sys.platform != 'win32':
90 encoding = get_filesystem_encoding()
91 # os.path.basename returns a bytes object, but NamedTemporaryFile
92 # will fail if the filename contains non ascii characters unless we
93 # use a unicode object
94 path_basename = lambda f: os.path.basename(fn).decode(encoding)
95 # the same for os.path.dirname
96 path_dirname = lambda f: os.path.dirname(fn).decode(encoding)
97 else:
98 path_basename = os.path.basename
99 path_dirname = os.path.dirname
100
101 args = {
102 'suffix': '.tmp',
103 'prefix': path_basename(fn) + '.',
104 'dir': path_dirname(fn),
105 'delete': False,
106 }
107
108 # In Python 2.x, json.dump expects a bytestream.
109 # In Python 3.x, it writes to a character stream
110 if sys.version_info < (3, 0):
111 args['mode'] = 'wb'
112 else:
113 args.update({
114 'mode': 'w',
115 'encoding': 'utf-8',
116 })
117
118 tf = tempfile.NamedTemporaryFile(**compat_kwargs(args))
119
120 try:
121 with tf:
122 json.dump(obj, tf)
123 if sys.platform == 'win32':
124 # Need to remove existing file on Windows, else os.rename raises
125 # WindowsError or FileExistsError.
126 try:
127 os.unlink(fn)
128 except OSError:
129 pass
130 os.rename(tf.name, fn)
131 except Exception:
132 try:
133 os.remove(tf.name)
134 except OSError:
135 pass
136 raise
137
138
139 if sys.version_info >= (2, 7):
140 def find_xpath_attr(node, xpath, key, val):
141 """ Find the xpath xpath[@key=val] """
142 assert re.match(r'^[a-zA-Z-]+$', key)
143 assert re.match(r'^[a-zA-Z0-9@\s:._-]*$', val)
144 expr = xpath + "[@%s='%s']" % (key, val)
145 return node.find(expr)
146 else:
147 def find_xpath_attr(node, xpath, key, val):
148 # Here comes the crazy part: In 2.6, if the xpath is a unicode,
149 # .//node does not match if a node is a direct child of . !
150 if isinstance(xpath, compat_str):
151 xpath = xpath.encode('ascii')
152
153 for f in node.findall(xpath):
154 if f.attrib.get(key) == val:
155 return f
156 return None
157
158 # On python2.6 the xml.etree.ElementTree.Element methods don't support
159 # the namespace parameter
160
161
162 def xpath_with_ns(path, ns_map):
163 components = [c.split(':') for c in path.split('/')]
164 replaced = []
165 for c in components:
166 if len(c) == 1:
167 replaced.append(c[0])
168 else:
169 ns, tag = c
170 replaced.append('{%s}%s' % (ns_map[ns], tag))
171 return '/'.join(replaced)
172
173
174 def xpath_text(node, xpath, name=None, fatal=False):
175 if sys.version_info < (2, 7): # Crazy 2.6
176 xpath = xpath.encode('ascii')
177
178 n = node.find(xpath)
179 if n is None or n.text is None:
180 if fatal:
181 name = xpath if name is None else name
182 raise ExtractorError('Could not find XML element %s' % name)
183 else:
184 return None
185 return n.text
186
187
188 def get_element_by_id(id, html):
189 """Return the content of the tag with the specified ID in the passed HTML document"""
190 return get_element_by_attribute("id", id, html)
191
192
193 def get_element_by_attribute(attribute, value, html):
194 """Return the content of the tag with the specified attribute in the passed HTML document"""
195
196 m = re.search(r'''(?xs)
197 <([a-zA-Z0-9:._-]+)
198 (?:\s+[a-zA-Z0-9:._-]+(?:=[a-zA-Z0-9:._-]+|="[^"]+"|='[^']+'))*?
199 \s+%s=['"]?%s['"]?
200 (?:\s+[a-zA-Z0-9:._-]+(?:=[a-zA-Z0-9:._-]+|="[^"]+"|='[^']+'))*?
201 \s*>
202 (?P<content>.*?)
203 </\1>
204 ''' % (re.escape(attribute), re.escape(value)), html)
205
206 if not m:
207 return None
208 res = m.group('content')
209
210 if res.startswith('"') or res.startswith("'"):
211 res = res[1:-1]
212
213 return unescapeHTML(res)
214
215
216 def clean_html(html):
217 """Clean an HTML snippet into a readable string"""
218
219 if html is None: # Convenience for sanitizing descriptions etc.
220 return html
221
222 # Newline vs <br />
223 html = html.replace('\n', ' ')
224 html = re.sub(r'\s*<\s*br\s*/?\s*>\s*', '\n', html)
225 html = re.sub(r'<\s*/\s*p\s*>\s*<\s*p[^>]*>', '\n', html)
226 # Strip html tags
227 html = re.sub('<.*?>', '', html)
228 # Replace html entities
229 html = unescapeHTML(html)
230 return html.strip()
231
232
233 def sanitize_open(filename, open_mode):
234 """Try to open the given filename, and slightly tweak it if this fails.
235
236 Attempts to open the given filename. If this fails, it tries to change
237 the filename slightly, step by step, until it's either able to open it
238 or it fails and raises a final exception, like the standard open()
239 function.
240
241 It returns the tuple (stream, definitive_file_name).
242 """
243 try:
244 if filename == '-':
245 if sys.platform == 'win32':
246 import msvcrt
247 msvcrt.setmode(sys.stdout.fileno(), os.O_BINARY)
248 return (sys.stdout.buffer if hasattr(sys.stdout, 'buffer') else sys.stdout, filename)
249 stream = open(encodeFilename(filename), open_mode)
250 return (stream, filename)
251 except (IOError, OSError) as err:
252 if err.errno in (errno.EACCES,):
253 raise
254
255 # In case of error, try to remove win32 forbidden chars
256 alt_filename = sanitize_path(filename)
257 if alt_filename == filename:
258 raise
259 else:
260 # An exception here should be caught in the caller
261 stream = open(encodeFilename(alt_filename), open_mode)
262 return (stream, alt_filename)
263
264
265 def timeconvert(timestr):
266 """Convert RFC 2822 defined time string into system timestamp"""
267 timestamp = None
268 timetuple = email.utils.parsedate_tz(timestr)
269 if timetuple is not None:
270 timestamp = email.utils.mktime_tz(timetuple)
271 return timestamp
272
273
274 def sanitize_filename(s, restricted=False, is_id=False):
275 """Sanitizes a string so it could be used as part of a filename.
276 If restricted is set, use a stricter subset of allowed characters.
277 Set is_id if this is not an arbitrary string, but an ID that should be kept if possible
278 """
279 def replace_insane(char):
280 if char == '?' or ord(char) < 32 or ord(char) == 127:
281 return ''
282 elif char == '"':
283 return '' if restricted else '\''
284 elif char == ':':
285 return '_-' if restricted else ' -'
286 elif char in '\\/|*<>':
287 return '_'
288 if restricted and (char in '!&\'()[]{}$;`^,#' or char.isspace()):
289 return '_'
290 if restricted and ord(char) > 127:
291 return '_'
292 return char
293
294 # Handle timestamps
295 s = re.sub(r'[0-9]+(?::[0-9]+)+', lambda m: m.group(0).replace(':', '_'), s)
296 result = ''.join(map(replace_insane, s))
297 if not is_id:
298 while '__' in result:
299 result = result.replace('__', '_')
300 result = result.strip('_')
301 # Common case of "Foreign band name - English song title"
302 if restricted and result.startswith('-_'):
303 result = result[2:]
304 if result.startswith('-'):
305 result = '_' + result[len('-'):]
306 result = result.lstrip('.')
307 if not result:
308 result = '_'
309 return result
310
311
312 def sanitize_path(s):
313 """Sanitizes and normalizes path on Windows"""
314 if sys.platform != 'win32':
315 return s
316 drive_or_unc, _ = os.path.splitdrive(s)
317 if sys.version_info < (2, 7) and not drive_or_unc:
318 drive_or_unc, _ = os.path.splitunc(s)
319 norm_path = os.path.normpath(remove_start(s, drive_or_unc)).split(os.path.sep)
320 if drive_or_unc:
321 norm_path.pop(0)
322 sanitized_path = [
323 path_part if path_part in ['.', '..'] else re.sub('(?:[/<>:"\\|\\\\?\\*]|\.$)', '#', path_part)
324 for path_part in norm_path]
325 if drive_or_unc:
326 sanitized_path.insert(0, drive_or_unc + os.path.sep)
327 return os.path.join(*sanitized_path)
328
329
330 def orderedSet(iterable):
331 """ Remove all duplicates from the input iterable """
332 res = []
333 for el in iterable:
334 if el not in res:
335 res.append(el)
336 return res
337
338
339 def _htmlentity_transform(entity):
340 """Transforms an HTML entity to a character."""
341 # Known non-numeric HTML entity
342 if entity in compat_html_entities.name2codepoint:
343 return compat_chr(compat_html_entities.name2codepoint[entity])
344
345 mobj = re.match(r'#(x[0-9a-fA-F]+|[0-9]+)', entity)
346 if mobj is not None:
347 numstr = mobj.group(1)
348 if numstr.startswith('x'):
349 base = 16
350 numstr = '0%s' % numstr
351 else:
352 base = 10
353 return compat_chr(int(numstr, base))
354
355 # Unknown entity in name, return its literal representation
356 return ('&%s;' % entity)
357
358
359 def unescapeHTML(s):
360 if s is None:
361 return None
362 assert type(s) == compat_str
363
364 return re.sub(
365 r'&([^;]+);', lambda m: _htmlentity_transform(m.group(1)), s)
366
367
368 def get_subprocess_encoding():
369 if sys.platform == 'win32' and sys.getwindowsversion()[0] >= 5:
370 # For subprocess calls, encode with locale encoding
371 # Refer to http://stackoverflow.com/a/9951851/35070
372 encoding = preferredencoding()
373 else:
374 encoding = sys.getfilesystemencoding()
375 if encoding is None:
376 encoding = 'utf-8'
377 return encoding
378
379
380 def encodeFilename(s, for_subprocess=False):
381 """
382 @param s The name of the file
383 """
384
385 assert type(s) == compat_str
386
387 # Python 3 has a Unicode API
388 if sys.version_info >= (3, 0):
389 return s
390
391 # Pass '' directly to use Unicode APIs on Windows 2000 and up
392 # (Detecting Windows NT 4 is tricky because 'major >= 4' would
393 # match Windows 9x series as well. Besides, NT 4 is obsolete.)
394 if not for_subprocess and sys.platform == 'win32' and sys.getwindowsversion()[0] >= 5:
395 return s
396
397 return s.encode(get_subprocess_encoding(), 'ignore')
398
399
400 def decodeFilename(b, for_subprocess=False):
401
402 if sys.version_info >= (3, 0):
403 return b
404
405 if not isinstance(b, bytes):
406 return b
407
408 return b.decode(get_subprocess_encoding(), 'ignore')
409
410
411 def encodeArgument(s):
412 if not isinstance(s, compat_str):
413 # Legacy code that uses byte strings
414 # Uncomment the following line after fixing all post processors
415 # assert False, 'Internal error: %r should be of type %r, is %r' % (s, compat_str, type(s))
416 s = s.decode('ascii')
417 return encodeFilename(s, True)
418
419
420 def decodeArgument(b):
421 return decodeFilename(b, True)
422
423
424 def decodeOption(optval):
425 if optval is None:
426 return optval
427 if isinstance(optval, bytes):
428 optval = optval.decode(preferredencoding())
429
430 assert isinstance(optval, compat_str)
431 return optval
432
433
434 def formatSeconds(secs):
435 if secs > 3600:
436 return '%d:%02d:%02d' % (secs // 3600, (secs % 3600) // 60, secs % 60)
437 elif secs > 60:
438 return '%d:%02d' % (secs // 60, secs % 60)
439 else:
440 return '%d' % secs
441
442
443 def make_HTTPS_handler(params, **kwargs):
444 opts_no_check_certificate = params.get('nocheckcertificate', False)
445 if hasattr(ssl, 'create_default_context'): # Python >= 3.4 or 2.7.9
446 context = ssl.create_default_context(ssl.Purpose.SERVER_AUTH)
447 if opts_no_check_certificate:
448 context.check_hostname = False
449 context.verify_mode = ssl.CERT_NONE
450 try:
451 return YoutubeDLHTTPSHandler(params, context=context, **kwargs)
452 except TypeError:
453 # Python 2.7.8
454 # (create_default_context present but HTTPSHandler has no context=)
455 pass
456
457 if sys.version_info < (3, 2):
458 return YoutubeDLHTTPSHandler(params, **kwargs)
459 else: # Python < 3.4
460 context = ssl.SSLContext(ssl.PROTOCOL_TLSv1)
461 context.verify_mode = (ssl.CERT_NONE
462 if opts_no_check_certificate
463 else ssl.CERT_REQUIRED)
464 context.set_default_verify_paths()
465 return YoutubeDLHTTPSHandler(params, context=context, **kwargs)
466
467
468 def bug_reports_message():
469 if ytdl_is_updateable():
470 update_cmd = 'type youtube-dl -U to update'
471 else:
472 update_cmd = 'see https://yt-dl.org/update on how to update'
473 msg = '; please report this issue on https://yt-dl.org/bug .'
474 msg += ' Make sure you are using the latest version; %s.' % update_cmd
475 msg += ' Be sure to call youtube-dl with the --verbose flag and include its complete output.'
476 return msg
477
478
479 class ExtractorError(Exception):
480 """Error during info extraction."""
481
482 def __init__(self, msg, tb=None, expected=False, cause=None, video_id=None):
483 """ tb, if given, is the original traceback (so that it can be printed out).
484 If expected is set, this is a normal error message and most likely not a bug in youtube-dl.
485 """
486
487 if sys.exc_info()[0] in (compat_urllib_error.URLError, socket.timeout, UnavailableVideoError):
488 expected = True
489 if video_id is not None:
490 msg = video_id + ': ' + msg
491 if cause:
492 msg += ' (caused by %r)' % cause
493 if not expected:
494 msg += bug_reports_message()
495 super(ExtractorError, self).__init__(msg)
496
497 self.traceback = tb
498 self.exc_info = sys.exc_info() # preserve original exception
499 self.cause = cause
500 self.video_id = video_id
501
502 def format_traceback(self):
503 if self.traceback is None:
504 return None
505 return ''.join(traceback.format_tb(self.traceback))
506
507
508 class UnsupportedError(ExtractorError):
509 def __init__(self, url):
510 super(UnsupportedError, self).__init__(
511 'Unsupported URL: %s' % url, expected=True)
512 self.url = url
513
514
515 class RegexNotFoundError(ExtractorError):
516 """Error when a regex didn't match"""
517 pass
518
519
520 class DownloadError(Exception):
521 """Download Error exception.
522
523 This exception may be thrown by FileDownloader objects if they are not
524 configured to continue on errors. They will contain the appropriate
525 error message.
526 """
527
528 def __init__(self, msg, exc_info=None):
529 """ exc_info, if given, is the original exception that caused the trouble (as returned by sys.exc_info()). """
530 super(DownloadError, self).__init__(msg)
531 self.exc_info = exc_info
532
533
534 class SameFileError(Exception):
535 """Same File exception.
536
537 This exception will be thrown by FileDownloader objects if they detect
538 multiple files would have to be downloaded to the same file on disk.
539 """
540 pass
541
542
543 class PostProcessingError(Exception):
544 """Post Processing exception.
545
546 This exception may be raised by PostProcessor's .run() method to
547 indicate an error in the postprocessing task.
548 """
549
550 def __init__(self, msg):
551 self.msg = msg
552
553
554 class MaxDownloadsReached(Exception):
555 """ --max-downloads limit has been reached. """
556 pass
557
558
559 class UnavailableVideoError(Exception):
560 """Unavailable Format exception.
561
562 This exception will be thrown when a video is requested
563 in a format that is not available for that video.
564 """
565 pass
566
567
568 class ContentTooShortError(Exception):
569 """Content Too Short exception.
570
571 This exception may be raised by FileDownloader objects when a file they
572 download is too small for what the server announced first, indicating
573 the connection was probably interrupted.
574 """
575 # Both in bytes
576 downloaded = None
577 expected = None
578
579 def __init__(self, downloaded, expected):
580 self.downloaded = downloaded
581 self.expected = expected
582
583
584 def _create_http_connection(ydl_handler, http_class, is_https, *args, **kwargs):
585 hc = http_class(*args, **kwargs)
586 source_address = ydl_handler._params.get('source_address')
587 if source_address is not None:
588 sa = (source_address, 0)
589 if hasattr(hc, 'source_address'): # Python 2.7+
590 hc.source_address = sa
591 else: # Python 2.6
592 def _hc_connect(self, *args, **kwargs):
593 sock = compat_socket_create_connection(
594 (self.host, self.port), self.timeout, sa)
595 if is_https:
596 self.sock = ssl.wrap_socket(
597 sock, self.key_file, self.cert_file,
598 ssl_version=ssl.PROTOCOL_TLSv1)
599 else:
600 self.sock = sock
601 hc.connect = functools.partial(_hc_connect, hc)
602
603 return hc
604
605
606 class YoutubeDLHandler(compat_urllib_request.HTTPHandler):
607 """Handler for HTTP requests and responses.
608
609 This class, when installed with an OpenerDirector, automatically adds
610 the standard headers to every HTTP request and handles gzipped and
611 deflated responses from web servers. If compression is to be avoided in
612 a particular request, the original request in the program code only has
613 to include the HTTP header "Youtubedl-No-Compression", which will be
614 removed before making the real request.
615
616 Part of this code was copied from:
617
618 http://techknack.net/python-urllib2-handlers/
619
620 Andrew Rowls, the author of that code, agreed to release it to the
621 public domain.
622 """
623
624 def __init__(self, params, *args, **kwargs):
625 compat_urllib_request.HTTPHandler.__init__(self, *args, **kwargs)
626 self._params = params
627
628 def http_open(self, req):
629 return self.do_open(functools.partial(
630 _create_http_connection, self, compat_http_client.HTTPConnection, False),
631 req)
632
633 @staticmethod
634 def deflate(data):
635 try:
636 return zlib.decompress(data, -zlib.MAX_WBITS)
637 except zlib.error:
638 return zlib.decompress(data)
639
640 @staticmethod
641 def addinfourl_wrapper(stream, headers, url, code):
642 if hasattr(compat_urllib_request.addinfourl, 'getcode'):
643 return compat_urllib_request.addinfourl(stream, headers, url, code)
644 ret = compat_urllib_request.addinfourl(stream, headers, url)
645 ret.code = code
646 return ret
647
648 def http_request(self, req):
649 for h, v in std_headers.items():
650 # Capitalize is needed because of Python bug 2275: http://bugs.python.org/issue2275
651 # The dict keys are capitalized because of this bug by urllib
652 if h.capitalize() not in req.headers:
653 req.add_header(h, v)
654 if 'Youtubedl-no-compression' in req.headers:
655 if 'Accept-encoding' in req.headers:
656 del req.headers['Accept-encoding']
657 del req.headers['Youtubedl-no-compression']
658
659 if sys.version_info < (2, 7) and '#' in req.get_full_url():
660 # Python 2.6 is brain-dead when it comes to fragments
661 req._Request__original = req._Request__original.partition('#')[0]
662 req._Request__r_type = req._Request__r_type.partition('#')[0]
663
664 return req
665
666 def http_response(self, req, resp):
667 old_resp = resp
668 # gzip
669 if resp.headers.get('Content-encoding', '') == 'gzip':
670 content = resp.read()
671 gz = gzip.GzipFile(fileobj=io.BytesIO(content), mode='rb')
672 try:
673 uncompressed = io.BytesIO(gz.read())
674 except IOError as original_ioerror:
675 # There may be junk add the end of the file
676 # See http://stackoverflow.com/q/4928560/35070 for details
677 for i in range(1, 1024):
678 try:
679 gz = gzip.GzipFile(fileobj=io.BytesIO(content[:-i]), mode='rb')
680 uncompressed = io.BytesIO(gz.read())
681 except IOError:
682 continue
683 break
684 else:
685 raise original_ioerror
686 resp = self.addinfourl_wrapper(uncompressed, old_resp.headers, old_resp.url, old_resp.code)
687 resp.msg = old_resp.msg
688 # deflate
689 if resp.headers.get('Content-encoding', '') == 'deflate':
690 gz = io.BytesIO(self.deflate(resp.read()))
691 resp = self.addinfourl_wrapper(gz, old_resp.headers, old_resp.url, old_resp.code)
692 resp.msg = old_resp.msg
693 return resp
694
695 https_request = http_request
696 https_response = http_response
697
698
699 class YoutubeDLHTTPSHandler(compat_urllib_request.HTTPSHandler):
700 def __init__(self, params, https_conn_class=None, *args, **kwargs):
701 compat_urllib_request.HTTPSHandler.__init__(self, *args, **kwargs)
702 self._https_conn_class = https_conn_class or compat_http_client.HTTPSConnection
703 self._params = params
704
705 def https_open(self, req):
706 kwargs = {}
707 if hasattr(self, '_context'): # python > 2.6
708 kwargs['context'] = self._context
709 if hasattr(self, '_check_hostname'): # python 3.x
710 kwargs['check_hostname'] = self._check_hostname
711 return self.do_open(functools.partial(
712 _create_http_connection, self, self._https_conn_class, True),
713 req, **kwargs)
714
715
716 def parse_iso8601(date_str, delimiter='T', timezone=None):
717 """ Return a UNIX timestamp from the given date """
718
719 if date_str is None:
720 return None
721
722 if timezone is None:
723 m = re.search(
724 r'(\.[0-9]+)?(?:Z$| ?(?P<sign>\+|-)(?P<hours>[0-9]{2}):?(?P<minutes>[0-9]{2})$)',
725 date_str)
726 if not m:
727 timezone = datetime.timedelta()
728 else:
729 date_str = date_str[:-len(m.group(0))]
730 if not m.group('sign'):
731 timezone = datetime.timedelta()
732 else:
733 sign = 1 if m.group('sign') == '+' else -1
734 timezone = datetime.timedelta(
735 hours=sign * int(m.group('hours')),
736 minutes=sign * int(m.group('minutes')))
737 date_format = '%Y-%m-%d{0}%H:%M:%S'.format(delimiter)
738 dt = datetime.datetime.strptime(date_str, date_format) - timezone
739 return calendar.timegm(dt.timetuple())
740
741
742 def unified_strdate(date_str, day_first=True):
743 """Return a string with the date in the format YYYYMMDD"""
744
745 if date_str is None:
746 return None
747 upload_date = None
748 # Replace commas
749 date_str = date_str.replace(',', ' ')
750 # %z (UTC offset) is only supported in python>=3.2
751 if not re.match(r'^[0-9]{1,2}-[0-9]{1,2}-[0-9]{4}$', date_str):
752 date_str = re.sub(r' ?(\+|-)[0-9]{2}:?[0-9]{2}$', '', date_str)
753 # Remove AM/PM + timezone
754 date_str = re.sub(r'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str)
755
756 format_expressions = [
757 '%d %B %Y',
758 '%d %b %Y',
759 '%B %d %Y',
760 '%b %d %Y',
761 '%b %dst %Y %I:%M%p',
762 '%b %dnd %Y %I:%M%p',
763 '%b %dth %Y %I:%M%p',
764 '%Y %m %d',
765 '%Y-%m-%d',
766 '%Y/%m/%d',
767 '%Y/%m/%d %H:%M:%S',
768 '%Y-%m-%d %H:%M:%S',
769 '%Y-%m-%d %H:%M:%S.%f',
770 '%d.%m.%Y %H:%M',
771 '%d.%m.%Y %H.%M',
772 '%Y-%m-%dT%H:%M:%SZ',
773 '%Y-%m-%dT%H:%M:%S.%fZ',
774 '%Y-%m-%dT%H:%M:%S.%f0Z',
775 '%Y-%m-%dT%H:%M:%S',
776 '%Y-%m-%dT%H:%M:%S.%f',
777 '%Y-%m-%dT%H:%M',
778 ]
779 if day_first:
780 format_expressions.extend([
781 '%d-%m-%Y',
782 '%d.%m.%Y',
783 '%d/%m/%Y',
784 '%d/%m/%y',
785 '%d/%m/%Y %H:%M:%S',
786 ])
787 else:
788 format_expressions.extend([
789 '%m-%d-%Y',
790 '%m.%d.%Y',
791 '%m/%d/%Y',
792 '%m/%d/%y',
793 '%m/%d/%Y %H:%M:%S',
794 ])
795 for expression in format_expressions:
796 try:
797 upload_date = datetime.datetime.strptime(date_str, expression).strftime('%Y%m%d')
798 except ValueError:
799 pass
800 if upload_date is None:
801 timetuple = email.utils.parsedate_tz(date_str)
802 if timetuple:
803 upload_date = datetime.datetime(*timetuple[:6]).strftime('%Y%m%d')
804 return upload_date
805
806
807 def determine_ext(url, default_ext='unknown_video'):
808 if url is None:
809 return default_ext
810 guess = url.partition('?')[0].rpartition('.')[2]
811 if re.match(r'^[A-Za-z0-9]+$', guess):
812 return guess
813 else:
814 return default_ext
815
816
817 def subtitles_filename(filename, sub_lang, sub_format):
818 return filename.rsplit('.', 1)[0] + '.' + sub_lang + '.' + sub_format
819
820
821 def date_from_str(date_str):
822 """
823 Return a datetime object from a string in the format YYYYMMDD or
824 (now|today)[+-][0-9](day|week|month|year)(s)?"""
825 today = datetime.date.today()
826 if date_str in ('now', 'today'):
827 return today
828 if date_str == 'yesterday':
829 return today - datetime.timedelta(days=1)
830 match = re.match('(now|today)(?P<sign>[+-])(?P<time>\d+)(?P<unit>day|week|month|year)(s)?', date_str)
831 if match is not None:
832 sign = match.group('sign')
833 time = int(match.group('time'))
834 if sign == '-':
835 time = -time
836 unit = match.group('unit')
837 # A bad aproximation?
838 if unit == 'month':
839 unit = 'day'
840 time *= 30
841 elif unit == 'year':
842 unit = 'day'
843 time *= 365
844 unit += 's'
845 delta = datetime.timedelta(**{unit: time})
846 return today + delta
847 return datetime.datetime.strptime(date_str, "%Y%m%d").date()
848
849
850 def hyphenate_date(date_str):
851 """
852 Convert a date in 'YYYYMMDD' format to 'YYYY-MM-DD' format"""
853 match = re.match(r'^(\d\d\d\d)(\d\d)(\d\d)$', date_str)
854 if match is not None:
855 return '-'.join(match.groups())
856 else:
857 return date_str
858
859
860 class DateRange(object):
861 """Represents a time interval between two dates"""
862
863 def __init__(self, start=None, end=None):
864 """start and end must be strings in the format accepted by date"""
865 if start is not None:
866 self.start = date_from_str(start)
867 else:
868 self.start = datetime.datetime.min.date()
869 if end is not None:
870 self.end = date_from_str(end)
871 else:
872 self.end = datetime.datetime.max.date()
873 if self.start > self.end:
874 raise ValueError('Date range: "%s" , the start date must be before the end date' % self)
875
876 @classmethod
877 def day(cls, day):
878 """Returns a range that only contains the given day"""
879 return cls(day, day)
880
881 def __contains__(self, date):
882 """Check if the date is in the range"""
883 if not isinstance(date, datetime.date):
884 date = date_from_str(date)
885 return self.start <= date <= self.end
886
887 def __str__(self):
888 return '%s - %s' % (self.start.isoformat(), self.end.isoformat())
889
890
891 def platform_name():
892 """ Returns the platform name as a compat_str """
893 res = platform.platform()
894 if isinstance(res, bytes):
895 res = res.decode(preferredencoding())
896
897 assert isinstance(res, compat_str)
898 return res
899
900
901 def _windows_write_string(s, out):
902 """ Returns True if the string was written using special methods,
903 False if it has yet to be written out."""
904 # Adapted from http://stackoverflow.com/a/3259271/35070
905
906 import ctypes
907 import ctypes.wintypes
908
909 WIN_OUTPUT_IDS = {
910 1: -11,
911 2: -12,
912 }
913
914 try:
915 fileno = out.fileno()
916 except AttributeError:
917 # If the output stream doesn't have a fileno, it's virtual
918 return False
919 except io.UnsupportedOperation:
920 # Some strange Windows pseudo files?
921 return False
922 if fileno not in WIN_OUTPUT_IDS:
923 return False
924
925 GetStdHandle = ctypes.WINFUNCTYPE(
926 ctypes.wintypes.HANDLE, ctypes.wintypes.DWORD)(
927 (b"GetStdHandle", ctypes.windll.kernel32))
928 h = GetStdHandle(WIN_OUTPUT_IDS[fileno])
929
930 WriteConsoleW = ctypes.WINFUNCTYPE(
931 ctypes.wintypes.BOOL, ctypes.wintypes.HANDLE, ctypes.wintypes.LPWSTR,
932 ctypes.wintypes.DWORD, ctypes.POINTER(ctypes.wintypes.DWORD),
933 ctypes.wintypes.LPVOID)((b"WriteConsoleW", ctypes.windll.kernel32))
934 written = ctypes.wintypes.DWORD(0)
935
936 GetFileType = ctypes.WINFUNCTYPE(ctypes.wintypes.DWORD, ctypes.wintypes.DWORD)((b"GetFileType", ctypes.windll.kernel32))
937 FILE_TYPE_CHAR = 0x0002
938 FILE_TYPE_REMOTE = 0x8000
939 GetConsoleMode = ctypes.WINFUNCTYPE(
940 ctypes.wintypes.BOOL, ctypes.wintypes.HANDLE,
941 ctypes.POINTER(ctypes.wintypes.DWORD))(
942 (b"GetConsoleMode", ctypes.windll.kernel32))
943 INVALID_HANDLE_VALUE = ctypes.wintypes.DWORD(-1).value
944
945 def not_a_console(handle):
946 if handle == INVALID_HANDLE_VALUE or handle is None:
947 return True
948 return ((GetFileType(handle) & ~FILE_TYPE_REMOTE) != FILE_TYPE_CHAR or
949 GetConsoleMode(handle, ctypes.byref(ctypes.wintypes.DWORD())) == 0)
950
951 if not_a_console(h):
952 return False
953
954 def next_nonbmp_pos(s):
955 try:
956 return next(i for i, c in enumerate(s) if ord(c) > 0xffff)
957 except StopIteration:
958 return len(s)
959
960 while s:
961 count = min(next_nonbmp_pos(s), 1024)
962
963 ret = WriteConsoleW(
964 h, s, count if count else 2, ctypes.byref(written), None)
965 if ret == 0:
966 raise OSError('Failed to write string')
967 if not count: # We just wrote a non-BMP character
968 assert written.value == 2
969 s = s[1:]
970 else:
971 assert written.value > 0
972 s = s[written.value:]
973 return True
974
975
976 def write_string(s, out=None, encoding=None):
977 if out is None:
978 out = sys.stderr
979 assert type(s) == compat_str
980
981 if sys.platform == 'win32' and encoding is None and hasattr(out, 'fileno'):
982 if _windows_write_string(s, out):
983 return
984
985 if ('b' in getattr(out, 'mode', '') or
986 sys.version_info[0] < 3): # Python 2 lies about mode of sys.stderr
987 byt = s.encode(encoding or preferredencoding(), 'ignore')
988 out.write(byt)
989 elif hasattr(out, 'buffer'):
990 enc = encoding or getattr(out, 'encoding', None) or preferredencoding()
991 byt = s.encode(enc, 'ignore')
992 out.buffer.write(byt)
993 else:
994 out.write(s)
995 out.flush()
996
997
998 def bytes_to_intlist(bs):
999 if not bs:
1000 return []
1001 if isinstance(bs[0], int): # Python 3
1002 return list(bs)
1003 else:
1004 return [ord(c) for c in bs]
1005
1006
1007 def intlist_to_bytes(xs):
1008 if not xs:
1009 return b''
1010 return struct_pack('%dB' % len(xs), *xs)
1011
1012
1013 # Cross-platform file locking
1014 if sys.platform == 'win32':
1015 import ctypes.wintypes
1016 import msvcrt
1017
1018 class OVERLAPPED(ctypes.Structure):
1019 _fields_ = [
1020 ('Internal', ctypes.wintypes.LPVOID),
1021 ('InternalHigh', ctypes.wintypes.LPVOID),
1022 ('Offset', ctypes.wintypes.DWORD),
1023 ('OffsetHigh', ctypes.wintypes.DWORD),
1024 ('hEvent', ctypes.wintypes.HANDLE),
1025 ]
1026
1027 kernel32 = ctypes.windll.kernel32
1028 LockFileEx = kernel32.LockFileEx
1029 LockFileEx.argtypes = [
1030 ctypes.wintypes.HANDLE, # hFile
1031 ctypes.wintypes.DWORD, # dwFlags
1032 ctypes.wintypes.DWORD, # dwReserved
1033 ctypes.wintypes.DWORD, # nNumberOfBytesToLockLow
1034 ctypes.wintypes.DWORD, # nNumberOfBytesToLockHigh
1035 ctypes.POINTER(OVERLAPPED) # Overlapped
1036 ]
1037 LockFileEx.restype = ctypes.wintypes.BOOL
1038 UnlockFileEx = kernel32.UnlockFileEx
1039 UnlockFileEx.argtypes = [
1040 ctypes.wintypes.HANDLE, # hFile
1041 ctypes.wintypes.DWORD, # dwReserved
1042 ctypes.wintypes.DWORD, # nNumberOfBytesToLockLow
1043 ctypes.wintypes.DWORD, # nNumberOfBytesToLockHigh
1044 ctypes.POINTER(OVERLAPPED) # Overlapped
1045 ]
1046 UnlockFileEx.restype = ctypes.wintypes.BOOL
1047 whole_low = 0xffffffff
1048 whole_high = 0x7fffffff
1049
1050 def _lock_file(f, exclusive):
1051 overlapped = OVERLAPPED()
1052 overlapped.Offset = 0
1053 overlapped.OffsetHigh = 0
1054 overlapped.hEvent = 0
1055 f._lock_file_overlapped_p = ctypes.pointer(overlapped)
1056 handle = msvcrt.get_osfhandle(f.fileno())
1057 if not LockFileEx(handle, 0x2 if exclusive else 0x0, 0,
1058 whole_low, whole_high, f._lock_file_overlapped_p):
1059 raise OSError('Locking file failed: %r' % ctypes.FormatError())
1060
1061 def _unlock_file(f):
1062 assert f._lock_file_overlapped_p
1063 handle = msvcrt.get_osfhandle(f.fileno())
1064 if not UnlockFileEx(handle, 0,
1065 whole_low, whole_high, f._lock_file_overlapped_p):
1066 raise OSError('Unlocking file failed: %r' % ctypes.FormatError())
1067
1068 else:
1069 import fcntl
1070
1071 def _lock_file(f, exclusive):
1072 fcntl.flock(f, fcntl.LOCK_EX if exclusive else fcntl.LOCK_SH)
1073
1074 def _unlock_file(f):
1075 fcntl.flock(f, fcntl.LOCK_UN)
1076
1077
1078 class locked_file(object):
1079 def __init__(self, filename, mode, encoding=None):
1080 assert mode in ['r', 'a', 'w']
1081 self.f = io.open(filename, mode, encoding=encoding)
1082 self.mode = mode
1083
1084 def __enter__(self):
1085 exclusive = self.mode != 'r'
1086 try:
1087 _lock_file(self.f, exclusive)
1088 except IOError:
1089 self.f.close()
1090 raise
1091 return self
1092
1093 def __exit__(self, etype, value, traceback):
1094 try:
1095 _unlock_file(self.f)
1096 finally:
1097 self.f.close()
1098
1099 def __iter__(self):
1100 return iter(self.f)
1101
1102 def write(self, *args):
1103 return self.f.write(*args)
1104
1105 def read(self, *args):
1106 return self.f.read(*args)
1107
1108
1109 def get_filesystem_encoding():
1110 encoding = sys.getfilesystemencoding()
1111 return encoding if encoding is not None else 'utf-8'
1112
1113
1114 def shell_quote(args):
1115 quoted_args = []
1116 encoding = get_filesystem_encoding()
1117 for a in args:
1118 if isinstance(a, bytes):
1119 # We may get a filename encoded with 'encodeFilename'
1120 a = a.decode(encoding)
1121 quoted_args.append(pipes.quote(a))
1122 return ' '.join(quoted_args)
1123
1124
1125 def smuggle_url(url, data):
1126 """ Pass additional data in a URL for internal use. """
1127
1128 sdata = compat_urllib_parse.urlencode(
1129 {'__youtubedl_smuggle': json.dumps(data)})
1130 return url + '#' + sdata
1131
1132
1133 def unsmuggle_url(smug_url, default=None):
1134 if '#__youtubedl_smuggle' not in smug_url:
1135 return smug_url, default
1136 url, _, sdata = smug_url.rpartition('#')
1137 jsond = compat_parse_qs(sdata)['__youtubedl_smuggle'][0]
1138 data = json.loads(jsond)
1139 return url, data
1140
1141
1142 def format_bytes(bytes):
1143 if bytes is None:
1144 return 'N/A'
1145 if type(bytes) is str:
1146 bytes = float(bytes)
1147 if bytes == 0.0:
1148 exponent = 0
1149 else:
1150 exponent = int(math.log(bytes, 1024.0))
1151 suffix = ['B', 'KiB', 'MiB', 'GiB', 'TiB', 'PiB', 'EiB', 'ZiB', 'YiB'][exponent]
1152 converted = float(bytes) / float(1024 ** exponent)
1153 return '%.2f%s' % (converted, suffix)
1154
1155
1156 def parse_filesize(s):
1157 if s is None:
1158 return None
1159
1160 # The lower-case forms are of course incorrect and inofficial,
1161 # but we support those too
1162 _UNIT_TABLE = {
1163 'B': 1,
1164 'b': 1,
1165 'KiB': 1024,
1166 'KB': 1000,
1167 'kB': 1024,
1168 'Kb': 1000,
1169 'MiB': 1024 ** 2,
1170 'MB': 1000 ** 2,
1171 'mB': 1024 ** 2,
1172 'Mb': 1000 ** 2,
1173 'GiB': 1024 ** 3,
1174 'GB': 1000 ** 3,
1175 'gB': 1024 ** 3,
1176 'Gb': 1000 ** 3,
1177 'TiB': 1024 ** 4,
1178 'TB': 1000 ** 4,
1179 'tB': 1024 ** 4,
1180 'Tb': 1000 ** 4,
1181 'PiB': 1024 ** 5,
1182 'PB': 1000 ** 5,
1183 'pB': 1024 ** 5,
1184 'Pb': 1000 ** 5,
1185 'EiB': 1024 ** 6,
1186 'EB': 1000 ** 6,
1187 'eB': 1024 ** 6,
1188 'Eb': 1000 ** 6,
1189 'ZiB': 1024 ** 7,
1190 'ZB': 1000 ** 7,
1191 'zB': 1024 ** 7,
1192 'Zb': 1000 ** 7,
1193 'YiB': 1024 ** 8,
1194 'YB': 1000 ** 8,
1195 'yB': 1024 ** 8,
1196 'Yb': 1000 ** 8,
1197 }
1198
1199 units_re = '|'.join(re.escape(u) for u in _UNIT_TABLE)
1200 m = re.match(
1201 r'(?P<num>[0-9]+(?:[,.][0-9]*)?)\s*(?P<unit>%s)' % units_re, s)
1202 if not m:
1203 return None
1204
1205 num_str = m.group('num').replace(',', '.')
1206 mult = _UNIT_TABLE[m.group('unit')]
1207 return int(float(num_str) * mult)
1208
1209
1210 def month_by_name(name):
1211 """ Return the number of a month by (locale-independently) English name """
1212
1213 try:
1214 return ENGLISH_MONTH_NAMES.index(name) + 1
1215 except ValueError:
1216 return None
1217
1218
1219 def month_by_abbreviation(abbrev):
1220 """ Return the number of a month by (locale-independently) English
1221 abbreviations """
1222
1223 try:
1224 return [s[:3] for s in ENGLISH_MONTH_NAMES].index(abbrev) + 1
1225 except ValueError:
1226 return None
1227
1228
1229 def fix_xml_ampersands(xml_str):
1230 """Replace all the '&' by '&amp;' in XML"""
1231 return re.sub(
1232 r'&(?!amp;|lt;|gt;|apos;|quot;|#x[0-9a-fA-F]{,4};|#[0-9]{,4};)',
1233 '&amp;',
1234 xml_str)
1235
1236
1237 def setproctitle(title):
1238 assert isinstance(title, compat_str)
1239 try:
1240 libc = ctypes.cdll.LoadLibrary("libc.so.6")
1241 except OSError:
1242 return
1243 title_bytes = title.encode('utf-8')
1244 buf = ctypes.create_string_buffer(len(title_bytes))
1245 buf.value = title_bytes
1246 try:
1247 libc.prctl(15, buf, 0, 0, 0)
1248 except AttributeError:
1249 return # Strange libc, just skip this
1250
1251
1252 def remove_start(s, start):
1253 if s.startswith(start):
1254 return s[len(start):]
1255 return s
1256
1257
1258 def remove_end(s, end):
1259 if s.endswith(end):
1260 return s[:-len(end)]
1261 return s
1262
1263
1264 def url_basename(url):
1265 path = compat_urlparse.urlparse(url).path
1266 return path.strip('/').split('/')[-1]
1267
1268
1269 class HEADRequest(compat_urllib_request.Request):
1270 def get_method(self):
1271 return "HEAD"
1272
1273
1274 def int_or_none(v, scale=1, default=None, get_attr=None, invscale=1):
1275 if get_attr:
1276 if v is not None:
1277 v = getattr(v, get_attr, None)
1278 if v == '':
1279 v = None
1280 return default if v is None else (int(v) * invscale // scale)
1281
1282
1283 def str_or_none(v, default=None):
1284 return default if v is None else compat_str(v)
1285
1286
1287 def str_to_int(int_str):
1288 """ A more relaxed version of int_or_none """
1289 if int_str is None:
1290 return None
1291 int_str = re.sub(r'[,\.\+]', '', int_str)
1292 return int(int_str)
1293
1294
1295 def float_or_none(v, scale=1, invscale=1, default=None):
1296 return default if v is None else (float(v) * invscale / scale)
1297
1298
1299 def parse_duration(s):
1300 if not isinstance(s, compat_basestring):
1301 return None
1302
1303 s = s.strip()
1304
1305 m = re.match(
1306 r'''(?ix)(?:P?T)?
1307 (?:
1308 (?P<only_mins>[0-9.]+)\s*(?:mins?|minutes?)\s*|
1309 (?P<only_hours>[0-9.]+)\s*(?:hours?)|
1310
1311 \s*(?P<hours_reversed>[0-9]+)\s*(?:[:h]|hours?)\s*(?P<mins_reversed>[0-9]+)\s*(?:[:m]|mins?|minutes?)\s*|
1312 (?:
1313 (?:
1314 (?:(?P<days>[0-9]+)\s*(?:[:d]|days?)\s*)?
1315 (?P<hours>[0-9]+)\s*(?:[:h]|hours?)\s*
1316 )?
1317 (?P<mins>[0-9]+)\s*(?:[:m]|mins?|minutes?)\s*
1318 )?
1319 (?P<secs>[0-9]+)(?P<ms>\.[0-9]+)?\s*(?:s|secs?|seconds?)?
1320 )$''', s)
1321 if not m:
1322 return None
1323 res = 0
1324 if m.group('only_mins'):
1325 return float_or_none(m.group('only_mins'), invscale=60)
1326 if m.group('only_hours'):
1327 return float_or_none(m.group('only_hours'), invscale=60 * 60)
1328 if m.group('secs'):
1329 res += int(m.group('secs'))
1330 if m.group('mins_reversed'):
1331 res += int(m.group('mins_reversed')) * 60
1332 if m.group('mins'):
1333 res += int(m.group('mins')) * 60
1334 if m.group('hours'):
1335 res += int(m.group('hours')) * 60 * 60
1336 if m.group('hours_reversed'):
1337 res += int(m.group('hours_reversed')) * 60 * 60
1338 if m.group('days'):
1339 res += int(m.group('days')) * 24 * 60 * 60
1340 if m.group('ms'):
1341 res += float(m.group('ms'))
1342 return res
1343
1344
1345 def prepend_extension(filename, ext, expected_real_ext=None):
1346 name, real_ext = os.path.splitext(filename)
1347 return (
1348 '{0}.{1}{2}'.format(name, ext, real_ext)
1349 if not expected_real_ext or real_ext[1:] == expected_real_ext
1350 else '{0}.{1}'.format(filename, ext))
1351
1352
1353 def replace_extension(filename, ext, expected_real_ext=None):
1354 name, real_ext = os.path.splitext(filename)
1355 return '{0}.{1}'.format(
1356 name if not expected_real_ext or real_ext[1:] == expected_real_ext else filename,
1357 ext)
1358
1359
1360 def check_executable(exe, args=[]):
1361 """ Checks if the given binary is installed somewhere in PATH, and returns its name.
1362 args can be a list of arguments for a short output (like -version) """
1363 try:
1364 subprocess.Popen([exe] + args, stdout=subprocess.PIPE, stderr=subprocess.PIPE).communicate()
1365 except OSError:
1366 return False
1367 return exe
1368
1369
1370 def get_exe_version(exe, args=['--version'],
1371 version_re=None, unrecognized='present'):
1372 """ Returns the version of the specified executable,
1373 or False if the executable is not present """
1374 try:
1375 out, _ = subprocess.Popen(
1376 [encodeArgument(exe)] + args,
1377 stdout=subprocess.PIPE, stderr=subprocess.STDOUT).communicate()
1378 except OSError:
1379 return False
1380 if isinstance(out, bytes): # Python 2.x
1381 out = out.decode('ascii', 'ignore')
1382 return detect_exe_version(out, version_re, unrecognized)
1383
1384
1385 def detect_exe_version(output, version_re=None, unrecognized='present'):
1386 assert isinstance(output, compat_str)
1387 if version_re is None:
1388 version_re = r'version\s+([-0-9._a-zA-Z]+)'
1389 m = re.search(version_re, output)
1390 if m:
1391 return m.group(1)
1392 else:
1393 return unrecognized
1394
1395
1396 class PagedList(object):
1397 def __len__(self):
1398 # This is only useful for tests
1399 return len(self.getslice())
1400
1401
1402 class OnDemandPagedList(PagedList):
1403 def __init__(self, pagefunc, pagesize):
1404 self._pagefunc = pagefunc
1405 self._pagesize = pagesize
1406
1407 def getslice(self, start=0, end=None):
1408 res = []
1409 for pagenum in itertools.count(start // self._pagesize):
1410 firstid = pagenum * self._pagesize
1411 nextfirstid = pagenum * self._pagesize + self._pagesize
1412 if start >= nextfirstid:
1413 continue
1414
1415 page_results = list(self._pagefunc(pagenum))
1416
1417 startv = (
1418 start % self._pagesize
1419 if firstid <= start < nextfirstid
1420 else 0)
1421
1422 endv = (
1423 ((end - 1) % self._pagesize) + 1
1424 if (end is not None and firstid <= end <= nextfirstid)
1425 else None)
1426
1427 if startv != 0 or endv is not None:
1428 page_results = page_results[startv:endv]
1429 res.extend(page_results)
1430
1431 # A little optimization - if current page is not "full", ie. does
1432 # not contain page_size videos then we can assume that this page
1433 # is the last one - there are no more ids on further pages -
1434 # i.e. no need to query again.
1435 if len(page_results) + startv < self._pagesize:
1436 break
1437
1438 # If we got the whole page, but the next page is not interesting,
1439 # break out early as well
1440 if end == nextfirstid:
1441 break
1442 return res
1443
1444
1445 class InAdvancePagedList(PagedList):
1446 def __init__(self, pagefunc, pagecount, pagesize):
1447 self._pagefunc = pagefunc
1448 self._pagecount = pagecount
1449 self._pagesize = pagesize
1450
1451 def getslice(self, start=0, end=None):
1452 res = []
1453 start_page = start // self._pagesize
1454 end_page = (
1455 self._pagecount if end is None else (end // self._pagesize + 1))
1456 skip_elems = start - start_page * self._pagesize
1457 only_more = None if end is None else end - start
1458 for pagenum in range(start_page, end_page):
1459 page = list(self._pagefunc(pagenum))
1460 if skip_elems:
1461 page = page[skip_elems:]
1462 skip_elems = None
1463 if only_more is not None:
1464 if len(page) < only_more:
1465 only_more -= len(page)
1466 else:
1467 page = page[:only_more]
1468 res.extend(page)
1469 break
1470 res.extend(page)
1471 return res
1472
1473
1474 def uppercase_escape(s):
1475 unicode_escape = codecs.getdecoder('unicode_escape')
1476 return re.sub(
1477 r'\\U[0-9a-fA-F]{8}',
1478 lambda m: unicode_escape(m.group(0))[0],
1479 s)
1480
1481
1482 def lowercase_escape(s):
1483 unicode_escape = codecs.getdecoder('unicode_escape')
1484 return re.sub(
1485 r'\\u[0-9a-fA-F]{4}',
1486 lambda m: unicode_escape(m.group(0))[0],
1487 s)
1488
1489
1490 def escape_rfc3986(s):
1491 """Escape non-ASCII characters as suggested by RFC 3986"""
1492 if sys.version_info < (3, 0) and isinstance(s, compat_str):
1493 s = s.encode('utf-8')
1494 return compat_urllib_parse.quote(s, b"%/;:@&=+$,!~*'()?#[]")
1495
1496
1497 def escape_url(url):
1498 """Escape URL as suggested by RFC 3986"""
1499 url_parsed = compat_urllib_parse_urlparse(url)
1500 return url_parsed._replace(
1501 path=escape_rfc3986(url_parsed.path),
1502 params=escape_rfc3986(url_parsed.params),
1503 query=escape_rfc3986(url_parsed.query),
1504 fragment=escape_rfc3986(url_parsed.fragment)
1505 ).geturl()
1506
1507 try:
1508 struct.pack('!I', 0)
1509 except TypeError:
1510 # In Python 2.6 (and some 2.7 versions), struct requires a bytes argument
1511 def struct_pack(spec, *args):
1512 if isinstance(spec, compat_str):
1513 spec = spec.encode('ascii')
1514 return struct.pack(spec, *args)
1515
1516 def struct_unpack(spec, *args):
1517 if isinstance(spec, compat_str):
1518 spec = spec.encode('ascii')
1519 return struct.unpack(spec, *args)
1520 else:
1521 struct_pack = struct.pack
1522 struct_unpack = struct.unpack
1523
1524
1525 def read_batch_urls(batch_fd):
1526 def fixup(url):
1527 if not isinstance(url, compat_str):
1528 url = url.decode('utf-8', 'replace')
1529 BOM_UTF8 = '\xef\xbb\xbf'
1530 if url.startswith(BOM_UTF8):
1531 url = url[len(BOM_UTF8):]
1532 url = url.strip()
1533 if url.startswith(('#', ';', ']')):
1534 return False
1535 return url
1536
1537 with contextlib.closing(batch_fd) as fd:
1538 return [url for url in map(fixup, fd) if url]
1539
1540
1541 def urlencode_postdata(*args, **kargs):
1542 return compat_urllib_parse.urlencode(*args, **kargs).encode('ascii')
1543
1544
1545 try:
1546 etree_iter = xml.etree.ElementTree.Element.iter
1547 except AttributeError: # Python <=2.6
1548 etree_iter = lambda n: n.findall('.//*')
1549
1550
1551 def parse_xml(s):
1552 class TreeBuilder(xml.etree.ElementTree.TreeBuilder):
1553 def doctype(self, name, pubid, system):
1554 pass # Ignore doctypes
1555
1556 parser = xml.etree.ElementTree.XMLParser(target=TreeBuilder())
1557 kwargs = {'parser': parser} if sys.version_info >= (2, 7) else {}
1558 tree = xml.etree.ElementTree.XML(s.encode('utf-8'), **kwargs)
1559 # Fix up XML parser in Python 2.x
1560 if sys.version_info < (3, 0):
1561 for n in etree_iter(tree):
1562 if n.text is not None:
1563 if not isinstance(n.text, compat_str):
1564 n.text = n.text.decode('utf-8')
1565 return tree
1566
1567
1568 US_RATINGS = {
1569 'G': 0,
1570 'PG': 10,
1571 'PG-13': 13,
1572 'R': 16,
1573 'NC': 18,
1574 }
1575
1576
1577 def parse_age_limit(s):
1578 if s is None:
1579 return None
1580 m = re.match(r'^(?P<age>\d{1,2})\+?$', s)
1581 return int(m.group('age')) if m else US_RATINGS.get(s, None)
1582
1583
1584 def strip_jsonp(code):
1585 return re.sub(
1586 r'(?s)^[a-zA-Z0-9_]+\s*\(\s*(.*)\);?\s*?(?://[^\n]*)*$', r'\1', code)
1587
1588
1589 def js_to_json(code):
1590 def fix_kv(m):
1591 v = m.group(0)
1592 if v in ('true', 'false', 'null'):
1593 return v
1594 if v.startswith('"'):
1595 return v
1596 if v.startswith("'"):
1597 v = v[1:-1]
1598 v = re.sub(r"\\\\|\\'|\"", lambda m: {
1599 '\\\\': '\\\\',
1600 "\\'": "'",
1601 '"': '\\"',
1602 }[m.group(0)], v)
1603 return '"%s"' % v
1604
1605 res = re.sub(r'''(?x)
1606 "(?:[^"\\]*(?:\\\\|\\['"nu]))*[^"\\]*"|
1607 '(?:[^'\\]*(?:\\\\|\\['"nu]))*[^'\\]*'|
1608 [a-zA-Z_][.a-zA-Z_0-9]*
1609 ''', fix_kv, code)
1610 res = re.sub(r',(\s*[\]}])', lambda m: m.group(1), res)
1611 return res
1612
1613
1614 def qualities(quality_ids):
1615 """ Get a numeric quality value out of a list of possible values """
1616 def q(qid):
1617 try:
1618 return quality_ids.index(qid)
1619 except ValueError:
1620 return -1
1621 return q
1622
1623
1624 DEFAULT_OUTTMPL = '%(title)s-%(id)s.%(ext)s'
1625
1626
1627 def limit_length(s, length):
1628 """ Add ellipses to overly long strings """
1629 if s is None:
1630 return None
1631 ELLIPSES = '...'
1632 if len(s) > length:
1633 return s[:length - len(ELLIPSES)] + ELLIPSES
1634 return s
1635
1636
1637 def version_tuple(v):
1638 return tuple(int(e) for e in re.split(r'[-.]', v))
1639
1640
1641 def is_outdated_version(version, limit, assume_new=True):
1642 if not version:
1643 return not assume_new
1644 try:
1645 return version_tuple(version) < version_tuple(limit)
1646 except ValueError:
1647 return not assume_new
1648
1649
1650 def ytdl_is_updateable():
1651 """ Returns if youtube-dl can be updated with -U """
1652 from zipimport import zipimporter
1653
1654 return isinstance(globals().get('__loader__'), zipimporter) or hasattr(sys, 'frozen')
1655
1656
1657 def args_to_str(args):
1658 # Get a short string representation for a subprocess command
1659 return ' '.join(shlex_quote(a) for a in args)
1660
1661
1662 def mimetype2ext(mt):
1663 _, _, res = mt.rpartition('/')
1664
1665 return {
1666 'x-ms-wmv': 'wmv',
1667 'x-mp4-fragmented': 'mp4',
1668 'ttml+xml': 'ttml',
1669 }.get(res, res)
1670
1671
1672 def urlhandle_detect_ext(url_handle):
1673 try:
1674 url_handle.headers
1675 getheader = lambda h: url_handle.headers[h]
1676 except AttributeError: # Python < 3
1677 getheader = url_handle.info().getheader
1678
1679 cd = getheader('Content-Disposition')
1680 if cd:
1681 m = re.match(r'attachment;\s*filename="(?P<filename>[^"]+)"', cd)
1682 if m:
1683 e = determine_ext(m.group('filename'), default_ext=None)
1684 if e:
1685 return e
1686
1687 return mimetype2ext(getheader('Content-Type'))
1688
1689
1690 def age_restricted(content_limit, age_limit):
1691 """ Returns True iff the content should be blocked """
1692
1693 if age_limit is None: # No limit set
1694 return False
1695 if content_limit is None:
1696 return False # Content available for everyone
1697 return age_limit < content_limit
1698
1699
1700 def is_html(first_bytes):
1701 """ Detect whether a file contains HTML by examining its first bytes. """
1702
1703 BOMS = [
1704 (b'\xef\xbb\xbf', 'utf-8'),
1705 (b'\x00\x00\xfe\xff', 'utf-32-be'),
1706 (b'\xff\xfe\x00\x00', 'utf-32-le'),
1707 (b'\xff\xfe', 'utf-16-le'),
1708 (b'\xfe\xff', 'utf-16-be'),
1709 ]
1710 for bom, enc in BOMS:
1711 if first_bytes.startswith(bom):
1712 s = first_bytes[len(bom):].decode(enc, 'replace')
1713 break
1714 else:
1715 s = first_bytes.decode('utf-8', 'replace')
1716
1717 return re.match(r'^\s*<', s)
1718
1719
1720 def determine_protocol(info_dict):
1721 protocol = info_dict.get('protocol')
1722 if protocol is not None:
1723 return protocol
1724
1725 url = info_dict['url']
1726 if url.startswith('rtmp'):
1727 return 'rtmp'
1728 elif url.startswith('mms'):
1729 return 'mms'
1730 elif url.startswith('rtsp'):
1731 return 'rtsp'
1732
1733 ext = determine_ext(url)
1734 if ext == 'm3u8':
1735 return 'm3u8'
1736 elif ext == 'f4m':
1737 return 'f4m'
1738
1739 return compat_urllib_parse_urlparse(url).scheme
1740
1741
1742 def render_table(header_row, data):
1743 """ Render a list of rows, each as a list of values """
1744 table = [header_row] + data
1745 max_lens = [max(len(compat_str(v)) for v in col) for col in zip(*table)]
1746 format_str = ' '.join('%-' + compat_str(ml + 1) + 's' for ml in max_lens[:-1]) + '%s'
1747 return '\n'.join(format_str % tuple(row) for row in table)
1748
1749
1750 def _match_one(filter_part, dct):
1751 COMPARISON_OPERATORS = {
1752 '<': operator.lt,
1753 '<=': operator.le,
1754 '>': operator.gt,
1755 '>=': operator.ge,
1756 '=': operator.eq,
1757 '!=': operator.ne,
1758 }
1759 operator_rex = re.compile(r'''(?x)\s*
1760 (?P<key>[a-z_]+)
1761 \s*(?P<op>%s)(?P<none_inclusive>\s*\?)?\s*
1762 (?:
1763 (?P<intval>[0-9.]+(?:[kKmMgGtTpPeEzZyY]i?[Bb]?)?)|
1764 (?P<strval>(?![0-9.])[a-z0-9A-Z]*)
1765 )
1766 \s*$
1767 ''' % '|'.join(map(re.escape, COMPARISON_OPERATORS.keys())))
1768 m = operator_rex.search(filter_part)
1769 if m:
1770 op = COMPARISON_OPERATORS[m.group('op')]
1771 if m.group('strval') is not None:
1772 if m.group('op') not in ('=', '!='):
1773 raise ValueError(
1774 'Operator %s does not support string values!' % m.group('op'))
1775 comparison_value = m.group('strval')
1776 else:
1777 try:
1778 comparison_value = int(m.group('intval'))
1779 except ValueError:
1780 comparison_value = parse_filesize(m.group('intval'))
1781 if comparison_value is None:
1782 comparison_value = parse_filesize(m.group('intval') + 'B')
1783 if comparison_value is None:
1784 raise ValueError(
1785 'Invalid integer value %r in filter part %r' % (
1786 m.group('intval'), filter_part))
1787 actual_value = dct.get(m.group('key'))
1788 if actual_value is None:
1789 return m.group('none_inclusive')
1790 return op(actual_value, comparison_value)
1791
1792 UNARY_OPERATORS = {
1793 '': lambda v: v is not None,
1794 '!': lambda v: v is None,
1795 }
1796 operator_rex = re.compile(r'''(?x)\s*
1797 (?P<op>%s)\s*(?P<key>[a-z_]+)
1798 \s*$
1799 ''' % '|'.join(map(re.escape, UNARY_OPERATORS.keys())))
1800 m = operator_rex.search(filter_part)
1801 if m:
1802 op = UNARY_OPERATORS[m.group('op')]
1803 actual_value = dct.get(m.group('key'))
1804 return op(actual_value)
1805
1806 raise ValueError('Invalid filter part %r' % filter_part)
1807
1808
1809 def match_str(filter_str, dct):
1810 """ Filter a dictionary with a simple string syntax. Returns True (=passes filter) or false """
1811
1812 return all(
1813 _match_one(filter_part, dct) for filter_part in filter_str.split('&'))
1814
1815
1816 def match_filter_func(filter_str):
1817 def _match_func(info_dict):
1818 if match_str(filter_str, info_dict):
1819 return None
1820 else:
1821 video_title = info_dict.get('title', info_dict.get('id', 'video'))
1822 return '%s does not pass filter %s, skipping ..' % (video_title, filter_str)
1823 return _match_func
1824
1825
1826 def parse_dfxp_time_expr(time_expr):
1827 if not time_expr:
1828 return 0.0
1829
1830 mobj = re.match(r'^(?P<time_offset>\d+(?:\.\d+)?)s?$', time_expr)
1831 if mobj:
1832 return float(mobj.group('time_offset'))
1833
1834 mobj = re.match(r'^(\d+):(\d\d):(\d\d(?:\.\d+)?)$', time_expr)
1835 if mobj:
1836 return 3600 * int(mobj.group(1)) + 60 * int(mobj.group(2)) + float(mobj.group(3))
1837
1838
1839 def srt_subtitles_timecode(seconds):
1840 return '%02d:%02d:%02d,%03d' % (seconds / 3600, (seconds % 3600) / 60, seconds % 60, (seconds % 1) * 1000)
1841
1842
1843 def dfxp2srt(dfxp_data):
1844 _x = functools.partial(xpath_with_ns, ns_map={'ttml': 'http://www.w3.org/ns/ttml'})
1845
1846 def parse_node(node):
1847 str_or_empty = functools.partial(str_or_none, default='')
1848
1849 out = str_or_empty(node.text)
1850
1851 for child in node:
1852 if child.tag in (_x('ttml:br'), 'br'):
1853 out += '\n' + str_or_empty(child.tail)
1854 elif child.tag in (_x('ttml:span'), 'span'):
1855 out += str_or_empty(parse_node(child))
1856 else:
1857 out += str_or_empty(xml.etree.ElementTree.tostring(child))
1858
1859 return out
1860
1861 dfxp = xml.etree.ElementTree.fromstring(dfxp_data.encode('utf-8'))
1862 out = []
1863 paras = dfxp.findall(_x('.//ttml:p')) or dfxp.findall('.//p')
1864
1865 if not paras:
1866 raise ValueError('Invalid dfxp/TTML subtitle')
1867
1868 for para, index in zip(paras, itertools.count(1)):
1869 begin_time = parse_dfxp_time_expr(para.attrib['begin'])
1870 end_time = parse_dfxp_time_expr(para.attrib.get('end'))
1871 if not end_time:
1872 end_time = begin_time + parse_dfxp_time_expr(para.attrib['dur'])
1873 out.append('%d\n%s --> %s\n%s\n\n' % (
1874 index,
1875 srt_subtitles_timecode(begin_time),
1876 srt_subtitles_timecode(end_time),
1877 parse_node(para)))
1878
1879 return ''.join(out)
1880
1881
1882 class PerRequestProxyHandler(compat_urllib_request.ProxyHandler):
1883 def __init__(self, proxies=None):
1884 # Set default handlers
1885 for type in ('http', 'https'):
1886 setattr(self, '%s_open' % type,
1887 lambda r, proxy='__noproxy__', type=type, meth=self.proxy_open:
1888 meth(r, proxy, type))
1889 return compat_urllib_request.ProxyHandler.__init__(self, proxies)
1890
1891 def proxy_open(self, req, proxy, type):
1892 req_proxy = req.headers.get('Ytdl-request-proxy')
1893 if req_proxy is not None:
1894 proxy = req_proxy
1895 del req.headers['Ytdl-request-proxy']
1896
1897 if proxy == '__noproxy__':
1898 return None # No Proxy
1899 return compat_urllib_request.ProxyHandler.proxy_open(
1900 self, req, proxy, type)