]> Raphaël G. Git Repositories - youtubedl/blob - youtube_dl/utils.py
6a3199fb992b72e70b6588d3999e9c6ec0a87890
[youtubedl] / youtube_dl / utils.py
1 #!/usr/bin/env python
2 # coding: utf-8
3
4 from __future__ import unicode_literals
5
6 import base64
7 import binascii
8 import calendar
9 import codecs
10 import contextlib
11 import ctypes
12 import datetime
13 import email.utils
14 import email.header
15 import errno
16 import functools
17 import gzip
18 import io
19 import itertools
20 import json
21 import locale
22 import math
23 import operator
24 import os
25 import platform
26 import random
27 import re
28 import socket
29 import ssl
30 import subprocess
31 import sys
32 import tempfile
33 import traceback
34 import xml.etree.ElementTree
35 import zlib
36
37 from .compat import (
38 compat_HTMLParseError,
39 compat_HTMLParser,
40 compat_basestring,
41 compat_chr,
42 compat_ctypes_WINFUNCTYPE,
43 compat_etree_fromstring,
44 compat_expanduser,
45 compat_html_entities,
46 compat_html_entities_html5,
47 compat_http_client,
48 compat_kwargs,
49 compat_os_name,
50 compat_parse_qs,
51 compat_shlex_quote,
52 compat_socket_create_connection,
53 compat_str,
54 compat_struct_pack,
55 compat_struct_unpack,
56 compat_urllib_error,
57 compat_urllib_parse,
58 compat_urllib_parse_urlencode,
59 compat_urllib_parse_urlparse,
60 compat_urllib_parse_unquote_plus,
61 compat_urllib_request,
62 compat_urlparse,
63 compat_xpath,
64 )
65
66 from .socks import (
67 ProxyType,
68 sockssocket,
69 )
70
71
72 def register_socks_protocols():
73 # "Register" SOCKS protocols
74 # In Python < 2.6.5, urlsplit() suffers from bug https://bugs.python.org/issue7904
75 # URLs with protocols not in urlparse.uses_netloc are not handled correctly
76 for scheme in ('socks', 'socks4', 'socks4a', 'socks5'):
77 if scheme not in compat_urlparse.uses_netloc:
78 compat_urlparse.uses_netloc.append(scheme)
79
80
81 # This is not clearly defined otherwise
82 compiled_regex_type = type(re.compile(''))
83
84 std_headers = {
85 'User-Agent': 'Mozilla/5.0 (X11; Linux x86_64; rv:59.0) Gecko/20100101 Firefox/59.0 (Chrome)',
86 'Accept-Charset': 'ISO-8859-1,utf-8;q=0.7,*;q=0.7',
87 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
88 'Accept-Encoding': 'gzip, deflate',
89 'Accept-Language': 'en-us,en;q=0.5',
90 }
91
92
93 USER_AGENTS = {
94 'Safari': 'Mozilla/5.0 (X11; Linux x86_64; rv:10.0) AppleWebKit/533.20.25 (KHTML, like Gecko) Version/5.0.4 Safari/533.20.27',
95 }
96
97
98 NO_DEFAULT = object()
99
100 ENGLISH_MONTH_NAMES = [
101 'January', 'February', 'March', 'April', 'May', 'June',
102 'July', 'August', 'September', 'October', 'November', 'December']
103
104 MONTH_NAMES = {
105 'en': ENGLISH_MONTH_NAMES,
106 'fr': [
107 'janvier', 'février', 'mars', 'avril', 'mai', 'juin',
108 'juillet', 'août', 'septembre', 'octobre', 'novembre', 'décembre'],
109 }
110
111 KNOWN_EXTENSIONS = (
112 'mp4', 'm4a', 'm4p', 'm4b', 'm4r', 'm4v', 'aac',
113 'flv', 'f4v', 'f4a', 'f4b',
114 'webm', 'ogg', 'ogv', 'oga', 'ogx', 'spx', 'opus',
115 'mkv', 'mka', 'mk3d',
116 'avi', 'divx',
117 'mov',
118 'asf', 'wmv', 'wma',
119 '3gp', '3g2',
120 'mp3',
121 'flac',
122 'ape',
123 'wav',
124 'f4f', 'f4m', 'm3u8', 'smil')
125
126 # needed for sanitizing filenames in restricted mode
127 ACCENT_CHARS = dict(zip('ÂÃÄÀÁÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖŐØŒÙÚÛÜŰÝÞßàáâãäåæçèéêëìíîïðñòóôõöőøœùúûüűýþÿ',
128 itertools.chain('AAAAAA', ['AE'], 'CEEEEIIIIDNOOOOOOO', ['OE'], 'UUUUUYP', ['ss'],
129 'aaaaaa', ['ae'], 'ceeeeiiiionooooooo', ['oe'], 'uuuuuypy')))
130
131 DATE_FORMATS = (
132 '%d %B %Y',
133 '%d %b %Y',
134 '%B %d %Y',
135 '%B %dst %Y',
136 '%B %dnd %Y',
137 '%B %dth %Y',
138 '%b %d %Y',
139 '%b %dst %Y',
140 '%b %dnd %Y',
141 '%b %dth %Y',
142 '%b %dst %Y %I:%M',
143 '%b %dnd %Y %I:%M',
144 '%b %dth %Y %I:%M',
145 '%Y %m %d',
146 '%Y-%m-%d',
147 '%Y/%m/%d',
148 '%Y/%m/%d %H:%M',
149 '%Y/%m/%d %H:%M:%S',
150 '%Y-%m-%d %H:%M',
151 '%Y-%m-%d %H:%M:%S',
152 '%Y-%m-%d %H:%M:%S.%f',
153 '%d.%m.%Y %H:%M',
154 '%d.%m.%Y %H.%M',
155 '%Y-%m-%dT%H:%M:%SZ',
156 '%Y-%m-%dT%H:%M:%S.%fZ',
157 '%Y-%m-%dT%H:%M:%S.%f0Z',
158 '%Y-%m-%dT%H:%M:%S',
159 '%Y-%m-%dT%H:%M:%S.%f',
160 '%Y-%m-%dT%H:%M',
161 '%b %d %Y at %H:%M',
162 '%b %d %Y at %H:%M:%S',
163 '%B %d %Y at %H:%M',
164 '%B %d %Y at %H:%M:%S',
165 )
166
167 DATE_FORMATS_DAY_FIRST = list(DATE_FORMATS)
168 DATE_FORMATS_DAY_FIRST.extend([
169 '%d-%m-%Y',
170 '%d.%m.%Y',
171 '%d.%m.%y',
172 '%d/%m/%Y',
173 '%d/%m/%y',
174 '%d/%m/%Y %H:%M:%S',
175 ])
176
177 DATE_FORMATS_MONTH_FIRST = list(DATE_FORMATS)
178 DATE_FORMATS_MONTH_FIRST.extend([
179 '%m-%d-%Y',
180 '%m.%d.%Y',
181 '%m/%d/%Y',
182 '%m/%d/%y',
183 '%m/%d/%Y %H:%M:%S',
184 ])
185
186 PACKED_CODES_RE = r"}\('(.+)',(\d+),(\d+),'([^']+)'\.split\('\|'\)"
187
188
189 def preferredencoding():
190 """Get preferred encoding.
191
192 Returns the best encoding scheme for the system, based on
193 locale.getpreferredencoding() and some further tweaks.
194 """
195 try:
196 pref = locale.getpreferredencoding()
197 'TEST'.encode(pref)
198 except Exception:
199 pref = 'UTF-8'
200
201 return pref
202
203
204 def write_json_file(obj, fn):
205 """ Encode obj as JSON and write it to fn, atomically if possible """
206
207 fn = encodeFilename(fn)
208 if sys.version_info < (3, 0) and sys.platform != 'win32':
209 encoding = get_filesystem_encoding()
210 # os.path.basename returns a bytes object, but NamedTemporaryFile
211 # will fail if the filename contains non ascii characters unless we
212 # use a unicode object
213 path_basename = lambda f: os.path.basename(fn).decode(encoding)
214 # the same for os.path.dirname
215 path_dirname = lambda f: os.path.dirname(fn).decode(encoding)
216 else:
217 path_basename = os.path.basename
218 path_dirname = os.path.dirname
219
220 args = {
221 'suffix': '.tmp',
222 'prefix': path_basename(fn) + '.',
223 'dir': path_dirname(fn),
224 'delete': False,
225 }
226
227 # In Python 2.x, json.dump expects a bytestream.
228 # In Python 3.x, it writes to a character stream
229 if sys.version_info < (3, 0):
230 args['mode'] = 'wb'
231 else:
232 args.update({
233 'mode': 'w',
234 'encoding': 'utf-8',
235 })
236
237 tf = tempfile.NamedTemporaryFile(**compat_kwargs(args))
238
239 try:
240 with tf:
241 json.dump(obj, tf)
242 if sys.platform == 'win32':
243 # Need to remove existing file on Windows, else os.rename raises
244 # WindowsError or FileExistsError.
245 try:
246 os.unlink(fn)
247 except OSError:
248 pass
249 os.rename(tf.name, fn)
250 except Exception:
251 try:
252 os.remove(tf.name)
253 except OSError:
254 pass
255 raise
256
257
258 if sys.version_info >= (2, 7):
259 def find_xpath_attr(node, xpath, key, val=None):
260 """ Find the xpath xpath[@key=val] """
261 assert re.match(r'^[a-zA-Z_-]+$', key)
262 expr = xpath + ('[@%s]' % key if val is None else "[@%s='%s']" % (key, val))
263 return node.find(expr)
264 else:
265 def find_xpath_attr(node, xpath, key, val=None):
266 for f in node.findall(compat_xpath(xpath)):
267 if key not in f.attrib:
268 continue
269 if val is None or f.attrib.get(key) == val:
270 return f
271 return None
272
273 # On python2.6 the xml.etree.ElementTree.Element methods don't support
274 # the namespace parameter
275
276
277 def xpath_with_ns(path, ns_map):
278 components = [c.split(':') for c in path.split('/')]
279 replaced = []
280 for c in components:
281 if len(c) == 1:
282 replaced.append(c[0])
283 else:
284 ns, tag = c
285 replaced.append('{%s}%s' % (ns_map[ns], tag))
286 return '/'.join(replaced)
287
288
289 def xpath_element(node, xpath, name=None, fatal=False, default=NO_DEFAULT):
290 def _find_xpath(xpath):
291 return node.find(compat_xpath(xpath))
292
293 if isinstance(xpath, (str, compat_str)):
294 n = _find_xpath(xpath)
295 else:
296 for xp in xpath:
297 n = _find_xpath(xp)
298 if n is not None:
299 break
300
301 if n is None:
302 if default is not NO_DEFAULT:
303 return default
304 elif fatal:
305 name = xpath if name is None else name
306 raise ExtractorError('Could not find XML element %s' % name)
307 else:
308 return None
309 return n
310
311
312 def xpath_text(node, xpath, name=None, fatal=False, default=NO_DEFAULT):
313 n = xpath_element(node, xpath, name, fatal=fatal, default=default)
314 if n is None or n == default:
315 return n
316 if n.text is None:
317 if default is not NO_DEFAULT:
318 return default
319 elif fatal:
320 name = xpath if name is None else name
321 raise ExtractorError('Could not find XML element\'s text %s' % name)
322 else:
323 return None
324 return n.text
325
326
327 def xpath_attr(node, xpath, key, name=None, fatal=False, default=NO_DEFAULT):
328 n = find_xpath_attr(node, xpath, key)
329 if n is None:
330 if default is not NO_DEFAULT:
331 return default
332 elif fatal:
333 name = '%s[@%s]' % (xpath, key) if name is None else name
334 raise ExtractorError('Could not find XML attribute %s' % name)
335 else:
336 return None
337 return n.attrib[key]
338
339
340 def get_element_by_id(id, html):
341 """Return the content of the tag with the specified ID in the passed HTML document"""
342 return get_element_by_attribute('id', id, html)
343
344
345 def get_element_by_class(class_name, html):
346 """Return the content of the first tag with the specified class in the passed HTML document"""
347 retval = get_elements_by_class(class_name, html)
348 return retval[0] if retval else None
349
350
351 def get_element_by_attribute(attribute, value, html, escape_value=True):
352 retval = get_elements_by_attribute(attribute, value, html, escape_value)
353 return retval[0] if retval else None
354
355
356 def get_elements_by_class(class_name, html):
357 """Return the content of all tags with the specified class in the passed HTML document as a list"""
358 return get_elements_by_attribute(
359 'class', r'[^\'"]*\b%s\b[^\'"]*' % re.escape(class_name),
360 html, escape_value=False)
361
362
363 def get_elements_by_attribute(attribute, value, html, escape_value=True):
364 """Return the content of the tag with the specified attribute in the passed HTML document"""
365
366 value = re.escape(value) if escape_value else value
367
368 retlist = []
369 for m in re.finditer(r'''(?xs)
370 <([a-zA-Z0-9:._-]+)
371 (?:\s+[a-zA-Z0-9:._-]+(?:=[a-zA-Z0-9:._-]*|="[^"]*"|='[^']*'|))*?
372 \s+%s=['"]?%s['"]?
373 (?:\s+[a-zA-Z0-9:._-]+(?:=[a-zA-Z0-9:._-]*|="[^"]*"|='[^']*'|))*?
374 \s*>
375 (?P<content>.*?)
376 </\1>
377 ''' % (re.escape(attribute), value), html):
378 res = m.group('content')
379
380 if res.startswith('"') or res.startswith("'"):
381 res = res[1:-1]
382
383 retlist.append(unescapeHTML(res))
384
385 return retlist
386
387
388 class HTMLAttributeParser(compat_HTMLParser):
389 """Trivial HTML parser to gather the attributes for a single element"""
390 def __init__(self):
391 self.attrs = {}
392 compat_HTMLParser.__init__(self)
393
394 def handle_starttag(self, tag, attrs):
395 self.attrs = dict(attrs)
396
397
398 def extract_attributes(html_element):
399 """Given a string for an HTML element such as
400 <el
401 a="foo" B="bar" c="&98;az" d=boz
402 empty= noval entity="&amp;"
403 sq='"' dq="'"
404 >
405 Decode and return a dictionary of attributes.
406 {
407 'a': 'foo', 'b': 'bar', c: 'baz', d: 'boz',
408 'empty': '', 'noval': None, 'entity': '&',
409 'sq': '"', 'dq': '\''
410 }.
411 NB HTMLParser is stricter in Python 2.6 & 3.2 than in later versions,
412 but the cases in the unit test will work for all of 2.6, 2.7, 3.2-3.5.
413 """
414 parser = HTMLAttributeParser()
415 try:
416 parser.feed(html_element)
417 parser.close()
418 # Older Python may throw HTMLParseError in case of malformed HTML
419 except compat_HTMLParseError:
420 pass
421 return parser.attrs
422
423
424 def clean_html(html):
425 """Clean an HTML snippet into a readable string"""
426
427 if html is None: # Convenience for sanitizing descriptions etc.
428 return html
429
430 # Newline vs <br />
431 html = html.replace('\n', ' ')
432 html = re.sub(r'(?u)\s*<\s*br\s*/?\s*>\s*', '\n', html)
433 html = re.sub(r'(?u)<\s*/\s*p\s*>\s*<\s*p[^>]*>', '\n', html)
434 # Strip html tags
435 html = re.sub('<.*?>', '', html)
436 # Replace html entities
437 html = unescapeHTML(html)
438 return html.strip()
439
440
441 def sanitize_open(filename, open_mode):
442 """Try to open the given filename, and slightly tweak it if this fails.
443
444 Attempts to open the given filename. If this fails, it tries to change
445 the filename slightly, step by step, until it's either able to open it
446 or it fails and raises a final exception, like the standard open()
447 function.
448
449 It returns the tuple (stream, definitive_file_name).
450 """
451 try:
452 if filename == '-':
453 if sys.platform == 'win32':
454 import msvcrt
455 msvcrt.setmode(sys.stdout.fileno(), os.O_BINARY)
456 return (sys.stdout.buffer if hasattr(sys.stdout, 'buffer') else sys.stdout, filename)
457 stream = open(encodeFilename(filename), open_mode)
458 return (stream, filename)
459 except (IOError, OSError) as err:
460 if err.errno in (errno.EACCES,):
461 raise
462
463 # In case of error, try to remove win32 forbidden chars
464 alt_filename = sanitize_path(filename)
465 if alt_filename == filename:
466 raise
467 else:
468 # An exception here should be caught in the caller
469 stream = open(encodeFilename(alt_filename), open_mode)
470 return (stream, alt_filename)
471
472
473 def timeconvert(timestr):
474 """Convert RFC 2822 defined time string into system timestamp"""
475 timestamp = None
476 timetuple = email.utils.parsedate_tz(timestr)
477 if timetuple is not None:
478 timestamp = email.utils.mktime_tz(timetuple)
479 return timestamp
480
481
482 def sanitize_filename(s, restricted=False, is_id=False):
483 """Sanitizes a string so it could be used as part of a filename.
484 If restricted is set, use a stricter subset of allowed characters.
485 Set is_id if this is not an arbitrary string, but an ID that should be kept
486 if possible.
487 """
488 def replace_insane(char):
489 if restricted and char in ACCENT_CHARS:
490 return ACCENT_CHARS[char]
491 if char == '?' or ord(char) < 32 or ord(char) == 127:
492 return ''
493 elif char == '"':
494 return '' if restricted else '\''
495 elif char == ':':
496 return '_-' if restricted else ' -'
497 elif char in '\\/|*<>':
498 return '_'
499 if restricted and (char in '!&\'()[]{}$;`^,#' or char.isspace()):
500 return '_'
501 if restricted and ord(char) > 127:
502 return '_'
503 return char
504
505 # Handle timestamps
506 s = re.sub(r'[0-9]+(?::[0-9]+)+', lambda m: m.group(0).replace(':', '_'), s)
507 result = ''.join(map(replace_insane, s))
508 if not is_id:
509 while '__' in result:
510 result = result.replace('__', '_')
511 result = result.strip('_')
512 # Common case of "Foreign band name - English song title"
513 if restricted and result.startswith('-_'):
514 result = result[2:]
515 if result.startswith('-'):
516 result = '_' + result[len('-'):]
517 result = result.lstrip('.')
518 if not result:
519 result = '_'
520 return result
521
522
523 def sanitize_path(s):
524 """Sanitizes and normalizes path on Windows"""
525 if sys.platform != 'win32':
526 return s
527 drive_or_unc, _ = os.path.splitdrive(s)
528 if sys.version_info < (2, 7) and not drive_or_unc:
529 drive_or_unc, _ = os.path.splitunc(s)
530 norm_path = os.path.normpath(remove_start(s, drive_or_unc)).split(os.path.sep)
531 if drive_or_unc:
532 norm_path.pop(0)
533 sanitized_path = [
534 path_part if path_part in ['.', '..'] else re.sub(r'(?:[/<>:"\|\\?\*]|[\s.]$)', '#', path_part)
535 for path_part in norm_path]
536 if drive_or_unc:
537 sanitized_path.insert(0, drive_or_unc + os.path.sep)
538 return os.path.join(*sanitized_path)
539
540
541 def sanitize_url(url):
542 # Prepend protocol-less URLs with `http:` scheme in order to mitigate
543 # the number of unwanted failures due to missing protocol
544 if url.startswith('//'):
545 return 'http:%s' % url
546 # Fix some common typos seen so far
547 COMMON_TYPOS = (
548 # https://github.com/rg3/youtube-dl/issues/15649
549 (r'^httpss://', r'https://'),
550 # https://bx1.be/lives/direct-tv/
551 (r'^rmtp([es]?)://', r'rtmp\1://'),
552 )
553 for mistake, fixup in COMMON_TYPOS:
554 if re.match(mistake, url):
555 return re.sub(mistake, fixup, url)
556 return url
557
558
559 def sanitized_Request(url, *args, **kwargs):
560 return compat_urllib_request.Request(sanitize_url(url), *args, **kwargs)
561
562
563 def expand_path(s):
564 """Expand shell variables and ~"""
565 return os.path.expandvars(compat_expanduser(s))
566
567
568 def orderedSet(iterable):
569 """ Remove all duplicates from the input iterable """
570 res = []
571 for el in iterable:
572 if el not in res:
573 res.append(el)
574 return res
575
576
577 def _htmlentity_transform(entity_with_semicolon):
578 """Transforms an HTML entity to a character."""
579 entity = entity_with_semicolon[:-1]
580
581 # Known non-numeric HTML entity
582 if entity in compat_html_entities.name2codepoint:
583 return compat_chr(compat_html_entities.name2codepoint[entity])
584
585 # TODO: HTML5 allows entities without a semicolon. For example,
586 # '&Eacuteric' should be decoded as 'Éric'.
587 if entity_with_semicolon in compat_html_entities_html5:
588 return compat_html_entities_html5[entity_with_semicolon]
589
590 mobj = re.match(r'#(x[0-9a-fA-F]+|[0-9]+)', entity)
591 if mobj is not None:
592 numstr = mobj.group(1)
593 if numstr.startswith('x'):
594 base = 16
595 numstr = '0%s' % numstr
596 else:
597 base = 10
598 # See https://github.com/rg3/youtube-dl/issues/7518
599 try:
600 return compat_chr(int(numstr, base))
601 except ValueError:
602 pass
603
604 # Unknown entity in name, return its literal representation
605 return '&%s;' % entity
606
607
608 def unescapeHTML(s):
609 if s is None:
610 return None
611 assert type(s) == compat_str
612
613 return re.sub(
614 r'&([^&;]+;)', lambda m: _htmlentity_transform(m.group(1)), s)
615
616
617 def get_subprocess_encoding():
618 if sys.platform == 'win32' and sys.getwindowsversion()[0] >= 5:
619 # For subprocess calls, encode with locale encoding
620 # Refer to http://stackoverflow.com/a/9951851/35070
621 encoding = preferredencoding()
622 else:
623 encoding = sys.getfilesystemencoding()
624 if encoding is None:
625 encoding = 'utf-8'
626 return encoding
627
628
629 def encodeFilename(s, for_subprocess=False):
630 """
631 @param s The name of the file
632 """
633
634 assert type(s) == compat_str
635
636 # Python 3 has a Unicode API
637 if sys.version_info >= (3, 0):
638 return s
639
640 # Pass '' directly to use Unicode APIs on Windows 2000 and up
641 # (Detecting Windows NT 4 is tricky because 'major >= 4' would
642 # match Windows 9x series as well. Besides, NT 4 is obsolete.)
643 if not for_subprocess and sys.platform == 'win32' and sys.getwindowsversion()[0] >= 5:
644 return s
645
646 # Jython assumes filenames are Unicode strings though reported as Python 2.x compatible
647 if sys.platform.startswith('java'):
648 return s
649
650 return s.encode(get_subprocess_encoding(), 'ignore')
651
652
653 def decodeFilename(b, for_subprocess=False):
654
655 if sys.version_info >= (3, 0):
656 return b
657
658 if not isinstance(b, bytes):
659 return b
660
661 return b.decode(get_subprocess_encoding(), 'ignore')
662
663
664 def encodeArgument(s):
665 if not isinstance(s, compat_str):
666 # Legacy code that uses byte strings
667 # Uncomment the following line after fixing all post processors
668 # assert False, 'Internal error: %r should be of type %r, is %r' % (s, compat_str, type(s))
669 s = s.decode('ascii')
670 return encodeFilename(s, True)
671
672
673 def decodeArgument(b):
674 return decodeFilename(b, True)
675
676
677 def decodeOption(optval):
678 if optval is None:
679 return optval
680 if isinstance(optval, bytes):
681 optval = optval.decode(preferredencoding())
682
683 assert isinstance(optval, compat_str)
684 return optval
685
686
687 def formatSeconds(secs):
688 if secs > 3600:
689 return '%d:%02d:%02d' % (secs // 3600, (secs % 3600) // 60, secs % 60)
690 elif secs > 60:
691 return '%d:%02d' % (secs // 60, secs % 60)
692 else:
693 return '%d' % secs
694
695
696 def make_HTTPS_handler(params, **kwargs):
697 opts_no_check_certificate = params.get('nocheckcertificate', False)
698 if hasattr(ssl, 'create_default_context'): # Python >= 3.4 or 2.7.9
699 context = ssl.create_default_context(ssl.Purpose.SERVER_AUTH)
700 if opts_no_check_certificate:
701 context.check_hostname = False
702 context.verify_mode = ssl.CERT_NONE
703 try:
704 return YoutubeDLHTTPSHandler(params, context=context, **kwargs)
705 except TypeError:
706 # Python 2.7.8
707 # (create_default_context present but HTTPSHandler has no context=)
708 pass
709
710 if sys.version_info < (3, 2):
711 return YoutubeDLHTTPSHandler(params, **kwargs)
712 else: # Python < 3.4
713 context = ssl.SSLContext(ssl.PROTOCOL_TLSv1)
714 context.verify_mode = (ssl.CERT_NONE
715 if opts_no_check_certificate
716 else ssl.CERT_REQUIRED)
717 context.set_default_verify_paths()
718 return YoutubeDLHTTPSHandler(params, context=context, **kwargs)
719
720
721 def bug_reports_message():
722 if ytdl_is_updateable():
723 update_cmd = 'type youtube-dl -U to update'
724 else:
725 update_cmd = 'see https://yt-dl.org/update on how to update'
726 msg = '; please report this issue on https://yt-dl.org/bug .'
727 msg += ' Make sure you are using the latest version; %s.' % update_cmd
728 msg += ' Be sure to call youtube-dl with the --verbose flag and include its complete output.'
729 return msg
730
731
732 class YoutubeDLError(Exception):
733 """Base exception for YoutubeDL errors."""
734 pass
735
736
737 class ExtractorError(YoutubeDLError):
738 """Error during info extraction."""
739
740 def __init__(self, msg, tb=None, expected=False, cause=None, video_id=None):
741 """ tb, if given, is the original traceback (so that it can be printed out).
742 If expected is set, this is a normal error message and most likely not a bug in youtube-dl.
743 """
744
745 if sys.exc_info()[0] in (compat_urllib_error.URLError, socket.timeout, UnavailableVideoError):
746 expected = True
747 if video_id is not None:
748 msg = video_id + ': ' + msg
749 if cause:
750 msg += ' (caused by %r)' % cause
751 if not expected:
752 msg += bug_reports_message()
753 super(ExtractorError, self).__init__(msg)
754
755 self.traceback = tb
756 self.exc_info = sys.exc_info() # preserve original exception
757 self.cause = cause
758 self.video_id = video_id
759
760 def format_traceback(self):
761 if self.traceback is None:
762 return None
763 return ''.join(traceback.format_tb(self.traceback))
764
765
766 class UnsupportedError(ExtractorError):
767 def __init__(self, url):
768 super(UnsupportedError, self).__init__(
769 'Unsupported URL: %s' % url, expected=True)
770 self.url = url
771
772
773 class RegexNotFoundError(ExtractorError):
774 """Error when a regex didn't match"""
775 pass
776
777
778 class GeoRestrictedError(ExtractorError):
779 """Geographic restriction Error exception.
780
781 This exception may be thrown when a video is not available from your
782 geographic location due to geographic restrictions imposed by a website.
783 """
784 def __init__(self, msg, countries=None):
785 super(GeoRestrictedError, self).__init__(msg, expected=True)
786 self.msg = msg
787 self.countries = countries
788
789
790 class DownloadError(YoutubeDLError):
791 """Download Error exception.
792
793 This exception may be thrown by FileDownloader objects if they are not
794 configured to continue on errors. They will contain the appropriate
795 error message.
796 """
797
798 def __init__(self, msg, exc_info=None):
799 """ exc_info, if given, is the original exception that caused the trouble (as returned by sys.exc_info()). """
800 super(DownloadError, self).__init__(msg)
801 self.exc_info = exc_info
802
803
804 class SameFileError(YoutubeDLError):
805 """Same File exception.
806
807 This exception will be thrown by FileDownloader objects if they detect
808 multiple files would have to be downloaded to the same file on disk.
809 """
810 pass
811
812
813 class PostProcessingError(YoutubeDLError):
814 """Post Processing exception.
815
816 This exception may be raised by PostProcessor's .run() method to
817 indicate an error in the postprocessing task.
818 """
819
820 def __init__(self, msg):
821 super(PostProcessingError, self).__init__(msg)
822 self.msg = msg
823
824
825 class MaxDownloadsReached(YoutubeDLError):
826 """ --max-downloads limit has been reached. """
827 pass
828
829
830 class UnavailableVideoError(YoutubeDLError):
831 """Unavailable Format exception.
832
833 This exception will be thrown when a video is requested
834 in a format that is not available for that video.
835 """
836 pass
837
838
839 class ContentTooShortError(YoutubeDLError):
840 """Content Too Short exception.
841
842 This exception may be raised by FileDownloader objects when a file they
843 download is too small for what the server announced first, indicating
844 the connection was probably interrupted.
845 """
846
847 def __init__(self, downloaded, expected):
848 super(ContentTooShortError, self).__init__(
849 'Downloaded {0} bytes, expected {1} bytes'.format(downloaded, expected)
850 )
851 # Both in bytes
852 self.downloaded = downloaded
853 self.expected = expected
854
855
856 class XAttrMetadataError(YoutubeDLError):
857 def __init__(self, code=None, msg='Unknown error'):
858 super(XAttrMetadataError, self).__init__(msg)
859 self.code = code
860 self.msg = msg
861
862 # Parsing code and msg
863 if (self.code in (errno.ENOSPC, errno.EDQUOT) or
864 'No space left' in self.msg or 'Disk quota excedded' in self.msg):
865 self.reason = 'NO_SPACE'
866 elif self.code == errno.E2BIG or 'Argument list too long' in self.msg:
867 self.reason = 'VALUE_TOO_LONG'
868 else:
869 self.reason = 'NOT_SUPPORTED'
870
871
872 class XAttrUnavailableError(YoutubeDLError):
873 pass
874
875
876 def _create_http_connection(ydl_handler, http_class, is_https, *args, **kwargs):
877 # Working around python 2 bug (see http://bugs.python.org/issue17849) by limiting
878 # expected HTTP responses to meet HTTP/1.0 or later (see also
879 # https://github.com/rg3/youtube-dl/issues/6727)
880 if sys.version_info < (3, 0):
881 kwargs['strict'] = True
882 hc = http_class(*args, **compat_kwargs(kwargs))
883 source_address = ydl_handler._params.get('source_address')
884 if source_address is not None:
885 sa = (source_address, 0)
886 if hasattr(hc, 'source_address'): # Python 2.7+
887 hc.source_address = sa
888 else: # Python 2.6
889 def _hc_connect(self, *args, **kwargs):
890 sock = compat_socket_create_connection(
891 (self.host, self.port), self.timeout, sa)
892 if is_https:
893 self.sock = ssl.wrap_socket(
894 sock, self.key_file, self.cert_file,
895 ssl_version=ssl.PROTOCOL_TLSv1)
896 else:
897 self.sock = sock
898 hc.connect = functools.partial(_hc_connect, hc)
899
900 return hc
901
902
903 def handle_youtubedl_headers(headers):
904 filtered_headers = headers
905
906 if 'Youtubedl-no-compression' in filtered_headers:
907 filtered_headers = dict((k, v) for k, v in filtered_headers.items() if k.lower() != 'accept-encoding')
908 del filtered_headers['Youtubedl-no-compression']
909
910 return filtered_headers
911
912
913 class YoutubeDLHandler(compat_urllib_request.HTTPHandler):
914 """Handler for HTTP requests and responses.
915
916 This class, when installed with an OpenerDirector, automatically adds
917 the standard headers to every HTTP request and handles gzipped and
918 deflated responses from web servers. If compression is to be avoided in
919 a particular request, the original request in the program code only has
920 to include the HTTP header "Youtubedl-no-compression", which will be
921 removed before making the real request.
922
923 Part of this code was copied from:
924
925 http://techknack.net/python-urllib2-handlers/
926
927 Andrew Rowls, the author of that code, agreed to release it to the
928 public domain.
929 """
930
931 def __init__(self, params, *args, **kwargs):
932 compat_urllib_request.HTTPHandler.__init__(self, *args, **kwargs)
933 self._params = params
934
935 def http_open(self, req):
936 conn_class = compat_http_client.HTTPConnection
937
938 socks_proxy = req.headers.get('Ytdl-socks-proxy')
939 if socks_proxy:
940 conn_class = make_socks_conn_class(conn_class, socks_proxy)
941 del req.headers['Ytdl-socks-proxy']
942
943 return self.do_open(functools.partial(
944 _create_http_connection, self, conn_class, False),
945 req)
946
947 @staticmethod
948 def deflate(data):
949 try:
950 return zlib.decompress(data, -zlib.MAX_WBITS)
951 except zlib.error:
952 return zlib.decompress(data)
953
954 def http_request(self, req):
955 # According to RFC 3986, URLs can not contain non-ASCII characters, however this is not
956 # always respected by websites, some tend to give out URLs with non percent-encoded
957 # non-ASCII characters (see telemb.py, ard.py [#3412])
958 # urllib chokes on URLs with non-ASCII characters (see http://bugs.python.org/issue3991)
959 # To work around aforementioned issue we will replace request's original URL with
960 # percent-encoded one
961 # Since redirects are also affected (e.g. http://www.southpark.de/alle-episoden/s18e09)
962 # the code of this workaround has been moved here from YoutubeDL.urlopen()
963 url = req.get_full_url()
964 url_escaped = escape_url(url)
965
966 # Substitute URL if any change after escaping
967 if url != url_escaped:
968 req = update_Request(req, url=url_escaped)
969
970 for h, v in std_headers.items():
971 # Capitalize is needed because of Python bug 2275: http://bugs.python.org/issue2275
972 # The dict keys are capitalized because of this bug by urllib
973 if h.capitalize() not in req.headers:
974 req.add_header(h, v)
975
976 req.headers = handle_youtubedl_headers(req.headers)
977
978 if sys.version_info < (2, 7) and '#' in req.get_full_url():
979 # Python 2.6 is brain-dead when it comes to fragments
980 req._Request__original = req._Request__original.partition('#')[0]
981 req._Request__r_type = req._Request__r_type.partition('#')[0]
982
983 return req
984
985 def http_response(self, req, resp):
986 old_resp = resp
987 # gzip
988 if resp.headers.get('Content-encoding', '') == 'gzip':
989 content = resp.read()
990 gz = gzip.GzipFile(fileobj=io.BytesIO(content), mode='rb')
991 try:
992 uncompressed = io.BytesIO(gz.read())
993 except IOError as original_ioerror:
994 # There may be junk add the end of the file
995 # See http://stackoverflow.com/q/4928560/35070 for details
996 for i in range(1, 1024):
997 try:
998 gz = gzip.GzipFile(fileobj=io.BytesIO(content[:-i]), mode='rb')
999 uncompressed = io.BytesIO(gz.read())
1000 except IOError:
1001 continue
1002 break
1003 else:
1004 raise original_ioerror
1005 resp = compat_urllib_request.addinfourl(uncompressed, old_resp.headers, old_resp.url, old_resp.code)
1006 resp.msg = old_resp.msg
1007 del resp.headers['Content-encoding']
1008 # deflate
1009 if resp.headers.get('Content-encoding', '') == 'deflate':
1010 gz = io.BytesIO(self.deflate(resp.read()))
1011 resp = compat_urllib_request.addinfourl(gz, old_resp.headers, old_resp.url, old_resp.code)
1012 resp.msg = old_resp.msg
1013 del resp.headers['Content-encoding']
1014 # Percent-encode redirect URL of Location HTTP header to satisfy RFC 3986 (see
1015 # https://github.com/rg3/youtube-dl/issues/6457).
1016 if 300 <= resp.code < 400:
1017 location = resp.headers.get('Location')
1018 if location:
1019 # As of RFC 2616 default charset is iso-8859-1 that is respected by python 3
1020 if sys.version_info >= (3, 0):
1021 location = location.encode('iso-8859-1').decode('utf-8')
1022 else:
1023 location = location.decode('utf-8')
1024 location_escaped = escape_url(location)
1025 if location != location_escaped:
1026 del resp.headers['Location']
1027 if sys.version_info < (3, 0):
1028 location_escaped = location_escaped.encode('utf-8')
1029 resp.headers['Location'] = location_escaped
1030 return resp
1031
1032 https_request = http_request
1033 https_response = http_response
1034
1035
1036 def make_socks_conn_class(base_class, socks_proxy):
1037 assert issubclass(base_class, (
1038 compat_http_client.HTTPConnection, compat_http_client.HTTPSConnection))
1039
1040 url_components = compat_urlparse.urlparse(socks_proxy)
1041 if url_components.scheme.lower() == 'socks5':
1042 socks_type = ProxyType.SOCKS5
1043 elif url_components.scheme.lower() in ('socks', 'socks4'):
1044 socks_type = ProxyType.SOCKS4
1045 elif url_components.scheme.lower() == 'socks4a':
1046 socks_type = ProxyType.SOCKS4A
1047
1048 def unquote_if_non_empty(s):
1049 if not s:
1050 return s
1051 return compat_urllib_parse_unquote_plus(s)
1052
1053 proxy_args = (
1054 socks_type,
1055 url_components.hostname, url_components.port or 1080,
1056 True, # Remote DNS
1057 unquote_if_non_empty(url_components.username),
1058 unquote_if_non_empty(url_components.password),
1059 )
1060
1061 class SocksConnection(base_class):
1062 def connect(self):
1063 self.sock = sockssocket()
1064 self.sock.setproxy(*proxy_args)
1065 if type(self.timeout) in (int, float):
1066 self.sock.settimeout(self.timeout)
1067 self.sock.connect((self.host, self.port))
1068
1069 if isinstance(self, compat_http_client.HTTPSConnection):
1070 if hasattr(self, '_context'): # Python > 2.6
1071 self.sock = self._context.wrap_socket(
1072 self.sock, server_hostname=self.host)
1073 else:
1074 self.sock = ssl.wrap_socket(self.sock)
1075
1076 return SocksConnection
1077
1078
1079 class YoutubeDLHTTPSHandler(compat_urllib_request.HTTPSHandler):
1080 def __init__(self, params, https_conn_class=None, *args, **kwargs):
1081 compat_urllib_request.HTTPSHandler.__init__(self, *args, **kwargs)
1082 self._https_conn_class = https_conn_class or compat_http_client.HTTPSConnection
1083 self._params = params
1084
1085 def https_open(self, req):
1086 kwargs = {}
1087 conn_class = self._https_conn_class
1088
1089 if hasattr(self, '_context'): # python > 2.6
1090 kwargs['context'] = self._context
1091 if hasattr(self, '_check_hostname'): # python 3.x
1092 kwargs['check_hostname'] = self._check_hostname
1093
1094 socks_proxy = req.headers.get('Ytdl-socks-proxy')
1095 if socks_proxy:
1096 conn_class = make_socks_conn_class(conn_class, socks_proxy)
1097 del req.headers['Ytdl-socks-proxy']
1098
1099 return self.do_open(functools.partial(
1100 _create_http_connection, self, conn_class, True),
1101 req, **kwargs)
1102
1103
1104 class YoutubeDLCookieProcessor(compat_urllib_request.HTTPCookieProcessor):
1105 def __init__(self, cookiejar=None):
1106 compat_urllib_request.HTTPCookieProcessor.__init__(self, cookiejar)
1107
1108 def http_response(self, request, response):
1109 # Python 2 will choke on next HTTP request in row if there are non-ASCII
1110 # characters in Set-Cookie HTTP header of last response (see
1111 # https://github.com/rg3/youtube-dl/issues/6769).
1112 # In order to at least prevent crashing we will percent encode Set-Cookie
1113 # header before HTTPCookieProcessor starts processing it.
1114 # if sys.version_info < (3, 0) and response.headers:
1115 # for set_cookie_header in ('Set-Cookie', 'Set-Cookie2'):
1116 # set_cookie = response.headers.get(set_cookie_header)
1117 # if set_cookie:
1118 # set_cookie_escaped = compat_urllib_parse.quote(set_cookie, b"%/;:@&=+$,!~*'()?#[] ")
1119 # if set_cookie != set_cookie_escaped:
1120 # del response.headers[set_cookie_header]
1121 # response.headers[set_cookie_header] = set_cookie_escaped
1122 return compat_urllib_request.HTTPCookieProcessor.http_response(self, request, response)
1123
1124 https_request = compat_urllib_request.HTTPCookieProcessor.http_request
1125 https_response = http_response
1126
1127
1128 def extract_timezone(date_str):
1129 m = re.search(
1130 r'^.{8,}?(?P<tz>Z$| ?(?P<sign>\+|-)(?P<hours>[0-9]{2}):?(?P<minutes>[0-9]{2})$)',
1131 date_str)
1132 if not m:
1133 timezone = datetime.timedelta()
1134 else:
1135 date_str = date_str[:-len(m.group('tz'))]
1136 if not m.group('sign'):
1137 timezone = datetime.timedelta()
1138 else:
1139 sign = 1 if m.group('sign') == '+' else -1
1140 timezone = datetime.timedelta(
1141 hours=sign * int(m.group('hours')),
1142 minutes=sign * int(m.group('minutes')))
1143 return timezone, date_str
1144
1145
1146 def parse_iso8601(date_str, delimiter='T', timezone=None):
1147 """ Return a UNIX timestamp from the given date """
1148
1149 if date_str is None:
1150 return None
1151
1152 date_str = re.sub(r'\.[0-9]+', '', date_str)
1153
1154 if timezone is None:
1155 timezone, date_str = extract_timezone(date_str)
1156
1157 try:
1158 date_format = '%Y-%m-%d{0}%H:%M:%S'.format(delimiter)
1159 dt = datetime.datetime.strptime(date_str, date_format) - timezone
1160 return calendar.timegm(dt.timetuple())
1161 except ValueError:
1162 pass
1163
1164
1165 def date_formats(day_first=True):
1166 return DATE_FORMATS_DAY_FIRST if day_first else DATE_FORMATS_MONTH_FIRST
1167
1168
1169 def unified_strdate(date_str, day_first=True):
1170 """Return a string with the date in the format YYYYMMDD"""
1171
1172 if date_str is None:
1173 return None
1174 upload_date = None
1175 # Replace commas
1176 date_str = date_str.replace(',', ' ')
1177 # Remove AM/PM + timezone
1178 date_str = re.sub(r'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str)
1179 _, date_str = extract_timezone(date_str)
1180
1181 for expression in date_formats(day_first):
1182 try:
1183 upload_date = datetime.datetime.strptime(date_str, expression).strftime('%Y%m%d')
1184 except ValueError:
1185 pass
1186 if upload_date is None:
1187 timetuple = email.utils.parsedate_tz(date_str)
1188 if timetuple:
1189 try:
1190 upload_date = datetime.datetime(*timetuple[:6]).strftime('%Y%m%d')
1191 except ValueError:
1192 pass
1193 if upload_date is not None:
1194 return compat_str(upload_date)
1195
1196
1197 def unified_timestamp(date_str, day_first=True):
1198 if date_str is None:
1199 return None
1200
1201 date_str = re.sub(r'[,|]', '', date_str)
1202
1203 pm_delta = 12 if re.search(r'(?i)PM', date_str) else 0
1204 timezone, date_str = extract_timezone(date_str)
1205
1206 # Remove AM/PM + timezone
1207 date_str = re.sub(r'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str)
1208
1209 # Remove unrecognized timezones from ISO 8601 alike timestamps
1210 m = re.search(r'\d{1,2}:\d{1,2}(?:\.\d+)?(?P<tz>\s*[A-Z]+)$', date_str)
1211 if m:
1212 date_str = date_str[:-len(m.group('tz'))]
1213
1214 # Python only supports microseconds, so remove nanoseconds
1215 m = re.search(r'^([0-9]{4,}-[0-9]{1,2}-[0-9]{1,2}T[0-9]{1,2}:[0-9]{1,2}:[0-9]{1,2}\.[0-9]{6})[0-9]+$', date_str)
1216 if m:
1217 date_str = m.group(1)
1218
1219 for expression in date_formats(day_first):
1220 try:
1221 dt = datetime.datetime.strptime(date_str, expression) - timezone + datetime.timedelta(hours=pm_delta)
1222 return calendar.timegm(dt.timetuple())
1223 except ValueError:
1224 pass
1225 timetuple = email.utils.parsedate_tz(date_str)
1226 if timetuple:
1227 return calendar.timegm(timetuple) + pm_delta * 3600
1228
1229
1230 def determine_ext(url, default_ext='unknown_video'):
1231 if url is None or '.' not in url:
1232 return default_ext
1233 guess = url.partition('?')[0].rpartition('.')[2]
1234 if re.match(r'^[A-Za-z0-9]+$', guess):
1235 return guess
1236 # Try extract ext from URLs like http://example.com/foo/bar.mp4/?download
1237 elif guess.rstrip('/') in KNOWN_EXTENSIONS:
1238 return guess.rstrip('/')
1239 else:
1240 return default_ext
1241
1242
1243 def subtitles_filename(filename, sub_lang, sub_format):
1244 return filename.rsplit('.', 1)[0] + '.' + sub_lang + '.' + sub_format
1245
1246
1247 def date_from_str(date_str):
1248 """
1249 Return a datetime object from a string in the format YYYYMMDD or
1250 (now|today)[+-][0-9](day|week|month|year)(s)?"""
1251 today = datetime.date.today()
1252 if date_str in ('now', 'today'):
1253 return today
1254 if date_str == 'yesterday':
1255 return today - datetime.timedelta(days=1)
1256 match = re.match(r'(now|today)(?P<sign>[+-])(?P<time>\d+)(?P<unit>day|week|month|year)(s)?', date_str)
1257 if match is not None:
1258 sign = match.group('sign')
1259 time = int(match.group('time'))
1260 if sign == '-':
1261 time = -time
1262 unit = match.group('unit')
1263 # A bad approximation?
1264 if unit == 'month':
1265 unit = 'day'
1266 time *= 30
1267 elif unit == 'year':
1268 unit = 'day'
1269 time *= 365
1270 unit += 's'
1271 delta = datetime.timedelta(**{unit: time})
1272 return today + delta
1273 return datetime.datetime.strptime(date_str, '%Y%m%d').date()
1274
1275
1276 def hyphenate_date(date_str):
1277 """
1278 Convert a date in 'YYYYMMDD' format to 'YYYY-MM-DD' format"""
1279 match = re.match(r'^(\d\d\d\d)(\d\d)(\d\d)$', date_str)
1280 if match is not None:
1281 return '-'.join(match.groups())
1282 else:
1283 return date_str
1284
1285
1286 class DateRange(object):
1287 """Represents a time interval between two dates"""
1288
1289 def __init__(self, start=None, end=None):
1290 """start and end must be strings in the format accepted by date"""
1291 if start is not None:
1292 self.start = date_from_str(start)
1293 else:
1294 self.start = datetime.datetime.min.date()
1295 if end is not None:
1296 self.end = date_from_str(end)
1297 else:
1298 self.end = datetime.datetime.max.date()
1299 if self.start > self.end:
1300 raise ValueError('Date range: "%s" , the start date must be before the end date' % self)
1301
1302 @classmethod
1303 def day(cls, day):
1304 """Returns a range that only contains the given day"""
1305 return cls(day, day)
1306
1307 def __contains__(self, date):
1308 """Check if the date is in the range"""
1309 if not isinstance(date, datetime.date):
1310 date = date_from_str(date)
1311 return self.start <= date <= self.end
1312
1313 def __str__(self):
1314 return '%s - %s' % (self.start.isoformat(), self.end.isoformat())
1315
1316
1317 def platform_name():
1318 """ Returns the platform name as a compat_str """
1319 res = platform.platform()
1320 if isinstance(res, bytes):
1321 res = res.decode(preferredencoding())
1322
1323 assert isinstance(res, compat_str)
1324 return res
1325
1326
1327 def _windows_write_string(s, out):
1328 """ Returns True if the string was written using special methods,
1329 False if it has yet to be written out."""
1330 # Adapted from http://stackoverflow.com/a/3259271/35070
1331
1332 import ctypes
1333 import ctypes.wintypes
1334
1335 WIN_OUTPUT_IDS = {
1336 1: -11,
1337 2: -12,
1338 }
1339
1340 try:
1341 fileno = out.fileno()
1342 except AttributeError:
1343 # If the output stream doesn't have a fileno, it's virtual
1344 return False
1345 except io.UnsupportedOperation:
1346 # Some strange Windows pseudo files?
1347 return False
1348 if fileno not in WIN_OUTPUT_IDS:
1349 return False
1350
1351 GetStdHandle = compat_ctypes_WINFUNCTYPE(
1352 ctypes.wintypes.HANDLE, ctypes.wintypes.DWORD)(
1353 ('GetStdHandle', ctypes.windll.kernel32))
1354 h = GetStdHandle(WIN_OUTPUT_IDS[fileno])
1355
1356 WriteConsoleW = compat_ctypes_WINFUNCTYPE(
1357 ctypes.wintypes.BOOL, ctypes.wintypes.HANDLE, ctypes.wintypes.LPWSTR,
1358 ctypes.wintypes.DWORD, ctypes.POINTER(ctypes.wintypes.DWORD),
1359 ctypes.wintypes.LPVOID)(('WriteConsoleW', ctypes.windll.kernel32))
1360 written = ctypes.wintypes.DWORD(0)
1361
1362 GetFileType = compat_ctypes_WINFUNCTYPE(ctypes.wintypes.DWORD, ctypes.wintypes.DWORD)(('GetFileType', ctypes.windll.kernel32))
1363 FILE_TYPE_CHAR = 0x0002
1364 FILE_TYPE_REMOTE = 0x8000
1365 GetConsoleMode = compat_ctypes_WINFUNCTYPE(
1366 ctypes.wintypes.BOOL, ctypes.wintypes.HANDLE,
1367 ctypes.POINTER(ctypes.wintypes.DWORD))(
1368 ('GetConsoleMode', ctypes.windll.kernel32))
1369 INVALID_HANDLE_VALUE = ctypes.wintypes.DWORD(-1).value
1370
1371 def not_a_console(handle):
1372 if handle == INVALID_HANDLE_VALUE or handle is None:
1373 return True
1374 return ((GetFileType(handle) & ~FILE_TYPE_REMOTE) != FILE_TYPE_CHAR or
1375 GetConsoleMode(handle, ctypes.byref(ctypes.wintypes.DWORD())) == 0)
1376
1377 if not_a_console(h):
1378 return False
1379
1380 def next_nonbmp_pos(s):
1381 try:
1382 return next(i for i, c in enumerate(s) if ord(c) > 0xffff)
1383 except StopIteration:
1384 return len(s)
1385
1386 while s:
1387 count = min(next_nonbmp_pos(s), 1024)
1388
1389 ret = WriteConsoleW(
1390 h, s, count if count else 2, ctypes.byref(written), None)
1391 if ret == 0:
1392 raise OSError('Failed to write string')
1393 if not count: # We just wrote a non-BMP character
1394 assert written.value == 2
1395 s = s[1:]
1396 else:
1397 assert written.value > 0
1398 s = s[written.value:]
1399 return True
1400
1401
1402 def write_string(s, out=None, encoding=None):
1403 if out is None:
1404 out = sys.stderr
1405 assert type(s) == compat_str
1406
1407 if sys.platform == 'win32' and encoding is None and hasattr(out, 'fileno'):
1408 if _windows_write_string(s, out):
1409 return
1410
1411 if ('b' in getattr(out, 'mode', '') or
1412 sys.version_info[0] < 3): # Python 2 lies about mode of sys.stderr
1413 byt = s.encode(encoding or preferredencoding(), 'ignore')
1414 out.write(byt)
1415 elif hasattr(out, 'buffer'):
1416 enc = encoding or getattr(out, 'encoding', None) or preferredencoding()
1417 byt = s.encode(enc, 'ignore')
1418 out.buffer.write(byt)
1419 else:
1420 out.write(s)
1421 out.flush()
1422
1423
1424 def bytes_to_intlist(bs):
1425 if not bs:
1426 return []
1427 if isinstance(bs[0], int): # Python 3
1428 return list(bs)
1429 else:
1430 return [ord(c) for c in bs]
1431
1432
1433 def intlist_to_bytes(xs):
1434 if not xs:
1435 return b''
1436 return compat_struct_pack('%dB' % len(xs), *xs)
1437
1438
1439 # Cross-platform file locking
1440 if sys.platform == 'win32':
1441 import ctypes.wintypes
1442 import msvcrt
1443
1444 class OVERLAPPED(ctypes.Structure):
1445 _fields_ = [
1446 ('Internal', ctypes.wintypes.LPVOID),
1447 ('InternalHigh', ctypes.wintypes.LPVOID),
1448 ('Offset', ctypes.wintypes.DWORD),
1449 ('OffsetHigh', ctypes.wintypes.DWORD),
1450 ('hEvent', ctypes.wintypes.HANDLE),
1451 ]
1452
1453 kernel32 = ctypes.windll.kernel32
1454 LockFileEx = kernel32.LockFileEx
1455 LockFileEx.argtypes = [
1456 ctypes.wintypes.HANDLE, # hFile
1457 ctypes.wintypes.DWORD, # dwFlags
1458 ctypes.wintypes.DWORD, # dwReserved
1459 ctypes.wintypes.DWORD, # nNumberOfBytesToLockLow
1460 ctypes.wintypes.DWORD, # nNumberOfBytesToLockHigh
1461 ctypes.POINTER(OVERLAPPED) # Overlapped
1462 ]
1463 LockFileEx.restype = ctypes.wintypes.BOOL
1464 UnlockFileEx = kernel32.UnlockFileEx
1465 UnlockFileEx.argtypes = [
1466 ctypes.wintypes.HANDLE, # hFile
1467 ctypes.wintypes.DWORD, # dwReserved
1468 ctypes.wintypes.DWORD, # nNumberOfBytesToLockLow
1469 ctypes.wintypes.DWORD, # nNumberOfBytesToLockHigh
1470 ctypes.POINTER(OVERLAPPED) # Overlapped
1471 ]
1472 UnlockFileEx.restype = ctypes.wintypes.BOOL
1473 whole_low = 0xffffffff
1474 whole_high = 0x7fffffff
1475
1476 def _lock_file(f, exclusive):
1477 overlapped = OVERLAPPED()
1478 overlapped.Offset = 0
1479 overlapped.OffsetHigh = 0
1480 overlapped.hEvent = 0
1481 f._lock_file_overlapped_p = ctypes.pointer(overlapped)
1482 handle = msvcrt.get_osfhandle(f.fileno())
1483 if not LockFileEx(handle, 0x2 if exclusive else 0x0, 0,
1484 whole_low, whole_high, f._lock_file_overlapped_p):
1485 raise OSError('Locking file failed: %r' % ctypes.FormatError())
1486
1487 def _unlock_file(f):
1488 assert f._lock_file_overlapped_p
1489 handle = msvcrt.get_osfhandle(f.fileno())
1490 if not UnlockFileEx(handle, 0,
1491 whole_low, whole_high, f._lock_file_overlapped_p):
1492 raise OSError('Unlocking file failed: %r' % ctypes.FormatError())
1493
1494 else:
1495 # Some platforms, such as Jython, is missing fcntl
1496 try:
1497 import fcntl
1498
1499 def _lock_file(f, exclusive):
1500 fcntl.flock(f, fcntl.LOCK_EX if exclusive else fcntl.LOCK_SH)
1501
1502 def _unlock_file(f):
1503 fcntl.flock(f, fcntl.LOCK_UN)
1504 except ImportError:
1505 UNSUPPORTED_MSG = 'file locking is not supported on this platform'
1506
1507 def _lock_file(f, exclusive):
1508 raise IOError(UNSUPPORTED_MSG)
1509
1510 def _unlock_file(f):
1511 raise IOError(UNSUPPORTED_MSG)
1512
1513
1514 class locked_file(object):
1515 def __init__(self, filename, mode, encoding=None):
1516 assert mode in ['r', 'a', 'w']
1517 self.f = io.open(filename, mode, encoding=encoding)
1518 self.mode = mode
1519
1520 def __enter__(self):
1521 exclusive = self.mode != 'r'
1522 try:
1523 _lock_file(self.f, exclusive)
1524 except IOError:
1525 self.f.close()
1526 raise
1527 return self
1528
1529 def __exit__(self, etype, value, traceback):
1530 try:
1531 _unlock_file(self.f)
1532 finally:
1533 self.f.close()
1534
1535 def __iter__(self):
1536 return iter(self.f)
1537
1538 def write(self, *args):
1539 return self.f.write(*args)
1540
1541 def read(self, *args):
1542 return self.f.read(*args)
1543
1544
1545 def get_filesystem_encoding():
1546 encoding = sys.getfilesystemencoding()
1547 return encoding if encoding is not None else 'utf-8'
1548
1549
1550 def shell_quote(args):
1551 quoted_args = []
1552 encoding = get_filesystem_encoding()
1553 for a in args:
1554 if isinstance(a, bytes):
1555 # We may get a filename encoded with 'encodeFilename'
1556 a = a.decode(encoding)
1557 quoted_args.append(compat_shlex_quote(a))
1558 return ' '.join(quoted_args)
1559
1560
1561 def smuggle_url(url, data):
1562 """ Pass additional data in a URL for internal use. """
1563
1564 url, idata = unsmuggle_url(url, {})
1565 data.update(idata)
1566 sdata = compat_urllib_parse_urlencode(
1567 {'__youtubedl_smuggle': json.dumps(data)})
1568 return url + '#' + sdata
1569
1570
1571 def unsmuggle_url(smug_url, default=None):
1572 if '#__youtubedl_smuggle' not in smug_url:
1573 return smug_url, default
1574 url, _, sdata = smug_url.rpartition('#')
1575 jsond = compat_parse_qs(sdata)['__youtubedl_smuggle'][0]
1576 data = json.loads(jsond)
1577 return url, data
1578
1579
1580 def format_bytes(bytes):
1581 if bytes is None:
1582 return 'N/A'
1583 if type(bytes) is str:
1584 bytes = float(bytes)
1585 if bytes == 0.0:
1586 exponent = 0
1587 else:
1588 exponent = int(math.log(bytes, 1024.0))
1589 suffix = ['B', 'KiB', 'MiB', 'GiB', 'TiB', 'PiB', 'EiB', 'ZiB', 'YiB'][exponent]
1590 converted = float(bytes) / float(1024 ** exponent)
1591 return '%.2f%s' % (converted, suffix)
1592
1593
1594 def lookup_unit_table(unit_table, s):
1595 units_re = '|'.join(re.escape(u) for u in unit_table)
1596 m = re.match(
1597 r'(?P<num>[0-9]+(?:[,.][0-9]*)?)\s*(?P<unit>%s)\b' % units_re, s)
1598 if not m:
1599 return None
1600 num_str = m.group('num').replace(',', '.')
1601 mult = unit_table[m.group('unit')]
1602 return int(float(num_str) * mult)
1603
1604
1605 def parse_filesize(s):
1606 if s is None:
1607 return None
1608
1609 # The lower-case forms are of course incorrect and unofficial,
1610 # but we support those too
1611 _UNIT_TABLE = {
1612 'B': 1,
1613 'b': 1,
1614 'bytes': 1,
1615 'KiB': 1024,
1616 'KB': 1000,
1617 'kB': 1024,
1618 'Kb': 1000,
1619 'kb': 1000,
1620 'kilobytes': 1000,
1621 'kibibytes': 1024,
1622 'MiB': 1024 ** 2,
1623 'MB': 1000 ** 2,
1624 'mB': 1024 ** 2,
1625 'Mb': 1000 ** 2,
1626 'mb': 1000 ** 2,
1627 'megabytes': 1000 ** 2,
1628 'mebibytes': 1024 ** 2,
1629 'GiB': 1024 ** 3,
1630 'GB': 1000 ** 3,
1631 'gB': 1024 ** 3,
1632 'Gb': 1000 ** 3,
1633 'gb': 1000 ** 3,
1634 'gigabytes': 1000 ** 3,
1635 'gibibytes': 1024 ** 3,
1636 'TiB': 1024 ** 4,
1637 'TB': 1000 ** 4,
1638 'tB': 1024 ** 4,
1639 'Tb': 1000 ** 4,
1640 'tb': 1000 ** 4,
1641 'terabytes': 1000 ** 4,
1642 'tebibytes': 1024 ** 4,
1643 'PiB': 1024 ** 5,
1644 'PB': 1000 ** 5,
1645 'pB': 1024 ** 5,
1646 'Pb': 1000 ** 5,
1647 'pb': 1000 ** 5,
1648 'petabytes': 1000 ** 5,
1649 'pebibytes': 1024 ** 5,
1650 'EiB': 1024 ** 6,
1651 'EB': 1000 ** 6,
1652 'eB': 1024 ** 6,
1653 'Eb': 1000 ** 6,
1654 'eb': 1000 ** 6,
1655 'exabytes': 1000 ** 6,
1656 'exbibytes': 1024 ** 6,
1657 'ZiB': 1024 ** 7,
1658 'ZB': 1000 ** 7,
1659 'zB': 1024 ** 7,
1660 'Zb': 1000 ** 7,
1661 'zb': 1000 ** 7,
1662 'zettabytes': 1000 ** 7,
1663 'zebibytes': 1024 ** 7,
1664 'YiB': 1024 ** 8,
1665 'YB': 1000 ** 8,
1666 'yB': 1024 ** 8,
1667 'Yb': 1000 ** 8,
1668 'yb': 1000 ** 8,
1669 'yottabytes': 1000 ** 8,
1670 'yobibytes': 1024 ** 8,
1671 }
1672
1673 return lookup_unit_table(_UNIT_TABLE, s)
1674
1675
1676 def parse_count(s):
1677 if s is None:
1678 return None
1679
1680 s = s.strip()
1681
1682 if re.match(r'^[\d,.]+$', s):
1683 return str_to_int(s)
1684
1685 _UNIT_TABLE = {
1686 'k': 1000,
1687 'K': 1000,
1688 'm': 1000 ** 2,
1689 'M': 1000 ** 2,
1690 'kk': 1000 ** 2,
1691 'KK': 1000 ** 2,
1692 }
1693
1694 return lookup_unit_table(_UNIT_TABLE, s)
1695
1696
1697 def parse_resolution(s):
1698 if s is None:
1699 return {}
1700
1701 mobj = re.search(r'\b(?P<w>\d+)\s*[xX×]\s*(?P<h>\d+)\b', s)
1702 if mobj:
1703 return {
1704 'width': int(mobj.group('w')),
1705 'height': int(mobj.group('h')),
1706 }
1707
1708 mobj = re.search(r'\b(\d+)[pPiI]\b', s)
1709 if mobj:
1710 return {'height': int(mobj.group(1))}
1711
1712 mobj = re.search(r'\b([48])[kK]\b', s)
1713 if mobj:
1714 return {'height': int(mobj.group(1)) * 540}
1715
1716 return {}
1717
1718
1719 def month_by_name(name, lang='en'):
1720 """ Return the number of a month by (locale-independently) English name """
1721
1722 month_names = MONTH_NAMES.get(lang, MONTH_NAMES['en'])
1723
1724 try:
1725 return month_names.index(name) + 1
1726 except ValueError:
1727 return None
1728
1729
1730 def month_by_abbreviation(abbrev):
1731 """ Return the number of a month by (locale-independently) English
1732 abbreviations """
1733
1734 try:
1735 return [s[:3] for s in ENGLISH_MONTH_NAMES].index(abbrev) + 1
1736 except ValueError:
1737 return None
1738
1739
1740 def fix_xml_ampersands(xml_str):
1741 """Replace all the '&' by '&amp;' in XML"""
1742 return re.sub(
1743 r'&(?!amp;|lt;|gt;|apos;|quot;|#x[0-9a-fA-F]{,4};|#[0-9]{,4};)',
1744 '&amp;',
1745 xml_str)
1746
1747
1748 def setproctitle(title):
1749 assert isinstance(title, compat_str)
1750
1751 # ctypes in Jython is not complete
1752 # http://bugs.jython.org/issue2148
1753 if sys.platform.startswith('java'):
1754 return
1755
1756 try:
1757 libc = ctypes.cdll.LoadLibrary('libc.so.6')
1758 except OSError:
1759 return
1760 except TypeError:
1761 # LoadLibrary in Windows Python 2.7.13 only expects
1762 # a bytestring, but since unicode_literals turns
1763 # every string into a unicode string, it fails.
1764 return
1765 title_bytes = title.encode('utf-8')
1766 buf = ctypes.create_string_buffer(len(title_bytes))
1767 buf.value = title_bytes
1768 try:
1769 libc.prctl(15, buf, 0, 0, 0)
1770 except AttributeError:
1771 return # Strange libc, just skip this
1772
1773
1774 def remove_start(s, start):
1775 return s[len(start):] if s is not None and s.startswith(start) else s
1776
1777
1778 def remove_end(s, end):
1779 return s[:-len(end)] if s is not None and s.endswith(end) else s
1780
1781
1782 def remove_quotes(s):
1783 if s is None or len(s) < 2:
1784 return s
1785 for quote in ('"', "'", ):
1786 if s[0] == quote and s[-1] == quote:
1787 return s[1:-1]
1788 return s
1789
1790
1791 def url_basename(url):
1792 path = compat_urlparse.urlparse(url).path
1793 return path.strip('/').split('/')[-1]
1794
1795
1796 def base_url(url):
1797 return re.match(r'https?://[^?#&]+/', url).group()
1798
1799
1800 def urljoin(base, path):
1801 if isinstance(path, bytes):
1802 path = path.decode('utf-8')
1803 if not isinstance(path, compat_str) or not path:
1804 return None
1805 if re.match(r'^(?:https?:)?//', path):
1806 return path
1807 if isinstance(base, bytes):
1808 base = base.decode('utf-8')
1809 if not isinstance(base, compat_str) or not re.match(
1810 r'^(?:https?:)?//', base):
1811 return None
1812 return compat_urlparse.urljoin(base, path)
1813
1814
1815 class HEADRequest(compat_urllib_request.Request):
1816 def get_method(self):
1817 return 'HEAD'
1818
1819
1820 class PUTRequest(compat_urllib_request.Request):
1821 def get_method(self):
1822 return 'PUT'
1823
1824
1825 def int_or_none(v, scale=1, default=None, get_attr=None, invscale=1):
1826 if get_attr:
1827 if v is not None:
1828 v = getattr(v, get_attr, None)
1829 if v == '':
1830 v = None
1831 if v is None:
1832 return default
1833 try:
1834 return int(v) * invscale // scale
1835 except ValueError:
1836 return default
1837
1838
1839 def str_or_none(v, default=None):
1840 return default if v is None else compat_str(v)
1841
1842
1843 def str_to_int(int_str):
1844 """ A more relaxed version of int_or_none """
1845 if int_str is None:
1846 return None
1847 int_str = re.sub(r'[,\.\+]', '', int_str)
1848 return int(int_str)
1849
1850
1851 def float_or_none(v, scale=1, invscale=1, default=None):
1852 if v is None:
1853 return default
1854 try:
1855 return float(v) * invscale / scale
1856 except ValueError:
1857 return default
1858
1859
1860 def bool_or_none(v, default=None):
1861 return v if isinstance(v, bool) else default
1862
1863
1864 def strip_or_none(v):
1865 return None if v is None else v.strip()
1866
1867
1868 def parse_duration(s):
1869 if not isinstance(s, compat_basestring):
1870 return None
1871
1872 s = s.strip()
1873
1874 days, hours, mins, secs, ms = [None] * 5
1875 m = re.match(r'(?:(?:(?:(?P<days>[0-9]+):)?(?P<hours>[0-9]+):)?(?P<mins>[0-9]+):)?(?P<secs>[0-9]+)(?P<ms>\.[0-9]+)?Z?$', s)
1876 if m:
1877 days, hours, mins, secs, ms = m.groups()
1878 else:
1879 m = re.match(
1880 r'''(?ix)(?:P?
1881 (?:
1882 [0-9]+\s*y(?:ears?)?\s*
1883 )?
1884 (?:
1885 [0-9]+\s*m(?:onths?)?\s*
1886 )?
1887 (?:
1888 [0-9]+\s*w(?:eeks?)?\s*
1889 )?
1890 (?:
1891 (?P<days>[0-9]+)\s*d(?:ays?)?\s*
1892 )?
1893 T)?
1894 (?:
1895 (?P<hours>[0-9]+)\s*h(?:ours?)?\s*
1896 )?
1897 (?:
1898 (?P<mins>[0-9]+)\s*m(?:in(?:ute)?s?)?\s*
1899 )?
1900 (?:
1901 (?P<secs>[0-9]+)(?P<ms>\.[0-9]+)?\s*s(?:ec(?:ond)?s?)?\s*
1902 )?Z?$''', s)
1903 if m:
1904 days, hours, mins, secs, ms = m.groups()
1905 else:
1906 m = re.match(r'(?i)(?:(?P<hours>[0-9.]+)\s*(?:hours?)|(?P<mins>[0-9.]+)\s*(?:mins?\.?|minutes?)\s*)Z?$', s)
1907 if m:
1908 hours, mins = m.groups()
1909 else:
1910 return None
1911
1912 duration = 0
1913 if secs:
1914 duration += float(secs)
1915 if mins:
1916 duration += float(mins) * 60
1917 if hours:
1918 duration += float(hours) * 60 * 60
1919 if days:
1920 duration += float(days) * 24 * 60 * 60
1921 if ms:
1922 duration += float(ms)
1923 return duration
1924
1925
1926 def prepend_extension(filename, ext, expected_real_ext=None):
1927 name, real_ext = os.path.splitext(filename)
1928 return (
1929 '{0}.{1}{2}'.format(name, ext, real_ext)
1930 if not expected_real_ext or real_ext[1:] == expected_real_ext
1931 else '{0}.{1}'.format(filename, ext))
1932
1933
1934 def replace_extension(filename, ext, expected_real_ext=None):
1935 name, real_ext = os.path.splitext(filename)
1936 return '{0}.{1}'.format(
1937 name if not expected_real_ext or real_ext[1:] == expected_real_ext else filename,
1938 ext)
1939
1940
1941 def check_executable(exe, args=[]):
1942 """ Checks if the given binary is installed somewhere in PATH, and returns its name.
1943 args can be a list of arguments for a short output (like -version) """
1944 try:
1945 subprocess.Popen([exe] + args, stdout=subprocess.PIPE, stderr=subprocess.PIPE).communicate()
1946 except OSError:
1947 return False
1948 return exe
1949
1950
1951 def get_exe_version(exe, args=['--version'],
1952 version_re=None, unrecognized='present'):
1953 """ Returns the version of the specified executable,
1954 or False if the executable is not present """
1955 try:
1956 # STDIN should be redirected too. On UNIX-like systems, ffmpeg triggers
1957 # SIGTTOU if youtube-dl is run in the background.
1958 # See https://github.com/rg3/youtube-dl/issues/955#issuecomment-209789656
1959 out, _ = subprocess.Popen(
1960 [encodeArgument(exe)] + args,
1961 stdin=subprocess.PIPE,
1962 stdout=subprocess.PIPE, stderr=subprocess.STDOUT).communicate()
1963 except OSError:
1964 return False
1965 if isinstance(out, bytes): # Python 2.x
1966 out = out.decode('ascii', 'ignore')
1967 return detect_exe_version(out, version_re, unrecognized)
1968
1969
1970 def detect_exe_version(output, version_re=None, unrecognized='present'):
1971 assert isinstance(output, compat_str)
1972 if version_re is None:
1973 version_re = r'version\s+([-0-9._a-zA-Z]+)'
1974 m = re.search(version_re, output)
1975 if m:
1976 return m.group(1)
1977 else:
1978 return unrecognized
1979
1980
1981 class PagedList(object):
1982 def __len__(self):
1983 # This is only useful for tests
1984 return len(self.getslice())
1985
1986
1987 class OnDemandPagedList(PagedList):
1988 def __init__(self, pagefunc, pagesize, use_cache=True):
1989 self._pagefunc = pagefunc
1990 self._pagesize = pagesize
1991 self._use_cache = use_cache
1992 if use_cache:
1993 self._cache = {}
1994
1995 def getslice(self, start=0, end=None):
1996 res = []
1997 for pagenum in itertools.count(start // self._pagesize):
1998 firstid = pagenum * self._pagesize
1999 nextfirstid = pagenum * self._pagesize + self._pagesize
2000 if start >= nextfirstid:
2001 continue
2002
2003 page_results = None
2004 if self._use_cache:
2005 page_results = self._cache.get(pagenum)
2006 if page_results is None:
2007 page_results = list(self._pagefunc(pagenum))
2008 if self._use_cache:
2009 self._cache[pagenum] = page_results
2010
2011 startv = (
2012 start % self._pagesize
2013 if firstid <= start < nextfirstid
2014 else 0)
2015
2016 endv = (
2017 ((end - 1) % self._pagesize) + 1
2018 if (end is not None and firstid <= end <= nextfirstid)
2019 else None)
2020
2021 if startv != 0 or endv is not None:
2022 page_results = page_results[startv:endv]
2023 res.extend(page_results)
2024
2025 # A little optimization - if current page is not "full", ie. does
2026 # not contain page_size videos then we can assume that this page
2027 # is the last one - there are no more ids on further pages -
2028 # i.e. no need to query again.
2029 if len(page_results) + startv < self._pagesize:
2030 break
2031
2032 # If we got the whole page, but the next page is not interesting,
2033 # break out early as well
2034 if end == nextfirstid:
2035 break
2036 return res
2037
2038
2039 class InAdvancePagedList(PagedList):
2040 def __init__(self, pagefunc, pagecount, pagesize):
2041 self._pagefunc = pagefunc
2042 self._pagecount = pagecount
2043 self._pagesize = pagesize
2044
2045 def getslice(self, start=0, end=None):
2046 res = []
2047 start_page = start // self._pagesize
2048 end_page = (
2049 self._pagecount if end is None else (end // self._pagesize + 1))
2050 skip_elems = start - start_page * self._pagesize
2051 only_more = None if end is None else end - start
2052 for pagenum in range(start_page, end_page):
2053 page = list(self._pagefunc(pagenum))
2054 if skip_elems:
2055 page = page[skip_elems:]
2056 skip_elems = None
2057 if only_more is not None:
2058 if len(page) < only_more:
2059 only_more -= len(page)
2060 else:
2061 page = page[:only_more]
2062 res.extend(page)
2063 break
2064 res.extend(page)
2065 return res
2066
2067
2068 def uppercase_escape(s):
2069 unicode_escape = codecs.getdecoder('unicode_escape')
2070 return re.sub(
2071 r'\\U[0-9a-fA-F]{8}',
2072 lambda m: unicode_escape(m.group(0))[0],
2073 s)
2074
2075
2076 def lowercase_escape(s):
2077 unicode_escape = codecs.getdecoder('unicode_escape')
2078 return re.sub(
2079 r'\\u[0-9a-fA-F]{4}',
2080 lambda m: unicode_escape(m.group(0))[0],
2081 s)
2082
2083
2084 def escape_rfc3986(s):
2085 """Escape non-ASCII characters as suggested by RFC 3986"""
2086 if sys.version_info < (3, 0) and isinstance(s, compat_str):
2087 s = s.encode('utf-8')
2088 return compat_urllib_parse.quote(s, b"%/;:@&=+$,!~*'()?#[]")
2089
2090
2091 def escape_url(url):
2092 """Escape URL as suggested by RFC 3986"""
2093 url_parsed = compat_urllib_parse_urlparse(url)
2094 return url_parsed._replace(
2095 netloc=url_parsed.netloc.encode('idna').decode('ascii'),
2096 path=escape_rfc3986(url_parsed.path),
2097 params=escape_rfc3986(url_parsed.params),
2098 query=escape_rfc3986(url_parsed.query),
2099 fragment=escape_rfc3986(url_parsed.fragment)
2100 ).geturl()
2101
2102
2103 def read_batch_urls(batch_fd):
2104 def fixup(url):
2105 if not isinstance(url, compat_str):
2106 url = url.decode('utf-8', 'replace')
2107 BOM_UTF8 = '\xef\xbb\xbf'
2108 if url.startswith(BOM_UTF8):
2109 url = url[len(BOM_UTF8):]
2110 url = url.strip()
2111 if url.startswith(('#', ';', ']')):
2112 return False
2113 return url
2114
2115 with contextlib.closing(batch_fd) as fd:
2116 return [url for url in map(fixup, fd) if url]
2117
2118
2119 def urlencode_postdata(*args, **kargs):
2120 return compat_urllib_parse_urlencode(*args, **kargs).encode('ascii')
2121
2122
2123 def update_url_query(url, query):
2124 if not query:
2125 return url
2126 parsed_url = compat_urlparse.urlparse(url)
2127 qs = compat_parse_qs(parsed_url.query)
2128 qs.update(query)
2129 return compat_urlparse.urlunparse(parsed_url._replace(
2130 query=compat_urllib_parse_urlencode(qs, True)))
2131
2132
2133 def update_Request(req, url=None, data=None, headers={}, query={}):
2134 req_headers = req.headers.copy()
2135 req_headers.update(headers)
2136 req_data = data or req.data
2137 req_url = update_url_query(url or req.get_full_url(), query)
2138 req_get_method = req.get_method()
2139 if req_get_method == 'HEAD':
2140 req_type = HEADRequest
2141 elif req_get_method == 'PUT':
2142 req_type = PUTRequest
2143 else:
2144 req_type = compat_urllib_request.Request
2145 new_req = req_type(
2146 req_url, data=req_data, headers=req_headers,
2147 origin_req_host=req.origin_req_host, unverifiable=req.unverifiable)
2148 if hasattr(req, 'timeout'):
2149 new_req.timeout = req.timeout
2150 return new_req
2151
2152
2153 def _multipart_encode_impl(data, boundary):
2154 content_type = 'multipart/form-data; boundary=%s' % boundary
2155
2156 out = b''
2157 for k, v in data.items():
2158 out += b'--' + boundary.encode('ascii') + b'\r\n'
2159 if isinstance(k, compat_str):
2160 k = k.encode('utf-8')
2161 if isinstance(v, compat_str):
2162 v = v.encode('utf-8')
2163 # RFC 2047 requires non-ASCII field names to be encoded, while RFC 7578
2164 # suggests sending UTF-8 directly. Firefox sends UTF-8, too
2165 content = b'Content-Disposition: form-data; name="' + k + b'"\r\n\r\n' + v + b'\r\n'
2166 if boundary.encode('ascii') in content:
2167 raise ValueError('Boundary overlaps with data')
2168 out += content
2169
2170 out += b'--' + boundary.encode('ascii') + b'--\r\n'
2171
2172 return out, content_type
2173
2174
2175 def multipart_encode(data, boundary=None):
2176 '''
2177 Encode a dict to RFC 7578-compliant form-data
2178
2179 data:
2180 A dict where keys and values can be either Unicode or bytes-like
2181 objects.
2182 boundary:
2183 If specified a Unicode object, it's used as the boundary. Otherwise
2184 a random boundary is generated.
2185
2186 Reference: https://tools.ietf.org/html/rfc7578
2187 '''
2188 has_specified_boundary = boundary is not None
2189
2190 while True:
2191 if boundary is None:
2192 boundary = '---------------' + str(random.randrange(0x0fffffff, 0xffffffff))
2193
2194 try:
2195 out, content_type = _multipart_encode_impl(data, boundary)
2196 break
2197 except ValueError:
2198 if has_specified_boundary:
2199 raise
2200 boundary = None
2201
2202 return out, content_type
2203
2204
2205 def dict_get(d, key_or_keys, default=None, skip_false_values=True):
2206 if isinstance(key_or_keys, (list, tuple)):
2207 for key in key_or_keys:
2208 if key not in d or d[key] is None or skip_false_values and not d[key]:
2209 continue
2210 return d[key]
2211 return default
2212 return d.get(key_or_keys, default)
2213
2214
2215 def try_get(src, getter, expected_type=None):
2216 if not isinstance(getter, (list, tuple)):
2217 getter = [getter]
2218 for get in getter:
2219 try:
2220 v = get(src)
2221 except (AttributeError, KeyError, TypeError, IndexError):
2222 pass
2223 else:
2224 if expected_type is None or isinstance(v, expected_type):
2225 return v
2226
2227
2228 def merge_dicts(*dicts):
2229 merged = {}
2230 for a_dict in dicts:
2231 for k, v in a_dict.items():
2232 if v is None:
2233 continue
2234 if (k not in merged or
2235 (isinstance(v, compat_str) and v and
2236 isinstance(merged[k], compat_str) and
2237 not merged[k])):
2238 merged[k] = v
2239 return merged
2240
2241
2242 def encode_compat_str(string, encoding=preferredencoding(), errors='strict'):
2243 return string if isinstance(string, compat_str) else compat_str(string, encoding, errors)
2244
2245
2246 US_RATINGS = {
2247 'G': 0,
2248 'PG': 10,
2249 'PG-13': 13,
2250 'R': 16,
2251 'NC': 18,
2252 }
2253
2254
2255 TV_PARENTAL_GUIDELINES = {
2256 'TV-Y': 0,
2257 'TV-Y7': 7,
2258 'TV-G': 0,
2259 'TV-PG': 0,
2260 'TV-14': 14,
2261 'TV-MA': 17,
2262 }
2263
2264
2265 def parse_age_limit(s):
2266 if type(s) == int:
2267 return s if 0 <= s <= 21 else None
2268 if not isinstance(s, compat_basestring):
2269 return None
2270 m = re.match(r'^(?P<age>\d{1,2})\+?$', s)
2271 if m:
2272 return int(m.group('age'))
2273 if s in US_RATINGS:
2274 return US_RATINGS[s]
2275 m = re.match(r'^TV[_-]?(%s)$' % '|'.join(k[3:] for k in TV_PARENTAL_GUIDELINES), s)
2276 if m:
2277 return TV_PARENTAL_GUIDELINES['TV-' + m.group(1)]
2278 return None
2279
2280
2281 def strip_jsonp(code):
2282 return re.sub(
2283 r'''(?sx)^
2284 (?:window\.)?(?P<func_name>[a-zA-Z0-9_.$]+)
2285 (?:\s*&&\s*(?P=func_name))?
2286 \s*\(\s*(?P<callback_data>.*)\);?
2287 \s*?(?://[^\n]*)*$''',
2288 r'\g<callback_data>', code)
2289
2290
2291 def js_to_json(code):
2292 COMMENT_RE = r'/\*(?:(?!\*/).)*?\*/|//[^\n]*'
2293 SKIP_RE = r'\s*(?:{comment})?\s*'.format(comment=COMMENT_RE)
2294 INTEGER_TABLE = (
2295 (r'(?s)^(0[xX][0-9a-fA-F]+){skip}:?$'.format(skip=SKIP_RE), 16),
2296 (r'(?s)^(0+[0-7]+){skip}:?$'.format(skip=SKIP_RE), 8),
2297 )
2298
2299 def fix_kv(m):
2300 v = m.group(0)
2301 if v in ('true', 'false', 'null'):
2302 return v
2303 elif v.startswith('/*') or v.startswith('//') or v == ',':
2304 return ""
2305
2306 if v[0] in ("'", '"'):
2307 v = re.sub(r'(?s)\\.|"', lambda m: {
2308 '"': '\\"',
2309 "\\'": "'",
2310 '\\\n': '',
2311 '\\x': '\\u00',
2312 }.get(m.group(0), m.group(0)), v[1:-1])
2313
2314 for regex, base in INTEGER_TABLE:
2315 im = re.match(regex, v)
2316 if im:
2317 i = int(im.group(1), base)
2318 return '"%d":' % i if v.endswith(':') else '%d' % i
2319
2320 return '"%s"' % v
2321
2322 return re.sub(r'''(?sx)
2323 "(?:[^"\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^"\\]*"|
2324 '(?:[^'\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^'\\]*'|
2325 {comment}|,(?={skip}[\]}}])|
2326 (?:(?<![0-9])[eE]|[a-df-zA-DF-Z_])[.a-zA-Z_0-9]*|
2327 \b(?:0[xX][0-9a-fA-F]+|0+[0-7]+)(?:{skip}:)?|
2328 [0-9]+(?={skip}:)
2329 '''.format(comment=COMMENT_RE, skip=SKIP_RE), fix_kv, code)
2330
2331
2332 def qualities(quality_ids):
2333 """ Get a numeric quality value out of a list of possible values """
2334 def q(qid):
2335 try:
2336 return quality_ids.index(qid)
2337 except ValueError:
2338 return -1
2339 return q
2340
2341
2342 DEFAULT_OUTTMPL = '%(title)s-%(id)s.%(ext)s'
2343
2344
2345 def limit_length(s, length):
2346 """ Add ellipses to overly long strings """
2347 if s is None:
2348 return None
2349 ELLIPSES = '...'
2350 if len(s) > length:
2351 return s[:length - len(ELLIPSES)] + ELLIPSES
2352 return s
2353
2354
2355 def version_tuple(v):
2356 return tuple(int(e) for e in re.split(r'[-.]', v))
2357
2358
2359 def is_outdated_version(version, limit, assume_new=True):
2360 if not version:
2361 return not assume_new
2362 try:
2363 return version_tuple(version) < version_tuple(limit)
2364 except ValueError:
2365 return not assume_new
2366
2367
2368 def ytdl_is_updateable():
2369 """ Returns if youtube-dl can be updated with -U """
2370 from zipimport import zipimporter
2371
2372 return isinstance(globals().get('__loader__'), zipimporter) or hasattr(sys, 'frozen')
2373
2374
2375 def args_to_str(args):
2376 # Get a short string representation for a subprocess command
2377 return ' '.join(compat_shlex_quote(a) for a in args)
2378
2379
2380 def error_to_compat_str(err):
2381 err_str = str(err)
2382 # On python 2 error byte string must be decoded with proper
2383 # encoding rather than ascii
2384 if sys.version_info[0] < 3:
2385 err_str = err_str.decode(preferredencoding())
2386 return err_str
2387
2388
2389 def mimetype2ext(mt):
2390 if mt is None:
2391 return None
2392
2393 ext = {
2394 'audio/mp4': 'm4a',
2395 # Per RFC 3003, audio/mpeg can be .mp1, .mp2 or .mp3. Here use .mp3 as
2396 # it's the most popular one
2397 'audio/mpeg': 'mp3',
2398 }.get(mt)
2399 if ext is not None:
2400 return ext
2401
2402 _, _, res = mt.rpartition('/')
2403 res = res.split(';')[0].strip().lower()
2404
2405 return {
2406 '3gpp': '3gp',
2407 'smptett+xml': 'tt',
2408 'ttaf+xml': 'dfxp',
2409 'ttml+xml': 'ttml',
2410 'x-flv': 'flv',
2411 'x-mp4-fragmented': 'mp4',
2412 'x-ms-sami': 'sami',
2413 'x-ms-wmv': 'wmv',
2414 'mpegurl': 'm3u8',
2415 'x-mpegurl': 'm3u8',
2416 'vnd.apple.mpegurl': 'm3u8',
2417 'dash+xml': 'mpd',
2418 'f4m+xml': 'f4m',
2419 'hds+xml': 'f4m',
2420 'vnd.ms-sstr+xml': 'ism',
2421 'quicktime': 'mov',
2422 'mp2t': 'ts',
2423 }.get(res, res)
2424
2425
2426 def parse_codecs(codecs_str):
2427 # http://tools.ietf.org/html/rfc6381
2428 if not codecs_str:
2429 return {}
2430 splited_codecs = list(filter(None, map(
2431 lambda str: str.strip(), codecs_str.strip().strip(',').split(','))))
2432 vcodec, acodec = None, None
2433 for full_codec in splited_codecs:
2434 codec = full_codec.split('.')[0]
2435 if codec in ('avc1', 'avc2', 'avc3', 'avc4', 'vp9', 'vp8', 'hev1', 'hev2', 'h263', 'h264', 'mp4v', 'hvc1'):
2436 if not vcodec:
2437 vcodec = full_codec
2438 elif codec in ('mp4a', 'opus', 'vorbis', 'mp3', 'aac', 'ac-3', 'ec-3', 'eac3', 'dtsc', 'dtse', 'dtsh', 'dtsl'):
2439 if not acodec:
2440 acodec = full_codec
2441 else:
2442 write_string('WARNING: Unknown codec %s\n' % full_codec, sys.stderr)
2443 if not vcodec and not acodec:
2444 if len(splited_codecs) == 2:
2445 return {
2446 'vcodec': vcodec,
2447 'acodec': acodec,
2448 }
2449 elif len(splited_codecs) == 1:
2450 return {
2451 'vcodec': 'none',
2452 'acodec': vcodec,
2453 }
2454 else:
2455 return {
2456 'vcodec': vcodec or 'none',
2457 'acodec': acodec or 'none',
2458 }
2459 return {}
2460
2461
2462 def urlhandle_detect_ext(url_handle):
2463 getheader = url_handle.headers.get
2464
2465 cd = getheader('Content-Disposition')
2466 if cd:
2467 m = re.match(r'attachment;\s*filename="(?P<filename>[^"]+)"', cd)
2468 if m:
2469 e = determine_ext(m.group('filename'), default_ext=None)
2470 if e:
2471 return e
2472
2473 return mimetype2ext(getheader('Content-Type'))
2474
2475
2476 def encode_data_uri(data, mime_type):
2477 return 'data:%s;base64,%s' % (mime_type, base64.b64encode(data).decode('ascii'))
2478
2479
2480 def age_restricted(content_limit, age_limit):
2481 """ Returns True iff the content should be blocked """
2482
2483 if age_limit is None: # No limit set
2484 return False
2485 if content_limit is None:
2486 return False # Content available for everyone
2487 return age_limit < content_limit
2488
2489
2490 def is_html(first_bytes):
2491 """ Detect whether a file contains HTML by examining its first bytes. """
2492
2493 BOMS = [
2494 (b'\xef\xbb\xbf', 'utf-8'),
2495 (b'\x00\x00\xfe\xff', 'utf-32-be'),
2496 (b'\xff\xfe\x00\x00', 'utf-32-le'),
2497 (b'\xff\xfe', 'utf-16-le'),
2498 (b'\xfe\xff', 'utf-16-be'),
2499 ]
2500 for bom, enc in BOMS:
2501 if first_bytes.startswith(bom):
2502 s = first_bytes[len(bom):].decode(enc, 'replace')
2503 break
2504 else:
2505 s = first_bytes.decode('utf-8', 'replace')
2506
2507 return re.match(r'^\s*<', s)
2508
2509
2510 def determine_protocol(info_dict):
2511 protocol = info_dict.get('protocol')
2512 if protocol is not None:
2513 return protocol
2514
2515 url = info_dict['url']
2516 if url.startswith('rtmp'):
2517 return 'rtmp'
2518 elif url.startswith('mms'):
2519 return 'mms'
2520 elif url.startswith('rtsp'):
2521 return 'rtsp'
2522
2523 ext = determine_ext(url)
2524 if ext == 'm3u8':
2525 return 'm3u8'
2526 elif ext == 'f4m':
2527 return 'f4m'
2528
2529 return compat_urllib_parse_urlparse(url).scheme
2530
2531
2532 def render_table(header_row, data):
2533 """ Render a list of rows, each as a list of values """
2534 table = [header_row] + data
2535 max_lens = [max(len(compat_str(v)) for v in col) for col in zip(*table)]
2536 format_str = ' '.join('%-' + compat_str(ml + 1) + 's' for ml in max_lens[:-1]) + '%s'
2537 return '\n'.join(format_str % tuple(row) for row in table)
2538
2539
2540 def _match_one(filter_part, dct):
2541 COMPARISON_OPERATORS = {
2542 '<': operator.lt,
2543 '<=': operator.le,
2544 '>': operator.gt,
2545 '>=': operator.ge,
2546 '=': operator.eq,
2547 '!=': operator.ne,
2548 }
2549 operator_rex = re.compile(r'''(?x)\s*
2550 (?P<key>[a-z_]+)
2551 \s*(?P<op>%s)(?P<none_inclusive>\s*\?)?\s*
2552 (?:
2553 (?P<intval>[0-9.]+(?:[kKmMgGtTpPeEzZyY]i?[Bb]?)?)|
2554 (?P<quote>["\'])(?P<quotedstrval>(?:\\.|(?!(?P=quote)|\\).)+?)(?P=quote)|
2555 (?P<strval>(?![0-9.])[a-z0-9A-Z]*)
2556 )
2557 \s*$
2558 ''' % '|'.join(map(re.escape, COMPARISON_OPERATORS.keys())))
2559 m = operator_rex.search(filter_part)
2560 if m:
2561 op = COMPARISON_OPERATORS[m.group('op')]
2562 actual_value = dct.get(m.group('key'))
2563 if (m.group('quotedstrval') is not None or
2564 m.group('strval') is not None or
2565 # If the original field is a string and matching comparisonvalue is
2566 # a number we should respect the origin of the original field
2567 # and process comparison value as a string (see
2568 # https://github.com/rg3/youtube-dl/issues/11082).
2569 actual_value is not None and m.group('intval') is not None and
2570 isinstance(actual_value, compat_str)):
2571 if m.group('op') not in ('=', '!='):
2572 raise ValueError(
2573 'Operator %s does not support string values!' % m.group('op'))
2574 comparison_value = m.group('quotedstrval') or m.group('strval') or m.group('intval')
2575 quote = m.group('quote')
2576 if quote is not None:
2577 comparison_value = comparison_value.replace(r'\%s' % quote, quote)
2578 else:
2579 try:
2580 comparison_value = int(m.group('intval'))
2581 except ValueError:
2582 comparison_value = parse_filesize(m.group('intval'))
2583 if comparison_value is None:
2584 comparison_value = parse_filesize(m.group('intval') + 'B')
2585 if comparison_value is None:
2586 raise ValueError(
2587 'Invalid integer value %r in filter part %r' % (
2588 m.group('intval'), filter_part))
2589 if actual_value is None:
2590 return m.group('none_inclusive')
2591 return op(actual_value, comparison_value)
2592
2593 UNARY_OPERATORS = {
2594 '': lambda v: (v is True) if isinstance(v, bool) else (v is not None),
2595 '!': lambda v: (v is False) if isinstance(v, bool) else (v is None),
2596 }
2597 operator_rex = re.compile(r'''(?x)\s*
2598 (?P<op>%s)\s*(?P<key>[a-z_]+)
2599 \s*$
2600 ''' % '|'.join(map(re.escape, UNARY_OPERATORS.keys())))
2601 m = operator_rex.search(filter_part)
2602 if m:
2603 op = UNARY_OPERATORS[m.group('op')]
2604 actual_value = dct.get(m.group('key'))
2605 return op(actual_value)
2606
2607 raise ValueError('Invalid filter part %r' % filter_part)
2608
2609
2610 def match_str(filter_str, dct):
2611 """ Filter a dictionary with a simple string syntax. Returns True (=passes filter) or false """
2612
2613 return all(
2614 _match_one(filter_part, dct) for filter_part in filter_str.split('&'))
2615
2616
2617 def match_filter_func(filter_str):
2618 def _match_func(info_dict):
2619 if match_str(filter_str, info_dict):
2620 return None
2621 else:
2622 video_title = info_dict.get('title', info_dict.get('id', 'video'))
2623 return '%s does not pass filter %s, skipping ..' % (video_title, filter_str)
2624 return _match_func
2625
2626
2627 def parse_dfxp_time_expr(time_expr):
2628 if not time_expr:
2629 return
2630
2631 mobj = re.match(r'^(?P<time_offset>\d+(?:\.\d+)?)s?$', time_expr)
2632 if mobj:
2633 return float(mobj.group('time_offset'))
2634
2635 mobj = re.match(r'^(\d+):(\d\d):(\d\d(?:(?:\.|:)\d+)?)$', time_expr)
2636 if mobj:
2637 return 3600 * int(mobj.group(1)) + 60 * int(mobj.group(2)) + float(mobj.group(3).replace(':', '.'))
2638
2639
2640 def srt_subtitles_timecode(seconds):
2641 return '%02d:%02d:%02d,%03d' % (seconds / 3600, (seconds % 3600) / 60, seconds % 60, (seconds % 1) * 1000)
2642
2643
2644 def dfxp2srt(dfxp_data):
2645 '''
2646 @param dfxp_data A bytes-like object containing DFXP data
2647 @returns A unicode object containing converted SRT data
2648 '''
2649 LEGACY_NAMESPACES = (
2650 (b'http://www.w3.org/ns/ttml', [
2651 b'http://www.w3.org/2004/11/ttaf1',
2652 b'http://www.w3.org/2006/04/ttaf1',
2653 b'http://www.w3.org/2006/10/ttaf1',
2654 ]),
2655 (b'http://www.w3.org/ns/ttml#styling', [
2656 b'http://www.w3.org/ns/ttml#style',
2657 ]),
2658 )
2659
2660 SUPPORTED_STYLING = [
2661 'color',
2662 'fontFamily',
2663 'fontSize',
2664 'fontStyle',
2665 'fontWeight',
2666 'textDecoration'
2667 ]
2668
2669 _x = functools.partial(xpath_with_ns, ns_map={
2670 'xml': 'http://www.w3.org/XML/1998/namespace',
2671 'ttml': 'http://www.w3.org/ns/ttml',
2672 'tts': 'http://www.w3.org/ns/ttml#styling',
2673 })
2674
2675 styles = {}
2676 default_style = {}
2677
2678 class TTMLPElementParser(object):
2679 _out = ''
2680 _unclosed_elements = []
2681 _applied_styles = []
2682
2683 def start(self, tag, attrib):
2684 if tag in (_x('ttml:br'), 'br'):
2685 self._out += '\n'
2686 else:
2687 unclosed_elements = []
2688 style = {}
2689 element_style_id = attrib.get('style')
2690 if default_style:
2691 style.update(default_style)
2692 if element_style_id:
2693 style.update(styles.get(element_style_id, {}))
2694 for prop in SUPPORTED_STYLING:
2695 prop_val = attrib.get(_x('tts:' + prop))
2696 if prop_val:
2697 style[prop] = prop_val
2698 if style:
2699 font = ''
2700 for k, v in sorted(style.items()):
2701 if self._applied_styles and self._applied_styles[-1].get(k) == v:
2702 continue
2703 if k == 'color':
2704 font += ' color="%s"' % v
2705 elif k == 'fontSize':
2706 font += ' size="%s"' % v
2707 elif k == 'fontFamily':
2708 font += ' face="%s"' % v
2709 elif k == 'fontWeight' and v == 'bold':
2710 self._out += '<b>'
2711 unclosed_elements.append('b')
2712 elif k == 'fontStyle' and v == 'italic':
2713 self._out += '<i>'
2714 unclosed_elements.append('i')
2715 elif k == 'textDecoration' and v == 'underline':
2716 self._out += '<u>'
2717 unclosed_elements.append('u')
2718 if font:
2719 self._out += '<font' + font + '>'
2720 unclosed_elements.append('font')
2721 applied_style = {}
2722 if self._applied_styles:
2723 applied_style.update(self._applied_styles[-1])
2724 applied_style.update(style)
2725 self._applied_styles.append(applied_style)
2726 self._unclosed_elements.append(unclosed_elements)
2727
2728 def end(self, tag):
2729 if tag not in (_x('ttml:br'), 'br'):
2730 unclosed_elements = self._unclosed_elements.pop()
2731 for element in reversed(unclosed_elements):
2732 self._out += '</%s>' % element
2733 if unclosed_elements and self._applied_styles:
2734 self._applied_styles.pop()
2735
2736 def data(self, data):
2737 self._out += data
2738
2739 def close(self):
2740 return self._out.strip()
2741
2742 def parse_node(node):
2743 target = TTMLPElementParser()
2744 parser = xml.etree.ElementTree.XMLParser(target=target)
2745 parser.feed(xml.etree.ElementTree.tostring(node))
2746 return parser.close()
2747
2748 for k, v in LEGACY_NAMESPACES:
2749 for ns in v:
2750 dfxp_data = dfxp_data.replace(ns, k)
2751
2752 dfxp = compat_etree_fromstring(dfxp_data)
2753 out = []
2754 paras = dfxp.findall(_x('.//ttml:p')) or dfxp.findall('.//p')
2755
2756 if not paras:
2757 raise ValueError('Invalid dfxp/TTML subtitle')
2758
2759 repeat = False
2760 while True:
2761 for style in dfxp.findall(_x('.//ttml:style')):
2762 style_id = style.get('id') or style.get(_x('xml:id'))
2763 if not style_id:
2764 continue
2765 parent_style_id = style.get('style')
2766 if parent_style_id:
2767 if parent_style_id not in styles:
2768 repeat = True
2769 continue
2770 styles[style_id] = styles[parent_style_id].copy()
2771 for prop in SUPPORTED_STYLING:
2772 prop_val = style.get(_x('tts:' + prop))
2773 if prop_val:
2774 styles.setdefault(style_id, {})[prop] = prop_val
2775 if repeat:
2776 repeat = False
2777 else:
2778 break
2779
2780 for p in ('body', 'div'):
2781 ele = xpath_element(dfxp, [_x('.//ttml:' + p), './/' + p])
2782 if ele is None:
2783 continue
2784 style = styles.get(ele.get('style'))
2785 if not style:
2786 continue
2787 default_style.update(style)
2788
2789 for para, index in zip(paras, itertools.count(1)):
2790 begin_time = parse_dfxp_time_expr(para.attrib.get('begin'))
2791 end_time = parse_dfxp_time_expr(para.attrib.get('end'))
2792 dur = parse_dfxp_time_expr(para.attrib.get('dur'))
2793 if begin_time is None:
2794 continue
2795 if not end_time:
2796 if not dur:
2797 continue
2798 end_time = begin_time + dur
2799 out.append('%d\n%s --> %s\n%s\n\n' % (
2800 index,
2801 srt_subtitles_timecode(begin_time),
2802 srt_subtitles_timecode(end_time),
2803 parse_node(para)))
2804
2805 return ''.join(out)
2806
2807
2808 def cli_option(params, command_option, param):
2809 param = params.get(param)
2810 if param:
2811 param = compat_str(param)
2812 return [command_option, param] if param is not None else []
2813
2814
2815 def cli_bool_option(params, command_option, param, true_value='true', false_value='false', separator=None):
2816 param = params.get(param)
2817 if param is None:
2818 return []
2819 assert isinstance(param, bool)
2820 if separator:
2821 return [command_option + separator + (true_value if param else false_value)]
2822 return [command_option, true_value if param else false_value]
2823
2824
2825 def cli_valueless_option(params, command_option, param, expected_value=True):
2826 param = params.get(param)
2827 return [command_option] if param == expected_value else []
2828
2829
2830 def cli_configuration_args(params, param, default=[]):
2831 ex_args = params.get(param)
2832 if ex_args is None:
2833 return default
2834 assert isinstance(ex_args, list)
2835 return ex_args
2836
2837
2838 class ISO639Utils(object):
2839 # See http://www.loc.gov/standards/iso639-2/ISO-639-2_utf-8.txt
2840 _lang_map = {
2841 'aa': 'aar',
2842 'ab': 'abk',
2843 'ae': 'ave',
2844 'af': 'afr',
2845 'ak': 'aka',
2846 'am': 'amh',
2847 'an': 'arg',
2848 'ar': 'ara',
2849 'as': 'asm',
2850 'av': 'ava',
2851 'ay': 'aym',
2852 'az': 'aze',
2853 'ba': 'bak',
2854 'be': 'bel',
2855 'bg': 'bul',
2856 'bh': 'bih',
2857 'bi': 'bis',
2858 'bm': 'bam',
2859 'bn': 'ben',
2860 'bo': 'bod',
2861 'br': 'bre',
2862 'bs': 'bos',
2863 'ca': 'cat',
2864 'ce': 'che',
2865 'ch': 'cha',
2866 'co': 'cos',
2867 'cr': 'cre',
2868 'cs': 'ces',
2869 'cu': 'chu',
2870 'cv': 'chv',
2871 'cy': 'cym',
2872 'da': 'dan',
2873 'de': 'deu',
2874 'dv': 'div',
2875 'dz': 'dzo',
2876 'ee': 'ewe',
2877 'el': 'ell',
2878 'en': 'eng',
2879 'eo': 'epo',
2880 'es': 'spa',
2881 'et': 'est',
2882 'eu': 'eus',
2883 'fa': 'fas',
2884 'ff': 'ful',
2885 'fi': 'fin',
2886 'fj': 'fij',
2887 'fo': 'fao',
2888 'fr': 'fra',
2889 'fy': 'fry',
2890 'ga': 'gle',
2891 'gd': 'gla',
2892 'gl': 'glg',
2893 'gn': 'grn',
2894 'gu': 'guj',
2895 'gv': 'glv',
2896 'ha': 'hau',
2897 'he': 'heb',
2898 'hi': 'hin',
2899 'ho': 'hmo',
2900 'hr': 'hrv',
2901 'ht': 'hat',
2902 'hu': 'hun',
2903 'hy': 'hye',
2904 'hz': 'her',
2905 'ia': 'ina',
2906 'id': 'ind',
2907 'ie': 'ile',
2908 'ig': 'ibo',
2909 'ii': 'iii',
2910 'ik': 'ipk',
2911 'io': 'ido',
2912 'is': 'isl',
2913 'it': 'ita',
2914 'iu': 'iku',
2915 'ja': 'jpn',
2916 'jv': 'jav',
2917 'ka': 'kat',
2918 'kg': 'kon',
2919 'ki': 'kik',
2920 'kj': 'kua',
2921 'kk': 'kaz',
2922 'kl': 'kal',
2923 'km': 'khm',
2924 'kn': 'kan',
2925 'ko': 'kor',
2926 'kr': 'kau',
2927 'ks': 'kas',
2928 'ku': 'kur',
2929 'kv': 'kom',
2930 'kw': 'cor',
2931 'ky': 'kir',
2932 'la': 'lat',
2933 'lb': 'ltz',
2934 'lg': 'lug',
2935 'li': 'lim',
2936 'ln': 'lin',
2937 'lo': 'lao',
2938 'lt': 'lit',
2939 'lu': 'lub',
2940 'lv': 'lav',
2941 'mg': 'mlg',
2942 'mh': 'mah',
2943 'mi': 'mri',
2944 'mk': 'mkd',
2945 'ml': 'mal',
2946 'mn': 'mon',
2947 'mr': 'mar',
2948 'ms': 'msa',
2949 'mt': 'mlt',
2950 'my': 'mya',
2951 'na': 'nau',
2952 'nb': 'nob',
2953 'nd': 'nde',
2954 'ne': 'nep',
2955 'ng': 'ndo',
2956 'nl': 'nld',
2957 'nn': 'nno',
2958 'no': 'nor',
2959 'nr': 'nbl',
2960 'nv': 'nav',
2961 'ny': 'nya',
2962 'oc': 'oci',
2963 'oj': 'oji',
2964 'om': 'orm',
2965 'or': 'ori',
2966 'os': 'oss',
2967 'pa': 'pan',
2968 'pi': 'pli',
2969 'pl': 'pol',
2970 'ps': 'pus',
2971 'pt': 'por',
2972 'qu': 'que',
2973 'rm': 'roh',
2974 'rn': 'run',
2975 'ro': 'ron',
2976 'ru': 'rus',
2977 'rw': 'kin',
2978 'sa': 'san',
2979 'sc': 'srd',
2980 'sd': 'snd',
2981 'se': 'sme',
2982 'sg': 'sag',
2983 'si': 'sin',
2984 'sk': 'slk',
2985 'sl': 'slv',
2986 'sm': 'smo',
2987 'sn': 'sna',
2988 'so': 'som',
2989 'sq': 'sqi',
2990 'sr': 'srp',
2991 'ss': 'ssw',
2992 'st': 'sot',
2993 'su': 'sun',
2994 'sv': 'swe',
2995 'sw': 'swa',
2996 'ta': 'tam',
2997 'te': 'tel',
2998 'tg': 'tgk',
2999 'th': 'tha',
3000 'ti': 'tir',
3001 'tk': 'tuk',
3002 'tl': 'tgl',
3003 'tn': 'tsn',
3004 'to': 'ton',
3005 'tr': 'tur',
3006 'ts': 'tso',
3007 'tt': 'tat',
3008 'tw': 'twi',
3009 'ty': 'tah',
3010 'ug': 'uig',
3011 'uk': 'ukr',
3012 'ur': 'urd',
3013 'uz': 'uzb',
3014 've': 'ven',
3015 'vi': 'vie',
3016 'vo': 'vol',
3017 'wa': 'wln',
3018 'wo': 'wol',
3019 'xh': 'xho',
3020 'yi': 'yid',
3021 'yo': 'yor',
3022 'za': 'zha',
3023 'zh': 'zho',
3024 'zu': 'zul',
3025 }
3026
3027 @classmethod
3028 def short2long(cls, code):
3029 """Convert language code from ISO 639-1 to ISO 639-2/T"""
3030 return cls._lang_map.get(code[:2])
3031
3032 @classmethod
3033 def long2short(cls, code):
3034 """Convert language code from ISO 639-2/T to ISO 639-1"""
3035 for short_name, long_name in cls._lang_map.items():
3036 if long_name == code:
3037 return short_name
3038
3039
3040 class ISO3166Utils(object):
3041 # From http://data.okfn.org/data/core/country-list
3042 _country_map = {
3043 'AF': 'Afghanistan',
3044 'AX': 'Åland Islands',
3045 'AL': 'Albania',
3046 'DZ': 'Algeria',
3047 'AS': 'American Samoa',
3048 'AD': 'Andorra',
3049 'AO': 'Angola',
3050 'AI': 'Anguilla',
3051 'AQ': 'Antarctica',
3052 'AG': 'Antigua and Barbuda',
3053 'AR': 'Argentina',
3054 'AM': 'Armenia',
3055 'AW': 'Aruba',
3056 'AU': 'Australia',
3057 'AT': 'Austria',
3058 'AZ': 'Azerbaijan',
3059 'BS': 'Bahamas',
3060 'BH': 'Bahrain',
3061 'BD': 'Bangladesh',
3062 'BB': 'Barbados',
3063 'BY': 'Belarus',
3064 'BE': 'Belgium',
3065 'BZ': 'Belize',
3066 'BJ': 'Benin',
3067 'BM': 'Bermuda',
3068 'BT': 'Bhutan',
3069 'BO': 'Bolivia, Plurinational State of',
3070 'BQ': 'Bonaire, Sint Eustatius and Saba',
3071 'BA': 'Bosnia and Herzegovina',
3072 'BW': 'Botswana',
3073 'BV': 'Bouvet Island',
3074 'BR': 'Brazil',
3075 'IO': 'British Indian Ocean Territory',
3076 'BN': 'Brunei Darussalam',
3077 'BG': 'Bulgaria',
3078 'BF': 'Burkina Faso',
3079 'BI': 'Burundi',
3080 'KH': 'Cambodia',
3081 'CM': 'Cameroon',
3082 'CA': 'Canada',
3083 'CV': 'Cape Verde',
3084 'KY': 'Cayman Islands',
3085 'CF': 'Central African Republic',
3086 'TD': 'Chad',
3087 'CL': 'Chile',
3088 'CN': 'China',
3089 'CX': 'Christmas Island',
3090 'CC': 'Cocos (Keeling) Islands',
3091 'CO': 'Colombia',
3092 'KM': 'Comoros',
3093 'CG': 'Congo',
3094 'CD': 'Congo, the Democratic Republic of the',
3095 'CK': 'Cook Islands',
3096 'CR': 'Costa Rica',
3097 'CI': 'Côte d\'Ivoire',
3098 'HR': 'Croatia',
3099 'CU': 'Cuba',
3100 'CW': 'Curaçao',
3101 'CY': 'Cyprus',
3102 'CZ': 'Czech Republic',
3103 'DK': 'Denmark',
3104 'DJ': 'Djibouti',
3105 'DM': 'Dominica',
3106 'DO': 'Dominican Republic',
3107 'EC': 'Ecuador',
3108 'EG': 'Egypt',
3109 'SV': 'El Salvador',
3110 'GQ': 'Equatorial Guinea',
3111 'ER': 'Eritrea',
3112 'EE': 'Estonia',
3113 'ET': 'Ethiopia',
3114 'FK': 'Falkland Islands (Malvinas)',
3115 'FO': 'Faroe Islands',
3116 'FJ': 'Fiji',
3117 'FI': 'Finland',
3118 'FR': 'France',
3119 'GF': 'French Guiana',
3120 'PF': 'French Polynesia',
3121 'TF': 'French Southern Territories',
3122 'GA': 'Gabon',
3123 'GM': 'Gambia',
3124 'GE': 'Georgia',
3125 'DE': 'Germany',
3126 'GH': 'Ghana',
3127 'GI': 'Gibraltar',
3128 'GR': 'Greece',
3129 'GL': 'Greenland',
3130 'GD': 'Grenada',
3131 'GP': 'Guadeloupe',
3132 'GU': 'Guam',
3133 'GT': 'Guatemala',
3134 'GG': 'Guernsey',
3135 'GN': 'Guinea',
3136 'GW': 'Guinea-Bissau',
3137 'GY': 'Guyana',
3138 'HT': 'Haiti',
3139 'HM': 'Heard Island and McDonald Islands',
3140 'VA': 'Holy See (Vatican City State)',
3141 'HN': 'Honduras',
3142 'HK': 'Hong Kong',
3143 'HU': 'Hungary',
3144 'IS': 'Iceland',
3145 'IN': 'India',
3146 'ID': 'Indonesia',
3147 'IR': 'Iran, Islamic Republic of',
3148 'IQ': 'Iraq',
3149 'IE': 'Ireland',
3150 'IM': 'Isle of Man',
3151 'IL': 'Israel',
3152 'IT': 'Italy',
3153 'JM': 'Jamaica',
3154 'JP': 'Japan',
3155 'JE': 'Jersey',
3156 'JO': 'Jordan',
3157 'KZ': 'Kazakhstan',
3158 'KE': 'Kenya',
3159 'KI': 'Kiribati',
3160 'KP': 'Korea, Democratic People\'s Republic of',
3161 'KR': 'Korea, Republic of',
3162 'KW': 'Kuwait',
3163 'KG': 'Kyrgyzstan',
3164 'LA': 'Lao People\'s Democratic Republic',
3165 'LV': 'Latvia',
3166 'LB': 'Lebanon',
3167 'LS': 'Lesotho',
3168 'LR': 'Liberia',
3169 'LY': 'Libya',
3170 'LI': 'Liechtenstein',
3171 'LT': 'Lithuania',
3172 'LU': 'Luxembourg',
3173 'MO': 'Macao',
3174 'MK': 'Macedonia, the Former Yugoslav Republic of',
3175 'MG': 'Madagascar',
3176 'MW': 'Malawi',
3177 'MY': 'Malaysia',
3178 'MV': 'Maldives',
3179 'ML': 'Mali',
3180 'MT': 'Malta',
3181 'MH': 'Marshall Islands',
3182 'MQ': 'Martinique',
3183 'MR': 'Mauritania',
3184 'MU': 'Mauritius',
3185 'YT': 'Mayotte',
3186 'MX': 'Mexico',
3187 'FM': 'Micronesia, Federated States of',
3188 'MD': 'Moldova, Republic of',
3189 'MC': 'Monaco',
3190 'MN': 'Mongolia',
3191 'ME': 'Montenegro',
3192 'MS': 'Montserrat',
3193 'MA': 'Morocco',
3194 'MZ': 'Mozambique',
3195 'MM': 'Myanmar',
3196 'NA': 'Namibia',
3197 'NR': 'Nauru',
3198 'NP': 'Nepal',
3199 'NL': 'Netherlands',
3200 'NC': 'New Caledonia',
3201 'NZ': 'New Zealand',
3202 'NI': 'Nicaragua',
3203 'NE': 'Niger',
3204 'NG': 'Nigeria',
3205 'NU': 'Niue',
3206 'NF': 'Norfolk Island',
3207 'MP': 'Northern Mariana Islands',
3208 'NO': 'Norway',
3209 'OM': 'Oman',
3210 'PK': 'Pakistan',
3211 'PW': 'Palau',
3212 'PS': 'Palestine, State of',
3213 'PA': 'Panama',
3214 'PG': 'Papua New Guinea',
3215 'PY': 'Paraguay',
3216 'PE': 'Peru',
3217 'PH': 'Philippines',
3218 'PN': 'Pitcairn',
3219 'PL': 'Poland',
3220 'PT': 'Portugal',
3221 'PR': 'Puerto Rico',
3222 'QA': 'Qatar',
3223 'RE': 'Réunion',
3224 'RO': 'Romania',
3225 'RU': 'Russian Federation',
3226 'RW': 'Rwanda',
3227 'BL': 'Saint Barthélemy',
3228 'SH': 'Saint Helena, Ascension and Tristan da Cunha',
3229 'KN': 'Saint Kitts and Nevis',
3230 'LC': 'Saint Lucia',
3231 'MF': 'Saint Martin (French part)',
3232 'PM': 'Saint Pierre and Miquelon',
3233 'VC': 'Saint Vincent and the Grenadines',
3234 'WS': 'Samoa',
3235 'SM': 'San Marino',
3236 'ST': 'Sao Tome and Principe',
3237 'SA': 'Saudi Arabia',
3238 'SN': 'Senegal',
3239 'RS': 'Serbia',
3240 'SC': 'Seychelles',
3241 'SL': 'Sierra Leone',
3242 'SG': 'Singapore',
3243 'SX': 'Sint Maarten (Dutch part)',
3244 'SK': 'Slovakia',
3245 'SI': 'Slovenia',
3246 'SB': 'Solomon Islands',
3247 'SO': 'Somalia',
3248 'ZA': 'South Africa',
3249 'GS': 'South Georgia and the South Sandwich Islands',
3250 'SS': 'South Sudan',
3251 'ES': 'Spain',
3252 'LK': 'Sri Lanka',
3253 'SD': 'Sudan',
3254 'SR': 'Suriname',
3255 'SJ': 'Svalbard and Jan Mayen',
3256 'SZ': 'Swaziland',
3257 'SE': 'Sweden',
3258 'CH': 'Switzerland',
3259 'SY': 'Syrian Arab Republic',
3260 'TW': 'Taiwan, Province of China',
3261 'TJ': 'Tajikistan',
3262 'TZ': 'Tanzania, United Republic of',
3263 'TH': 'Thailand',
3264 'TL': 'Timor-Leste',
3265 'TG': 'Togo',
3266 'TK': 'Tokelau',
3267 'TO': 'Tonga',
3268 'TT': 'Trinidad and Tobago',
3269 'TN': 'Tunisia',
3270 'TR': 'Turkey',
3271 'TM': 'Turkmenistan',
3272 'TC': 'Turks and Caicos Islands',
3273 'TV': 'Tuvalu',
3274 'UG': 'Uganda',
3275 'UA': 'Ukraine',
3276 'AE': 'United Arab Emirates',
3277 'GB': 'United Kingdom',
3278 'US': 'United States',
3279 'UM': 'United States Minor Outlying Islands',
3280 'UY': 'Uruguay',
3281 'UZ': 'Uzbekistan',
3282 'VU': 'Vanuatu',
3283 'VE': 'Venezuela, Bolivarian Republic of',
3284 'VN': 'Viet Nam',
3285 'VG': 'Virgin Islands, British',
3286 'VI': 'Virgin Islands, U.S.',
3287 'WF': 'Wallis and Futuna',
3288 'EH': 'Western Sahara',
3289 'YE': 'Yemen',
3290 'ZM': 'Zambia',
3291 'ZW': 'Zimbabwe',
3292 }
3293
3294 @classmethod
3295 def short2full(cls, code):
3296 """Convert an ISO 3166-2 country code to the corresponding full name"""
3297 return cls._country_map.get(code.upper())
3298
3299
3300 class GeoUtils(object):
3301 # Major IPv4 address blocks per country
3302 _country_ip_map = {
3303 'AD': '85.94.160.0/19',
3304 'AE': '94.200.0.0/13',
3305 'AF': '149.54.0.0/17',
3306 'AG': '209.59.64.0/18',
3307 'AI': '204.14.248.0/21',
3308 'AL': '46.99.0.0/16',
3309 'AM': '46.70.0.0/15',
3310 'AO': '105.168.0.0/13',
3311 'AP': '159.117.192.0/21',
3312 'AR': '181.0.0.0/12',
3313 'AS': '202.70.112.0/20',
3314 'AT': '84.112.0.0/13',
3315 'AU': '1.128.0.0/11',
3316 'AW': '181.41.0.0/18',
3317 'AZ': '5.191.0.0/16',
3318 'BA': '31.176.128.0/17',
3319 'BB': '65.48.128.0/17',
3320 'BD': '114.130.0.0/16',
3321 'BE': '57.0.0.0/8',
3322 'BF': '129.45.128.0/17',
3323 'BG': '95.42.0.0/15',
3324 'BH': '37.131.0.0/17',
3325 'BI': '154.117.192.0/18',
3326 'BJ': '137.255.0.0/16',
3327 'BL': '192.131.134.0/24',
3328 'BM': '196.12.64.0/18',
3329 'BN': '156.31.0.0/16',
3330 'BO': '161.56.0.0/16',
3331 'BQ': '161.0.80.0/20',
3332 'BR': '152.240.0.0/12',
3333 'BS': '24.51.64.0/18',
3334 'BT': '119.2.96.0/19',
3335 'BW': '168.167.0.0/16',
3336 'BY': '178.120.0.0/13',
3337 'BZ': '179.42.192.0/18',
3338 'CA': '99.224.0.0/11',
3339 'CD': '41.243.0.0/16',
3340 'CF': '196.32.200.0/21',
3341 'CG': '197.214.128.0/17',
3342 'CH': '85.0.0.0/13',
3343 'CI': '154.232.0.0/14',
3344 'CK': '202.65.32.0/19',
3345 'CL': '152.172.0.0/14',
3346 'CM': '165.210.0.0/15',
3347 'CN': '36.128.0.0/10',
3348 'CO': '181.240.0.0/12',
3349 'CR': '201.192.0.0/12',
3350 'CU': '152.206.0.0/15',
3351 'CV': '165.90.96.0/19',
3352 'CW': '190.88.128.0/17',
3353 'CY': '46.198.0.0/15',
3354 'CZ': '88.100.0.0/14',
3355 'DE': '53.0.0.0/8',
3356 'DJ': '197.241.0.0/17',
3357 'DK': '87.48.0.0/12',
3358 'DM': '192.243.48.0/20',
3359 'DO': '152.166.0.0/15',
3360 'DZ': '41.96.0.0/12',
3361 'EC': '186.68.0.0/15',
3362 'EE': '90.190.0.0/15',
3363 'EG': '156.160.0.0/11',
3364 'ER': '196.200.96.0/20',
3365 'ES': '88.0.0.0/11',
3366 'ET': '196.188.0.0/14',
3367 'EU': '2.16.0.0/13',
3368 'FI': '91.152.0.0/13',
3369 'FJ': '144.120.0.0/16',
3370 'FM': '119.252.112.0/20',
3371 'FO': '88.85.32.0/19',
3372 'FR': '90.0.0.0/9',
3373 'GA': '41.158.0.0/15',
3374 'GB': '25.0.0.0/8',
3375 'GD': '74.122.88.0/21',
3376 'GE': '31.146.0.0/16',
3377 'GF': '161.22.64.0/18',
3378 'GG': '62.68.160.0/19',
3379 'GH': '45.208.0.0/14',
3380 'GI': '85.115.128.0/19',
3381 'GL': '88.83.0.0/19',
3382 'GM': '160.182.0.0/15',
3383 'GN': '197.149.192.0/18',
3384 'GP': '104.250.0.0/19',
3385 'GQ': '105.235.224.0/20',
3386 'GR': '94.64.0.0/13',
3387 'GT': '168.234.0.0/16',
3388 'GU': '168.123.0.0/16',
3389 'GW': '197.214.80.0/20',
3390 'GY': '181.41.64.0/18',
3391 'HK': '113.252.0.0/14',
3392 'HN': '181.210.0.0/16',
3393 'HR': '93.136.0.0/13',
3394 'HT': '148.102.128.0/17',
3395 'HU': '84.0.0.0/14',
3396 'ID': '39.192.0.0/10',
3397 'IE': '87.32.0.0/12',
3398 'IL': '79.176.0.0/13',
3399 'IM': '5.62.80.0/20',
3400 'IN': '117.192.0.0/10',
3401 'IO': '203.83.48.0/21',
3402 'IQ': '37.236.0.0/14',
3403 'IR': '2.176.0.0/12',
3404 'IS': '82.221.0.0/16',
3405 'IT': '79.0.0.0/10',
3406 'JE': '87.244.64.0/18',
3407 'JM': '72.27.0.0/17',
3408 'JO': '176.29.0.0/16',
3409 'JP': '126.0.0.0/8',
3410 'KE': '105.48.0.0/12',
3411 'KG': '158.181.128.0/17',
3412 'KH': '36.37.128.0/17',
3413 'KI': '103.25.140.0/22',
3414 'KM': '197.255.224.0/20',
3415 'KN': '198.32.32.0/19',
3416 'KP': '175.45.176.0/22',
3417 'KR': '175.192.0.0/10',
3418 'KW': '37.36.0.0/14',
3419 'KY': '64.96.0.0/15',
3420 'KZ': '2.72.0.0/13',
3421 'LA': '115.84.64.0/18',
3422 'LB': '178.135.0.0/16',
3423 'LC': '192.147.231.0/24',
3424 'LI': '82.117.0.0/19',
3425 'LK': '112.134.0.0/15',
3426 'LR': '41.86.0.0/19',
3427 'LS': '129.232.0.0/17',
3428 'LT': '78.56.0.0/13',
3429 'LU': '188.42.0.0/16',
3430 'LV': '46.109.0.0/16',
3431 'LY': '41.252.0.0/14',
3432 'MA': '105.128.0.0/11',
3433 'MC': '88.209.64.0/18',
3434 'MD': '37.246.0.0/16',
3435 'ME': '178.175.0.0/17',
3436 'MF': '74.112.232.0/21',
3437 'MG': '154.126.0.0/17',
3438 'MH': '117.103.88.0/21',
3439 'MK': '77.28.0.0/15',
3440 'ML': '154.118.128.0/18',
3441 'MM': '37.111.0.0/17',
3442 'MN': '49.0.128.0/17',
3443 'MO': '60.246.0.0/16',
3444 'MP': '202.88.64.0/20',
3445 'MQ': '109.203.224.0/19',
3446 'MR': '41.188.64.0/18',
3447 'MS': '208.90.112.0/22',
3448 'MT': '46.11.0.0/16',
3449 'MU': '105.16.0.0/12',
3450 'MV': '27.114.128.0/18',
3451 'MW': '105.234.0.0/16',
3452 'MX': '187.192.0.0/11',
3453 'MY': '175.136.0.0/13',
3454 'MZ': '197.218.0.0/15',
3455 'NA': '41.182.0.0/16',
3456 'NC': '101.101.0.0/18',
3457 'NE': '197.214.0.0/18',
3458 'NF': '203.17.240.0/22',
3459 'NG': '105.112.0.0/12',
3460 'NI': '186.76.0.0/15',
3461 'NL': '145.96.0.0/11',
3462 'NO': '84.208.0.0/13',
3463 'NP': '36.252.0.0/15',
3464 'NR': '203.98.224.0/19',
3465 'NU': '49.156.48.0/22',
3466 'NZ': '49.224.0.0/14',
3467 'OM': '5.36.0.0/15',
3468 'PA': '186.72.0.0/15',
3469 'PE': '186.160.0.0/14',
3470 'PF': '123.50.64.0/18',
3471 'PG': '124.240.192.0/19',
3472 'PH': '49.144.0.0/13',
3473 'PK': '39.32.0.0/11',
3474 'PL': '83.0.0.0/11',
3475 'PM': '70.36.0.0/20',
3476 'PR': '66.50.0.0/16',
3477 'PS': '188.161.0.0/16',
3478 'PT': '85.240.0.0/13',
3479 'PW': '202.124.224.0/20',
3480 'PY': '181.120.0.0/14',
3481 'QA': '37.210.0.0/15',
3482 'RE': '139.26.0.0/16',
3483 'RO': '79.112.0.0/13',
3484 'RS': '178.220.0.0/14',
3485 'RU': '5.136.0.0/13',
3486 'RW': '105.178.0.0/15',
3487 'SA': '188.48.0.0/13',
3488 'SB': '202.1.160.0/19',
3489 'SC': '154.192.0.0/11',
3490 'SD': '154.96.0.0/13',
3491 'SE': '78.64.0.0/12',
3492 'SG': '152.56.0.0/14',
3493 'SI': '188.196.0.0/14',
3494 'SK': '78.98.0.0/15',
3495 'SL': '197.215.0.0/17',
3496 'SM': '89.186.32.0/19',
3497 'SN': '41.82.0.0/15',
3498 'SO': '197.220.64.0/19',
3499 'SR': '186.179.128.0/17',
3500 'SS': '105.235.208.0/21',
3501 'ST': '197.159.160.0/19',
3502 'SV': '168.243.0.0/16',
3503 'SX': '190.102.0.0/20',
3504 'SY': '5.0.0.0/16',
3505 'SZ': '41.84.224.0/19',
3506 'TC': '65.255.48.0/20',
3507 'TD': '154.68.128.0/19',
3508 'TG': '196.168.0.0/14',
3509 'TH': '171.96.0.0/13',
3510 'TJ': '85.9.128.0/18',
3511 'TK': '27.96.24.0/21',
3512 'TL': '180.189.160.0/20',
3513 'TM': '95.85.96.0/19',
3514 'TN': '197.0.0.0/11',
3515 'TO': '175.176.144.0/21',
3516 'TR': '78.160.0.0/11',
3517 'TT': '186.44.0.0/15',
3518 'TV': '202.2.96.0/19',
3519 'TW': '120.96.0.0/11',
3520 'TZ': '156.156.0.0/14',
3521 'UA': '93.72.0.0/13',
3522 'UG': '154.224.0.0/13',
3523 'US': '3.0.0.0/8',
3524 'UY': '167.56.0.0/13',
3525 'UZ': '82.215.64.0/18',
3526 'VA': '212.77.0.0/19',
3527 'VC': '24.92.144.0/20',
3528 'VE': '186.88.0.0/13',
3529 'VG': '172.103.64.0/18',
3530 'VI': '146.226.0.0/16',
3531 'VN': '14.160.0.0/11',
3532 'VU': '202.80.32.0/20',
3533 'WF': '117.20.32.0/21',
3534 'WS': '202.4.32.0/19',
3535 'YE': '134.35.0.0/16',
3536 'YT': '41.242.116.0/22',
3537 'ZA': '41.0.0.0/11',
3538 'ZM': '165.56.0.0/13',
3539 'ZW': '41.85.192.0/19',
3540 }
3541
3542 @classmethod
3543 def random_ipv4(cls, code_or_block):
3544 if len(code_or_block) == 2:
3545 block = cls._country_ip_map.get(code_or_block.upper())
3546 if not block:
3547 return None
3548 else:
3549 block = code_or_block
3550 addr, preflen = block.split('/')
3551 addr_min = compat_struct_unpack('!L', socket.inet_aton(addr))[0]
3552 addr_max = addr_min | (0xffffffff >> int(preflen))
3553 return compat_str(socket.inet_ntoa(
3554 compat_struct_pack('!L', random.randint(addr_min, addr_max))))
3555
3556
3557 class PerRequestProxyHandler(compat_urllib_request.ProxyHandler):
3558 def __init__(self, proxies=None):
3559 # Set default handlers
3560 for type in ('http', 'https'):
3561 setattr(self, '%s_open' % type,
3562 lambda r, proxy='__noproxy__', type=type, meth=self.proxy_open:
3563 meth(r, proxy, type))
3564 return compat_urllib_request.ProxyHandler.__init__(self, proxies)
3565
3566 def proxy_open(self, req, proxy, type):
3567 req_proxy = req.headers.get('Ytdl-request-proxy')
3568 if req_proxy is not None:
3569 proxy = req_proxy
3570 del req.headers['Ytdl-request-proxy']
3571
3572 if proxy == '__noproxy__':
3573 return None # No Proxy
3574 if compat_urlparse.urlparse(proxy).scheme.lower() in ('socks', 'socks4', 'socks4a', 'socks5'):
3575 req.add_header('Ytdl-socks-proxy', proxy)
3576 # youtube-dl's http/https handlers do wrapping the socket with socks
3577 return None
3578 return compat_urllib_request.ProxyHandler.proxy_open(
3579 self, req, proxy, type)
3580
3581
3582 # Both long_to_bytes and bytes_to_long are adapted from PyCrypto, which is
3583 # released into Public Domain
3584 # https://github.com/dlitz/pycrypto/blob/master/lib/Crypto/Util/number.py#L387
3585
3586 def long_to_bytes(n, blocksize=0):
3587 """long_to_bytes(n:long, blocksize:int) : string
3588 Convert a long integer to a byte string.
3589
3590 If optional blocksize is given and greater than zero, pad the front of the
3591 byte string with binary zeros so that the length is a multiple of
3592 blocksize.
3593 """
3594 # after much testing, this algorithm was deemed to be the fastest
3595 s = b''
3596 n = int(n)
3597 while n > 0:
3598 s = compat_struct_pack('>I', n & 0xffffffff) + s
3599 n = n >> 32
3600 # strip off leading zeros
3601 for i in range(len(s)):
3602 if s[i] != b'\000'[0]:
3603 break
3604 else:
3605 # only happens when n == 0
3606 s = b'\000'
3607 i = 0
3608 s = s[i:]
3609 # add back some pad bytes. this could be done more efficiently w.r.t. the
3610 # de-padding being done above, but sigh...
3611 if blocksize > 0 and len(s) % blocksize:
3612 s = (blocksize - len(s) % blocksize) * b'\000' + s
3613 return s
3614
3615
3616 def bytes_to_long(s):
3617 """bytes_to_long(string) : long
3618 Convert a byte string to a long integer.
3619
3620 This is (essentially) the inverse of long_to_bytes().
3621 """
3622 acc = 0
3623 length = len(s)
3624 if length % 4:
3625 extra = (4 - length % 4)
3626 s = b'\000' * extra + s
3627 length = length + extra
3628 for i in range(0, length, 4):
3629 acc = (acc << 32) + compat_struct_unpack('>I', s[i:i + 4])[0]
3630 return acc
3631
3632
3633 def ohdave_rsa_encrypt(data, exponent, modulus):
3634 '''
3635 Implement OHDave's RSA algorithm. See http://www.ohdave.com/rsa/
3636
3637 Input:
3638 data: data to encrypt, bytes-like object
3639 exponent, modulus: parameter e and N of RSA algorithm, both integer
3640 Output: hex string of encrypted data
3641
3642 Limitation: supports one block encryption only
3643 '''
3644
3645 payload = int(binascii.hexlify(data[::-1]), 16)
3646 encrypted = pow(payload, exponent, modulus)
3647 return '%x' % encrypted
3648
3649
3650 def pkcs1pad(data, length):
3651 """
3652 Padding input data with PKCS#1 scheme
3653
3654 @param {int[]} data input data
3655 @param {int} length target length
3656 @returns {int[]} padded data
3657 """
3658 if len(data) > length - 11:
3659 raise ValueError('Input data too long for PKCS#1 padding')
3660
3661 pseudo_random = [random.randint(0, 254) for _ in range(length - len(data) - 3)]
3662 return [0, 2] + pseudo_random + [0] + data
3663
3664
3665 def encode_base_n(num, n, table=None):
3666 FULL_TABLE = '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ'
3667 if not table:
3668 table = FULL_TABLE[:n]
3669
3670 if n > len(table):
3671 raise ValueError('base %d exceeds table length %d' % (n, len(table)))
3672
3673 if num == 0:
3674 return table[0]
3675
3676 ret = ''
3677 while num:
3678 ret = table[num % n] + ret
3679 num = num // n
3680 return ret
3681
3682
3683 def decode_packed_codes(code):
3684 mobj = re.search(PACKED_CODES_RE, code)
3685 obfucasted_code, base, count, symbols = mobj.groups()
3686 base = int(base)
3687 count = int(count)
3688 symbols = symbols.split('|')
3689 symbol_table = {}
3690
3691 while count:
3692 count -= 1
3693 base_n_count = encode_base_n(count, base)
3694 symbol_table[base_n_count] = symbols[count] or base_n_count
3695
3696 return re.sub(
3697 r'\b(\w+)\b', lambda mobj: symbol_table[mobj.group(0)],
3698 obfucasted_code)
3699
3700
3701 def parse_m3u8_attributes(attrib):
3702 info = {}
3703 for (key, val) in re.findall(r'(?P<key>[A-Z0-9-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)', attrib):
3704 if val.startswith('"'):
3705 val = val[1:-1]
3706 info[key] = val
3707 return info
3708
3709
3710 def urshift(val, n):
3711 return val >> n if val >= 0 else (val + 0x100000000) >> n
3712
3713
3714 # Based on png2str() written by @gdkchan and improved by @yokrysty
3715 # Originally posted at https://github.com/rg3/youtube-dl/issues/9706
3716 def decode_png(png_data):
3717 # Reference: https://www.w3.org/TR/PNG/
3718 header = png_data[8:]
3719
3720 if png_data[:8] != b'\x89PNG\x0d\x0a\x1a\x0a' or header[4:8] != b'IHDR':
3721 raise IOError('Not a valid PNG file.')
3722
3723 int_map = {1: '>B', 2: '>H', 4: '>I'}
3724 unpack_integer = lambda x: compat_struct_unpack(int_map[len(x)], x)[0]
3725
3726 chunks = []
3727
3728 while header:
3729 length = unpack_integer(header[:4])
3730 header = header[4:]
3731
3732 chunk_type = header[:4]
3733 header = header[4:]
3734
3735 chunk_data = header[:length]
3736 header = header[length:]
3737
3738 header = header[4:] # Skip CRC
3739
3740 chunks.append({
3741 'type': chunk_type,
3742 'length': length,
3743 'data': chunk_data
3744 })
3745
3746 ihdr = chunks[0]['data']
3747
3748 width = unpack_integer(ihdr[:4])
3749 height = unpack_integer(ihdr[4:8])
3750
3751 idat = b''
3752
3753 for chunk in chunks:
3754 if chunk['type'] == b'IDAT':
3755 idat += chunk['data']
3756
3757 if not idat:
3758 raise IOError('Unable to read PNG data.')
3759
3760 decompressed_data = bytearray(zlib.decompress(idat))
3761
3762 stride = width * 3
3763 pixels = []
3764
3765 def _get_pixel(idx):
3766 x = idx % stride
3767 y = idx // stride
3768 return pixels[y][x]
3769
3770 for y in range(height):
3771 basePos = y * (1 + stride)
3772 filter_type = decompressed_data[basePos]
3773
3774 current_row = []
3775
3776 pixels.append(current_row)
3777
3778 for x in range(stride):
3779 color = decompressed_data[1 + basePos + x]
3780 basex = y * stride + x
3781 left = 0
3782 up = 0
3783
3784 if x > 2:
3785 left = _get_pixel(basex - 3)
3786 if y > 0:
3787 up = _get_pixel(basex - stride)
3788
3789 if filter_type == 1: # Sub
3790 color = (color + left) & 0xff
3791 elif filter_type == 2: # Up
3792 color = (color + up) & 0xff
3793 elif filter_type == 3: # Average
3794 color = (color + ((left + up) >> 1)) & 0xff
3795 elif filter_type == 4: # Paeth
3796 a = left
3797 b = up
3798 c = 0
3799
3800 if x > 2 and y > 0:
3801 c = _get_pixel(basex - stride - 3)
3802
3803 p = a + b - c
3804
3805 pa = abs(p - a)
3806 pb = abs(p - b)
3807 pc = abs(p - c)
3808
3809 if pa <= pb and pa <= pc:
3810 color = (color + a) & 0xff
3811 elif pb <= pc:
3812 color = (color + b) & 0xff
3813 else:
3814 color = (color + c) & 0xff
3815
3816 current_row.append(color)
3817
3818 return width, height, pixels
3819
3820
3821 def write_xattr(path, key, value):
3822 # This mess below finds the best xattr tool for the job
3823 try:
3824 # try the pyxattr module...
3825 import xattr
3826
3827 if hasattr(xattr, 'set'): # pyxattr
3828 # Unicode arguments are not supported in python-pyxattr until
3829 # version 0.5.0
3830 # See https://github.com/rg3/youtube-dl/issues/5498
3831 pyxattr_required_version = '0.5.0'
3832 if version_tuple(xattr.__version__) < version_tuple(pyxattr_required_version):
3833 # TODO: fallback to CLI tools
3834 raise XAttrUnavailableError(
3835 'python-pyxattr is detected but is too old. '
3836 'youtube-dl requires %s or above while your version is %s. '
3837 'Falling back to other xattr implementations' % (
3838 pyxattr_required_version, xattr.__version__))
3839
3840 setxattr = xattr.set
3841 else: # xattr
3842 setxattr = xattr.setxattr
3843
3844 try:
3845 setxattr(path, key, value)
3846 except EnvironmentError as e:
3847 raise XAttrMetadataError(e.errno, e.strerror)
3848
3849 except ImportError:
3850 if compat_os_name == 'nt':
3851 # Write xattrs to NTFS Alternate Data Streams:
3852 # http://en.wikipedia.org/wiki/NTFS#Alternate_data_streams_.28ADS.29
3853 assert ':' not in key
3854 assert os.path.exists(path)
3855
3856 ads_fn = path + ':' + key
3857 try:
3858 with open(ads_fn, 'wb') as f:
3859 f.write(value)
3860 except EnvironmentError as e:
3861 raise XAttrMetadataError(e.errno, e.strerror)
3862 else:
3863 user_has_setfattr = check_executable('setfattr', ['--version'])
3864 user_has_xattr = check_executable('xattr', ['-h'])
3865
3866 if user_has_setfattr or user_has_xattr:
3867
3868 value = value.decode('utf-8')
3869 if user_has_setfattr:
3870 executable = 'setfattr'
3871 opts = ['-n', key, '-v', value]
3872 elif user_has_xattr:
3873 executable = 'xattr'
3874 opts = ['-w', key, value]
3875
3876 cmd = ([encodeFilename(executable, True)] +
3877 [encodeArgument(o) for o in opts] +
3878 [encodeFilename(path, True)])
3879
3880 try:
3881 p = subprocess.Popen(
3882 cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE, stdin=subprocess.PIPE)
3883 except EnvironmentError as e:
3884 raise XAttrMetadataError(e.errno, e.strerror)
3885 stdout, stderr = p.communicate()
3886 stderr = stderr.decode('utf-8', 'replace')
3887 if p.returncode != 0:
3888 raise XAttrMetadataError(p.returncode, stderr)
3889
3890 else:
3891 # On Unix, and can't find pyxattr, setfattr, or xattr.
3892 if sys.platform.startswith('linux'):
3893 raise XAttrUnavailableError(
3894 "Couldn't find a tool to set the xattrs. "
3895 "Install either the python 'pyxattr' or 'xattr' "
3896 "modules, or the GNU 'attr' package "
3897 "(which contains the 'setfattr' tool).")
3898 else:
3899 raise XAttrUnavailableError(
3900 "Couldn't find a tool to set the xattrs. "
3901 "Install either the python 'xattr' module, "
3902 "or the 'xattr' binary.")
3903
3904
3905 def random_birthday(year_field, month_field, day_field):
3906 return {
3907 year_field: str(random.randint(1950, 1995)),
3908 month_field: str(random.randint(1, 12)),
3909 day_field: str(random.randint(1, 31)),
3910 }