import ctypes
import datetime
import email.utils
+import email.header
import errno
import functools
import gzip
import math
import operator
import os
-import pipes
import platform
import random
import re
import zlib
from .compat import (
+ compat_HTMLParseError,
compat_HTMLParser,
compat_basestring,
compat_chr,
+ compat_ctypes_WINFUNCTYPE,
compat_etree_fromstring,
+ compat_expanduser,
compat_html_entities,
compat_html_entities_html5,
compat_http_client,
compiled_regex_type = type(re.compile(''))
std_headers = {
- 'User-Agent': 'Mozilla/5.0 (X11; Linux x86_64; rv:10.0) Gecko/20150101 Firefox/47.0 (Chrome)',
+ 'User-Agent': 'Mozilla/5.0 (X11; Linux x86_64; rv:59.0) Gecko/20100101 Firefox/59.0 (Chrome)',
'Accept-Charset': 'ISO-8859-1,utf-8;q=0.7,*;q=0.7',
'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
'Accept-Encoding': 'gzip, deflate',
'%Y-%m-%dT%H:%M',
'%b %d %Y at %H:%M',
'%b %d %Y at %H:%M:%S',
+ '%B %d %Y at %H:%M',
+ '%B %d %Y at %H:%M:%S',
)
DATE_FORMATS_DAY_FIRST = list(DATE_FORMATS)
retlist = []
for m in re.finditer(r'''(?xs)
<([a-zA-Z0-9:._-]+)
- (?:\s+[a-zA-Z0-9:._-]+(?:=[a-zA-Z0-9:._-]*|="[^"]*"|='[^']*'))*?
+ (?:\s+[a-zA-Z0-9:._-]+(?:=[a-zA-Z0-9:._-]*|="[^"]*"|='[^']*'|))*?
\s+%s=['"]?%s['"]?
- (?:\s+[a-zA-Z0-9:._-]+(?:=[a-zA-Z0-9:._-]*|="[^"]*"|='[^']*'))*?
+ (?:\s+[a-zA-Z0-9:._-]+(?:=[a-zA-Z0-9:._-]*|="[^"]*"|='[^']*'|))*?
\s*>
(?P<content>.*?)
</\1>
but the cases in the unit test will work for all of 2.6, 2.7, 3.2-3.5.
"""
parser = HTMLAttributeParser()
- parser.feed(html_element)
- parser.close()
+ try:
+ parser.feed(html_element)
+ parser.close()
+ # Older Python may throw HTMLParseError in case of malformed HTML
+ except compat_HTMLParseError:
+ pass
return parser.attrs
# Newline vs <br />
html = html.replace('\n', ' ')
- html = re.sub(r'\s*<\s*br\s*/?\s*>\s*', '\n', html)
- html = re.sub(r'<\s*/\s*p\s*>\s*<\s*p[^>]*>', '\n', html)
+ html = re.sub(r'(?u)\s*<\s*br\s*/?\s*>\s*', '\n', html)
+ html = re.sub(r'(?u)<\s*/\s*p\s*>\s*<\s*p[^>]*>', '\n', html)
# Strip html tags
html = re.sub('<.*?>', '', html)
# Replace html entities
return os.path.join(*sanitized_path)
-# Prepend protocol-less URLs with `http:` scheme in order to mitigate the number of
-# unwanted failures due to missing protocol
def sanitize_url(url):
- return 'http:%s' % url if url.startswith('//') else url
+ # Prepend protocol-less URLs with `http:` scheme in order to mitigate
+ # the number of unwanted failures due to missing protocol
+ if url.startswith('//'):
+ return 'http:%s' % url
+ # Fix some common typos seen so far
+ COMMON_TYPOS = (
+ # https://github.com/rg3/youtube-dl/issues/15649
+ (r'^httpss://', r'https://'),
+ # https://bx1.be/lives/direct-tv/
+ (r'^rmtp([es]?)://', r'rtmp\1://'),
+ )
+ for mistake, fixup in COMMON_TYPOS:
+ if re.match(mistake, url):
+ return re.sub(mistake, fixup, url)
+ return url
def sanitized_Request(url, *args, **kwargs):
return compat_urllib_request.Request(sanitize_url(url), *args, **kwargs)
+def expand_path(s):
+ """Expand shell variables and ~"""
+ return os.path.expandvars(compat_expanduser(s))
+
+
def orderedSet(iterable):
""" Remove all duplicates from the input iterable """
res = []
assert type(s) == compat_str
return re.sub(
- r'&([^;]+;)', lambda m: _htmlentity_transform(m.group(1)), s)
+ r'&([^&;]+;)', lambda m: _htmlentity_transform(m.group(1)), s)
def get_subprocess_encoding():
# expected HTTP responses to meet HTTP/1.0 or later (see also
# https://github.com/rg3/youtube-dl/issues/6727)
if sys.version_info < (3, 0):
- kwargs[b'strict'] = True
- hc = http_class(*args, **kwargs)
+ kwargs['strict'] = True
+ hc = http_class(*args, **compat_kwargs(kwargs))
source_address = ydl_handler._params.get('source_address')
if source_address is not None:
sa = (source_address, 0)
except zlib.error:
return zlib.decompress(data)
- @staticmethod
- def addinfourl_wrapper(stream, headers, url, code):
- if hasattr(compat_urllib_request.addinfourl, 'getcode'):
- return compat_urllib_request.addinfourl(stream, headers, url, code)
- ret = compat_urllib_request.addinfourl(stream, headers, url)
- ret.code = code
- return ret
-
def http_request(self, req):
# According to RFC 3986, URLs can not contain non-ASCII characters, however this is not
# always respected by websites, some tend to give out URLs with non percent-encoded
break
else:
raise original_ioerror
- resp = self.addinfourl_wrapper(uncompressed, old_resp.headers, old_resp.url, old_resp.code)
+ resp = compat_urllib_request.addinfourl(uncompressed, old_resp.headers, old_resp.url, old_resp.code)
resp.msg = old_resp.msg
del resp.headers['Content-encoding']
# deflate
if resp.headers.get('Content-encoding', '') == 'deflate':
gz = io.BytesIO(self.deflate(resp.read()))
- resp = self.addinfourl_wrapper(gz, old_resp.headers, old_resp.url, old_resp.code)
+ resp = compat_urllib_request.addinfourl(gz, old_resp.headers, old_resp.url, old_resp.code)
resp.msg = old_resp.msg
del resp.headers['Content-encoding']
# Percent-encode redirect URL of Location HTTP header to satisfy RFC 3986 (see
if date_str is None:
return None
- date_str = date_str.replace(',', ' ')
+ date_str = re.sub(r'[,|]', '', date_str)
pm_delta = 12 if re.search(r'(?i)PM', date_str) else 0
timezone, date_str = extract_timezone(date_str)
# Remove AM/PM + timezone
date_str = re.sub(r'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str)
+ # Remove unrecognized timezones from ISO 8601 alike timestamps
+ m = re.search(r'\d{1,2}:\d{1,2}(?:\.\d+)?(?P<tz>\s*[A-Z]+)$', date_str)
+ if m:
+ date_str = date_str[:-len(m.group('tz'))]
+
+ # Python only supports microseconds, so remove nanoseconds
+ m = re.search(r'^([0-9]{4,}-[0-9]{1,2}-[0-9]{1,2}T[0-9]{1,2}:[0-9]{1,2}:[0-9]{1,2}\.[0-9]{6})[0-9]+$', date_str)
+ if m:
+ date_str = m.group(1)
+
for expression in date_formats(day_first):
try:
dt = datetime.datetime.strptime(date_str, expression) - timezone + datetime.timedelta(hours=pm_delta)
if fileno not in WIN_OUTPUT_IDS:
return False
- GetStdHandle = ctypes.WINFUNCTYPE(
+ GetStdHandle = compat_ctypes_WINFUNCTYPE(
ctypes.wintypes.HANDLE, ctypes.wintypes.DWORD)(
- (b'GetStdHandle', ctypes.windll.kernel32))
+ ('GetStdHandle', ctypes.windll.kernel32))
h = GetStdHandle(WIN_OUTPUT_IDS[fileno])
- WriteConsoleW = ctypes.WINFUNCTYPE(
+ WriteConsoleW = compat_ctypes_WINFUNCTYPE(
ctypes.wintypes.BOOL, ctypes.wintypes.HANDLE, ctypes.wintypes.LPWSTR,
ctypes.wintypes.DWORD, ctypes.POINTER(ctypes.wintypes.DWORD),
- ctypes.wintypes.LPVOID)((b'WriteConsoleW', ctypes.windll.kernel32))
+ ctypes.wintypes.LPVOID)(('WriteConsoleW', ctypes.windll.kernel32))
written = ctypes.wintypes.DWORD(0)
- GetFileType = ctypes.WINFUNCTYPE(ctypes.wintypes.DWORD, ctypes.wintypes.DWORD)((b'GetFileType', ctypes.windll.kernel32))
+ GetFileType = compat_ctypes_WINFUNCTYPE(ctypes.wintypes.DWORD, ctypes.wintypes.DWORD)(('GetFileType', ctypes.windll.kernel32))
FILE_TYPE_CHAR = 0x0002
FILE_TYPE_REMOTE = 0x8000
- GetConsoleMode = ctypes.WINFUNCTYPE(
+ GetConsoleMode = compat_ctypes_WINFUNCTYPE(
ctypes.wintypes.BOOL, ctypes.wintypes.HANDLE,
ctypes.POINTER(ctypes.wintypes.DWORD))(
- (b'GetConsoleMode', ctypes.windll.kernel32))
+ ('GetConsoleMode', ctypes.windll.kernel32))
INVALID_HANDLE_VALUE = ctypes.wintypes.DWORD(-1).value
def not_a_console(handle):
if isinstance(a, bytes):
# We may get a filename encoded with 'encodeFilename'
a = a.decode(encoding)
- quoted_args.append(pipes.quote(a))
+ quoted_args.append(compat_shlex_quote(a))
return ' '.join(quoted_args)
return lookup_unit_table(_UNIT_TABLE, s)
+def parse_resolution(s):
+ if s is None:
+ return {}
+
+ mobj = re.search(r'\b(?P<w>\d+)\s*[xX×]\s*(?P<h>\d+)\b', s)
+ if mobj:
+ return {
+ 'width': int(mobj.group('w')),
+ 'height': int(mobj.group('h')),
+ }
+
+ mobj = re.search(r'\b(\d+)[pPiI]\b', s)
+ if mobj:
+ return {'height': int(mobj.group(1))}
+
+ mobj = re.search(r'\b([48])[kK]\b', s)
+ if mobj:
+ return {'height': int(mobj.group(1)) * 540}
+
+ return {}
+
+
def month_by_name(name, lang='en'):
""" Return the number of a month by (locale-independently) English name """
return default
+def bool_or_none(v, default=None):
+ return v if isinstance(v, bool) else default
+
+
def strip_or_none(v):
return None if v is None else v.strip()
days, hours, mins, secs, ms = m.groups()
else:
m = re.match(
- r'''(?ix)(?:P?T)?
+ r'''(?ix)(?:P?
+ (?:
+ [0-9]+\s*y(?:ears?)?\s*
+ )?
+ (?:
+ [0-9]+\s*m(?:onths?)?\s*
+ )?
+ (?:
+ [0-9]+\s*w(?:eeks?)?\s*
+ )?
(?:
(?P<days>[0-9]+)\s*d(?:ays?)?\s*
)?
+ T)?
(?:
(?P<hours>[0-9]+)\s*h(?:ours?)?\s*
)?
class OnDemandPagedList(PagedList):
- def __init__(self, pagefunc, pagesize, use_cache=False):
+ def __init__(self, pagefunc, pagesize, use_cache=True):
self._pagefunc = pagefunc
self._pagesize = pagesize
self._use_cache = use_cache
return new_req
+def _multipart_encode_impl(data, boundary):
+ content_type = 'multipart/form-data; boundary=%s' % boundary
+
+ out = b''
+ for k, v in data.items():
+ out += b'--' + boundary.encode('ascii') + b'\r\n'
+ if isinstance(k, compat_str):
+ k = k.encode('utf-8')
+ if isinstance(v, compat_str):
+ v = v.encode('utf-8')
+ # RFC 2047 requires non-ASCII field names to be encoded, while RFC 7578
+ # suggests sending UTF-8 directly. Firefox sends UTF-8, too
+ content = b'Content-Disposition: form-data; name="' + k + b'"\r\n\r\n' + v + b'\r\n'
+ if boundary.encode('ascii') in content:
+ raise ValueError('Boundary overlaps with data')
+ out += content
+
+ out += b'--' + boundary.encode('ascii') + b'--\r\n'
+
+ return out, content_type
+
+
+def multipart_encode(data, boundary=None):
+ '''
+ Encode a dict to RFC 7578-compliant form-data
+
+ data:
+ A dict where keys and values can be either Unicode or bytes-like
+ objects.
+ boundary:
+ If specified a Unicode object, it's used as the boundary. Otherwise
+ a random boundary is generated.
+
+ Reference: https://tools.ietf.org/html/rfc7578
+ '''
+ has_specified_boundary = boundary is not None
+
+ while True:
+ if boundary is None:
+ boundary = '---------------' + str(random.randrange(0x0fffffff, 0xffffffff))
+
+ try:
+ out, content_type = _multipart_encode_impl(data, boundary)
+ break
+ except ValueError:
+ if has_specified_boundary:
+ raise
+ boundary = None
+
+ return out, content_type
+
+
def dict_get(d, key_or_keys, default=None, skip_false_values=True):
if isinstance(key_or_keys, (list, tuple)):
for key in key_or_keys:
def try_get(src, getter, expected_type=None):
- try:
- v = getter(src)
- except (AttributeError, KeyError, TypeError, IndexError):
- pass
- else:
- if expected_type is None or isinstance(v, expected_type):
- return v
+ if not isinstance(getter, (list, tuple)):
+ getter = [getter]
+ for get in getter:
+ try:
+ v = get(src)
+ except (AttributeError, KeyError, TypeError, IndexError):
+ pass
+ else:
+ if expected_type is None or isinstance(v, expected_type):
+ return v
def encode_compat_str(string, encoding=preferredencoding(), errors='strict'):
def strip_jsonp(code):
return re.sub(
- r'(?s)^[a-zA-Z0-9_.$]+\s*\(\s*(.*)\);?\s*?(?://[^\n]*)*$', r'\1', code)
+ r'''(?sx)^
+ (?:window\.)?(?P<func_name>[a-zA-Z0-9_.$]+)
+ (?:\s*&&\s*(?P=func_name))?
+ \s*\(\s*(?P<callback_data>.*)\);?
+ \s*?(?://[^\n]*)*$''',
+ r'\g<callback_data>', code)
def js_to_json(code):
"(?:[^"\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^"\\]*"|
'(?:[^'\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^'\\]*'|
{comment}|,(?={skip}[\]}}])|
- [a-zA-Z_][.a-zA-Z_0-9]*|
+ (?:(?<![0-9])[eE]|[a-df-zA-DF-Z_])[.a-zA-Z_0-9]*|
\b(?:0[xX][0-9a-fA-F]+|0+[0-7]+)(?:{skip}:)?|
[0-9]+(?={skip}:)
'''.format(comment=COMMENT_RE, skip=SKIP_RE), fix_kv, code)
return {
'3gpp': '3gp',
'smptett+xml': 'tt',
- 'srt': 'srt',
'ttaf+xml': 'dfxp',
'ttml+xml': 'ttml',
- 'vtt': 'vtt',
'x-flv': 'flv',
'x-mp4-fragmented': 'mp4',
+ 'x-ms-sami': 'sami',
'x-ms-wmv': 'wmv',
'mpegurl': 'm3u8',
'x-mpegurl': 'm3u8',
'vnd.apple.mpegurl': 'm3u8',
'dash+xml': 'mpd',
- 'f4m': 'f4m',
'f4m+xml': 'f4m',
'hds+xml': 'f4m',
'vnd.ms-sstr+xml': 'ism',
'quicktime': 'mov',
+ 'mp2t': 'ts',
}.get(res, res)
vcodec, acodec = None, None
for full_codec in splited_codecs:
codec = full_codec.split('.')[0]
- if codec in ('avc1', 'avc2', 'avc3', 'avc4', 'vp9', 'vp8', 'hev1', 'hev2', 'h263', 'h264', 'mp4v'):
+ if codec in ('avc1', 'avc2', 'avc3', 'avc4', 'vp9', 'vp8', 'hev1', 'hev2', 'h263', 'h264', 'mp4v', 'hvc1'):
if not vcodec:
vcodec = full_codec
- elif codec in ('mp4a', 'opus', 'vorbis', 'mp3', 'aac', 'ac-3'):
+ elif codec in ('mp4a', 'opus', 'vorbis', 'mp3', 'aac', 'ac-3', 'ec-3', 'eac3', 'dtsc', 'dtse', 'dtsh', 'dtsl'):
if not acodec:
acodec = full_codec
else:
- write_string('WARNING: Unknown codec %s' % full_codec, sys.stderr)
+ write_string('WARNING: Unknown codec %s\n' % full_codec, sys.stderr)
if not vcodec and not acodec:
if len(splited_codecs) == 2:
return {
def dfxp2srt(dfxp_data):
+ '''
+ @param dfxp_data A bytes-like object containing DFXP data
+ @returns A unicode object containing converted SRT data
+ '''
+ LEGACY_NAMESPACES = (
+ (b'http://www.w3.org/ns/ttml', [
+ b'http://www.w3.org/2004/11/ttaf1',
+ b'http://www.w3.org/2006/04/ttaf1',
+ b'http://www.w3.org/2006/10/ttaf1',
+ ]),
+ (b'http://www.w3.org/ns/ttml#styling', [
+ b'http://www.w3.org/ns/ttml#style',
+ ]),
+ )
+
+ SUPPORTED_STYLING = [
+ 'color',
+ 'fontFamily',
+ 'fontSize',
+ 'fontStyle',
+ 'fontWeight',
+ 'textDecoration'
+ ]
+
_x = functools.partial(xpath_with_ns, ns_map={
'ttml': 'http://www.w3.org/ns/ttml',
- 'ttaf1': 'http://www.w3.org/2006/10/ttaf1',
- 'ttaf1_0604': 'http://www.w3.org/2006/04/ttaf1',
+ 'tts': 'http://www.w3.org/ns/ttml#styling',
})
+ styles = {}
+ default_style = {}
+
class TTMLPElementParser(object):
- out = ''
+ _out = ''
+ _unclosed_elements = []
+ _applied_styles = []
def start(self, tag, attrib):
- if tag in (_x('ttml:br'), _x('ttaf1:br'), 'br'):
- self.out += '\n'
+ if tag in (_x('ttml:br'), 'br'):
+ self._out += '\n'
+ else:
+ unclosed_elements = []
+ style = {}
+ element_style_id = attrib.get('style')
+ if default_style:
+ style.update(default_style)
+ if element_style_id:
+ style.update(styles.get(element_style_id, {}))
+ for prop in SUPPORTED_STYLING:
+ prop_val = attrib.get(_x('tts:' + prop))
+ if prop_val:
+ style[prop] = prop_val
+ if style:
+ font = ''
+ for k, v in sorted(style.items()):
+ if self._applied_styles and self._applied_styles[-1].get(k) == v:
+ continue
+ if k == 'color':
+ font += ' color="%s"' % v
+ elif k == 'fontSize':
+ font += ' size="%s"' % v
+ elif k == 'fontFamily':
+ font += ' face="%s"' % v
+ elif k == 'fontWeight' and v == 'bold':
+ self._out += '<b>'
+ unclosed_elements.append('b')
+ elif k == 'fontStyle' and v == 'italic':
+ self._out += '<i>'
+ unclosed_elements.append('i')
+ elif k == 'textDecoration' and v == 'underline':
+ self._out += '<u>'
+ unclosed_elements.append('u')
+ if font:
+ self._out += '<font' + font + '>'
+ unclosed_elements.append('font')
+ applied_style = {}
+ if self._applied_styles:
+ applied_style.update(self._applied_styles[-1])
+ applied_style.update(style)
+ self._applied_styles.append(applied_style)
+ self._unclosed_elements.append(unclosed_elements)
def end(self, tag):
- pass
+ if tag not in (_x('ttml:br'), 'br'):
+ unclosed_elements = self._unclosed_elements.pop()
+ for element in reversed(unclosed_elements):
+ self._out += '</%s>' % element
+ if unclosed_elements and self._applied_styles:
+ self._applied_styles.pop()
def data(self, data):
- self.out += data
+ self._out += data
def close(self):
- return self.out.strip()
+ return self._out.strip()
def parse_node(node):
target = TTMLPElementParser()
parser.feed(xml.etree.ElementTree.tostring(node))
return parser.close()
- dfxp = compat_etree_fromstring(dfxp_data.encode('utf-8'))
+ for k, v in LEGACY_NAMESPACES:
+ for ns in v:
+ dfxp_data = dfxp_data.replace(ns, k)
+
+ dfxp = compat_etree_fromstring(dfxp_data)
out = []
- paras = dfxp.findall(_x('.//ttml:p')) or dfxp.findall(_x('.//ttaf1:p')) or dfxp.findall(_x('.//ttaf1_0604:p')) or dfxp.findall('.//p')
+ paras = dfxp.findall(_x('.//ttml:p')) or dfxp.findall('.//p')
if not paras:
raise ValueError('Invalid dfxp/TTML subtitle')
+ repeat = False
+ while True:
+ for style in dfxp.findall(_x('.//ttml:style')):
+ style_id = style.get('id')
+ parent_style_id = style.get('style')
+ if parent_style_id:
+ if parent_style_id not in styles:
+ repeat = True
+ continue
+ styles[style_id] = styles[parent_style_id].copy()
+ for prop in SUPPORTED_STYLING:
+ prop_val = style.get(_x('tts:' + prop))
+ if prop_val:
+ styles.setdefault(style_id, {})[prop] = prop_val
+ if repeat:
+ repeat = False
+ else:
+ break
+
+ for p in ('body', 'div'):
+ ele = xpath_element(dfxp, [_x('.//ttml:' + p), './/' + p])
+ if ele is None:
+ continue
+ style = styles.get(ele.get('style'))
+ if not style:
+ continue
+ default_style.update(style)
+
for para, index in zip(paras, itertools.count(1)):
begin_time = parse_dfxp_time_expr(para.attrib.get('begin'))
end_time = parse_dfxp_time_expr(para.attrib.get('end'))
def cli_bool_option(params, command_option, param, true_value='true', false_value='false', separator=None):
param = params.get(param)
+ if param is None:
+ return []
assert isinstance(param, bool)
if separator:
return [command_option + separator + (true_value if param else false_value)]
"Couldn't find a tool to set the xattrs. "
"Install either the python 'xattr' module, "
"or the 'xattr' binary.")
+
+
+def random_birthday(year_field, month_field, day_field):
+ return {
+ year_field: str(random.randint(1950, 1995)),
+ month_field: str(random.randint(1, 12)),
+ day_field: str(random.randint(1, 31)),
+ }