X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;f=youtube_dl%2Futils.py;h=6c84bfe0ffc5667d53a88fbae3ce76e76e36d909;hb=5552c9eb0fece567f7dda13810939fca32d7d65a;hp=4e76b6b7be709e4d6274f65a8ef8175b436cddc2;hpb=773f291dcbce486fefe24e1abd29735d374d0a9e;p=youtube-dl
diff --git a/youtube_dl/utils.py b/youtube_dl/utils.py
index 4e76b6b7b..6c84bfe0f 100644
--- a/youtube_dl/utils.py
+++ b/youtube_dl/utils.py
@@ -11,6 +11,7 @@ import contextlib
import ctypes
import datetime
import email.utils
+import email.header
import errno
import functools
import gzip
@@ -39,6 +40,7 @@ from .compat import (
compat_basestring,
compat_chr,
compat_etree_fromstring,
+ compat_expanduser,
compat_html_entities,
compat_html_entities_html5,
compat_http_client,
@@ -420,8 +422,8 @@ def clean_html(html):
# Newline vs
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
@@ -473,7 +475,8 @@ def timeconvert(timestr):
def sanitize_filename(s, restricted=False, is_id=False):
"""Sanitizes a string so it could be used as part of a filename.
If restricted is set, use a stricter subset of allowed characters.
- Set is_id if this is not an arbitrary string, but an ID that should be kept if possible
+ Set is_id if this is not an arbitrary string, but an ID that should be kept
+ if possible.
"""
def replace_insane(char):
if restricted and char in ACCENT_CHARS:
@@ -538,6 +541,11 @@ 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 = []
@@ -1187,6 +1195,11 @@ def unified_timestamp(date_str, day_first=True):
# 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\s*[A-Z]+)$', date_str)
+ if m:
+ date_str = date_str[:-len(m.group('tz'))]
+
for expression in date_formats(day_first):
try:
dt = datetime.datetime.strptime(date_str, expression) - timezone + datetime.timedelta(hours=pm_delta)
@@ -1747,11 +1760,16 @@ def base_url(url):
def urljoin(base, path):
+ if isinstance(path, bytes):
+ path = path.decode('utf-8')
if not isinstance(path, compat_str) or not path:
return None
if re.match(r'^(?:https?:)?//', path):
return path
- if not isinstance(base, compat_str) or not re.match(r'^(?:https?:)?//', base):
+ if isinstance(base, bytes):
+ base = base.decode('utf-8')
+ if not isinstance(base, compat_str) or not re.match(
+ r'^(?:https?:)?//', base):
return None
return compat_urlparse.urljoin(base, path)
@@ -2080,6 +2098,58 @@ def update_Request(req, url=None, data=None, headers={}, query={}):
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:
@@ -2091,13 +2161,16 @@ def dict_get(d, key_or_keys, default=None, skip_false_values=True):
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'):
@@ -2138,7 +2211,12 @@ def parse_age_limit(s):
def strip_jsonp(code):
return re.sub(
- r'(?s)^[a-zA-Z0-9_.$]+\s*\(\s*(.*)\);?\s*?(?://[^\n]*)*$', r'\1', code)
+ r'''(?sx)^
+ (?:window\.)?(?P[a-zA-Z0-9_.$]+)
+ (?:\s*&&\s*(?P=func_name))?
+ \s*\(\s*(?P.*)\);?
+ \s*?(?://[^\n]*)*$''',
+ r'\g', code)
def js_to_json(code):
@@ -2258,10 +2336,8 @@ def mimetype2ext(mt):
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-wmv': 'wmv',
@@ -2269,11 +2345,11 @@ def mimetype2ext(mt):
'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)
@@ -2289,11 +2365,11 @@ def parse_codecs(codecs_str):
if codec in ('avc1', 'avc2', 'avc3', 'avc4', 'vp9', 'vp8', 'hev1', 'hev2', 'h263', 'h264', 'mp4v'):
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 {
@@ -2496,27 +2572,97 @@ def srt_subtitles_timecode(seconds):
def dfxp2srt(dfxp_data):
+ LEGACY_NAMESPACES = (
+ ('http://www.w3.org/ns/ttml', [
+ 'http://www.w3.org/2004/11/ttaf1',
+ 'http://www.w3.org/2006/04/ttaf1',
+ 'http://www.w3.org/2006/10/ttaf1',
+ ]),
+ ('http://www.w3.org/ns/ttml#styling', [
+ '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 += ''
+ unclosed_elements.append('b')
+ elif k == 'fontStyle' and v == 'italic':
+ self._out += ''
+ unclosed_elements.append('i')
+ elif k == 'textDecoration' and v == 'underline':
+ self._out += ''
+ unclosed_elements.append('u')
+ if font:
+ self._out += ''
+ 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 += '' % 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()
@@ -2524,13 +2670,45 @@ def dfxp2srt(dfxp_data):
parser.feed(xml.etree.ElementTree.tostring(node))
return parser.close()
+ for k, v in LEGACY_NAMESPACES:
+ for ns in v:
+ dfxp_data = dfxp_data.replace(ns, k)
+
dfxp = compat_etree_fromstring(dfxp_data.encode('utf-8'))
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'))
@@ -3290,8 +3468,8 @@ class GeoUtils(object):
addr, preflen = block.split('/')
addr_min = compat_struct_unpack('!L', socket.inet_aton(addr))[0]
addr_max = addr_min | (0xffffffff >> int(preflen))
- return socket.inet_ntoa(
- compat_struct_pack('!I', random.randint(addr_min, addr_max)))
+ return compat_str(socket.inet_ntoa(
+ compat_struct_pack('!L', random.randint(addr_min, addr_max))))
class PerRequestProxyHandler(compat_urllib_request.ProxyHandler):
@@ -3319,6 +3497,57 @@ class PerRequestProxyHandler(compat_urllib_request.ProxyHandler):
self, req, proxy, type)
+# Both long_to_bytes and bytes_to_long are adapted from PyCrypto, which is
+# released into Public Domain
+# https://github.com/dlitz/pycrypto/blob/master/lib/Crypto/Util/number.py#L387
+
+def long_to_bytes(n, blocksize=0):
+ """long_to_bytes(n:long, blocksize:int) : string
+ Convert a long integer to a byte string.
+
+ If optional blocksize is given and greater than zero, pad the front of the
+ byte string with binary zeros so that the length is a multiple of
+ blocksize.
+ """
+ # after much testing, this algorithm was deemed to be the fastest
+ s = b''
+ n = int(n)
+ while n > 0:
+ s = compat_struct_pack('>I', n & 0xffffffff) + s
+ n = n >> 32
+ # strip off leading zeros
+ for i in range(len(s)):
+ if s[i] != b'\000'[0]:
+ break
+ else:
+ # only happens when n == 0
+ s = b'\000'
+ i = 0
+ s = s[i:]
+ # add back some pad bytes. this could be done more efficiently w.r.t. the
+ # de-padding being done above, but sigh...
+ if blocksize > 0 and len(s) % blocksize:
+ s = (blocksize - len(s) % blocksize) * b'\000' + s
+ return s
+
+
+def bytes_to_long(s):
+ """bytes_to_long(string) : long
+ Convert a byte string to a long integer.
+
+ This is (essentially) the inverse of long_to_bytes().
+ """
+ acc = 0
+ length = len(s)
+ if length % 4:
+ extra = (4 - length % 4)
+ s = b'\000' * extra + s
+ length = length + extra
+ for i in range(0, length, 4):
+ acc = (acc << 32) + compat_struct_unpack('>I', s[i:i + 4])[0]
+ return acc
+
+
def ohdave_rsa_encrypt(data, exponent, modulus):
'''
Implement OHDave's RSA algorithm. See http://www.ohdave.com/rsa/
@@ -3336,6 +3565,21 @@ def ohdave_rsa_encrypt(data, exponent, modulus):
return '%x' % encrypted
+def pkcs1pad(data, length):
+ """
+ Padding input data with PKCS#1 scheme
+
+ @param {int[]} data input data
+ @param {int} length target length
+ @returns {int[]} padded data
+ """
+ if len(data) > length - 11:
+ raise ValueError('Input data too long for PKCS#1 padding')
+
+ pseudo_random = [random.randint(0, 254) for _ in range(length - len(data) - 3)]
+ return [0, 2] + pseudo_random + [0] + data
+
+
def encode_base_n(num, n, table=None):
FULL_TABLE = '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ'
if not table:
@@ -3574,3 +3818,11 @@ def write_xattr(path, key, 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)),
+ }