X-Git-Url: http://git.bitcoin.ninja/index.cgi?p=youtube-dl;a=blobdiff_plain;f=youtube_dl%2Futils.py;h=2fe9cf585db817e1d86831c71ef28502b4a16ee5;hp=cbf7639c56393f21bd259ee30e294d1db704fb6b;hb=c384d537f882efab10a78a56ce6dcb0a30f54b47;hpb=18a0defab063523cd76a30be2dd5a80e9f9172d6 diff --git a/youtube_dl/utils.py b/youtube_dl/utils.py index cbf7639c5..2fe9cf585 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 @@ -21,7 +22,6 @@ import locale import math import operator import os -import pipes import platform import random import re @@ -35,10 +35,13 @@ import xml.etree.ElementTree 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, @@ -157,6 +160,8 @@ DATE_FORMATS = ( '%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) @@ -363,9 +368,9 @@ def get_elements_by_attribute(attribute, value, html, escape_value=True): 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.*?) @@ -407,8 +412,12 @@ def extract_attributes(html_element): 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 @@ -420,8 +429,8 @@ 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 +482,8 @@ 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 +548,11 @@ 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 = [] @@ -584,7 +599,7 @@ def unescapeHTML(s): 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(): @@ -924,14 +939,6 @@ class YoutubeDLHandler(compat_urllib_request.HTTPHandler): 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 @@ -983,13 +990,13 @@ class YoutubeDLHandler(compat_urllib_request.HTTPHandler): 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 @@ -1179,7 +1186,7 @@ def unified_timestamp(date_str, day_first=True): 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) @@ -1187,6 +1194,11 @@ 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) @@ -1319,24 +1331,24 @@ def _windows_write_string(s, out): 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): @@ -1525,7 +1537,7 @@ def shell_quote(args): 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) @@ -1747,11 +1759,16 @@ 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) @@ -1801,6 +1818,10 @@ def float_or_none(v, scale=1, invscale=1, default=None): 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() @@ -1817,10 +1838,20 @@ def parse_duration(s): 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[0-9]+)\s*d(?:ays?)?\s* )? + T)? (?: (?P[0-9]+)\s*h(?:ours?)?\s* )? @@ -1915,7 +1946,7 @@ class PagedList(object): 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 @@ -2080,6 +2111,58 @@ 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 +2174,16 @@ 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 +2224,12 @@ 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): @@ -2176,7 +2267,7 @@ def js_to_json(code): "(?:[^"\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^"\\]*"| '(?:[^'\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^'\\]*'| {comment}|,(?={skip}[\]}}])| - [a-zA-Z_][.a-zA-Z_0-9]*| + (?:(?' + 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 +2688,45 @@ def dfxp2srt(dfxp_data): 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')) @@ -2559,6 +2755,8 @@ def cli_option(params, command_option, param): 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)] @@ -3291,7 +3489,7 @@ class GeoUtils(object): addr_min = compat_struct_unpack('!L', socket.inet_aton(addr))[0] addr_max = addr_min | (0xffffffff >> int(preflen)) return compat_str(socket.inet_ntoa( - compat_struct_pack('!I', random.randint(addr_min, addr_max)))) + compat_struct_pack('!L', random.randint(addr_min, addr_max)))) class PerRequestProxyHandler(compat_urllib_request.ProxyHandler): @@ -3319,6 +3517,57 @@ 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 +3585,21 @@ 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 +3838,11 @@ 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)), + }