X-Git-Url: https://git.rapsys.eu/youtubedl/blobdiff_plain/9fb40a4e5f59ba243dee2edad23a2c5555338113..faab51d51a7cb39acf832a7bf8ced3463a6b965a:/youtube_dl/utils.py?ds=sidebyside diff --git a/youtube_dl/utils.py b/youtube_dl/utils.py index 17b8379..2843a3d 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 +35,12 @@ import xml.etree.ElementTree import zlib from .compat import ( + compat_HTMLParseError, compat_HTMLParser, compat_basestring, compat_chr, compat_etree_fromstring, + compat_expanduser, compat_html_entities, compat_html_entities_html5, compat_http_client, @@ -157,6 +159,8 @@ 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 +367,9 @@ 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 +411,12 @@ 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 +428,8 @@ 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 +481,8 @@ 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 +547,11 @@ 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 +598,7 @@ 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 +938,6 @@ 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 +989,13 @@ 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 +1185,7 @@ 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 +1193,11 @@ 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) @@ -1525,7 +1536,7 @@ 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 +1758,16 @@ 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 +1817,10 @@ 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 +1837,20 @@ 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 +1945,7 @@ 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 +2110,58 @@ 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 +2173,16 @@ 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 +2223,12 @@ 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): @@ -2258,22 +2348,21 @@ 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-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) @@ -2286,14 +2375,14 @@ def parse_codecs(codecs_str): 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 { @@ -2496,27 +2585,101 @@ def srt_subtitles_timecode(seconds): 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 += '' + 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 +2687,45 @@ 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 +2754,8 @@ 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)] @@ -3319,6 +3516,57 @@ 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 +3584,21 @@ 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 +3837,11 @@ 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)), + }