X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;f=youtube_dl%2Futils.py;h=9c1c0e0bdf55040dc3b82d4393d50c3623e3b0de;hb=dfb1b1468cef4ddc7ecc43776abce03763f8e426;hp=c761ea22a4580216bb2484bd08ec7e8de1d724e4;hpb=f78546272cf7c4b10c8003870728ab69bec982fc;p=youtube-dl diff --git a/youtube_dl/utils.py b/youtube_dl/utils.py index c761ea22a..9c1c0e0bd 100644 --- a/youtube_dl/utils.py +++ b/youtube_dl/utils.py @@ -70,6 +70,21 @@ ENGLISH_MONTH_NAMES = [ 'January', 'February', 'March', 'April', 'May', 'June', 'July', 'August', 'September', 'October', 'November', 'December'] +KNOWN_EXTENSIONS = ( + 'mp4', 'm4a', 'm4p', 'm4b', 'm4r', 'm4v', 'aac', + 'flv', 'f4v', 'f4a', 'f4b', + 'webm', 'ogg', 'ogv', 'oga', 'ogx', 'spx', 'opus', + 'mkv', 'mka', 'mk3d', + 'avi', 'divx', + 'mov', + 'asf', 'wmv', 'wma', + '3gp', '3g2', + 'mp3', + 'flac', + 'ape', + 'wav', + 'f4f', 'f4m', 'm3u8', 'smil') + def preferredencoding(): """Get preferred encoding. @@ -179,10 +194,19 @@ def xpath_with_ns(path, ns_map): def xpath_element(node, xpath, name=None, fatal=False, default=NO_DEFAULT): - if sys.version_info < (2, 7): # Crazy 2.6 - xpath = xpath.encode('ascii') + def _find_xpath(xpath): + if sys.version_info < (2, 7): # Crazy 2.6 + xpath = xpath.encode('ascii') + return node.find(xpath) + + if isinstance(xpath, (str, compat_str)): + n = _find_xpath(xpath) + else: + for xp in xpath: + n = _find_xpath(xp) + if n is not None: + break - n = node.find(xpath) if n is None: if default is not NO_DEFAULT: return default @@ -357,13 +381,20 @@ def sanitize_path(s): if drive_or_unc: norm_path.pop(0) sanitized_path = [ - path_part if path_part in ['.', '..'] else re.sub('(?:[/<>:"\\|\\\\?\\*]|\.$)', '#', path_part) + path_part if path_part in ['.', '..'] else re.sub('(?:[/<>:"\\|\\\\?\\*]|[\s.]$)', '#', path_part) for path_part in norm_path] if drive_or_unc: sanitized_path.insert(0, drive_or_unc + os.path.sep) 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 sanitized_Request(url, *args, **kwargs): + return compat_urllib_request.Request( + 'http:%s' % url if url.startswith('//') else url, *args, **kwargs) + + def orderedSet(iterable): """ Remove all duplicates from the input iterable """ res = [] @@ -387,10 +418,14 @@ def _htmlentity_transform(entity): numstr = '0%s' % numstr else: base = 10 - return compat_chr(int(numstr, base)) + # See https://github.com/rg3/youtube-dl/issues/7518 + try: + return compat_chr(int(numstr, base)) + except ValueError: + pass # Unknown entity in name, return its literal representation - return ('&%s;' % entity) + return '&%s;' % entity def unescapeHTML(s): @@ -643,6 +678,16 @@ def _create_http_connection(ydl_handler, http_class, is_https, *args, **kwargs): return hc +def handle_youtubedl_headers(headers): + filtered_headers = headers + + if 'Youtubedl-no-compression' in filtered_headers: + filtered_headers = dict((k, v) for k, v in filtered_headers.items() if k.lower() != 'accept-encoding') + del filtered_headers['Youtubedl-no-compression'] + + return filtered_headers + + class YoutubeDLHandler(compat_urllib_request.HTTPHandler): """Handler for HTTP requests and responses. @@ -650,7 +695,7 @@ class YoutubeDLHandler(compat_urllib_request.HTTPHandler): the standard headers to every HTTP request and handles gzipped and deflated responses from web servers. If compression is to be avoided in a particular request, the original request in the program code only has - to include the HTTP header "Youtubedl-No-Compression", which will be + to include the HTTP header "Youtubedl-no-compression", which will be removed before making the real request. Part of this code was copied from: @@ -711,10 +756,8 @@ class YoutubeDLHandler(compat_urllib_request.HTTPHandler): # The dict keys are capitalized because of this bug by urllib if h.capitalize() not in req.headers: req.add_header(h, v) - if 'Youtubedl-no-compression' in req.headers: - if 'Accept-encoding' in req.headers: - del req.headers['Accept-encoding'] - del req.headers['Youtubedl-no-compression'] + + req.headers = handle_youtubedl_headers(req.headers) if sys.version_info < (2, 7) and '#' in req.get_full_url(): # Python 2.6 is brain-dead when it comes to fragments @@ -745,11 +788,13 @@ class YoutubeDLHandler(compat_urllib_request.HTTPHandler): raise original_ioerror resp = self.addinfourl_wrapper(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.msg = old_resp.msg + del resp.headers['Content-encoding'] # Percent-encode redirect URL of Location HTTP header to satisfy RFC 3986 (see # https://github.com/rg3/youtube-dl/issues/6457). if 300 <= resp.code < 400: @@ -815,9 +860,11 @@ def parse_iso8601(date_str, delimiter='T', timezone=None): if date_str is None: return None + date_str = re.sub(r'\.[0-9]+', '', date_str) + if timezone is None: m = re.search( - r'(\.[0-9]+)?(?:Z$| ?(?P\+|-)(?P[0-9]{2}):?(?P[0-9]{2})$)', + r'(?:Z$| ?(?P\+|-)(?P[0-9]{2}):?(?P[0-9]{2})$)', date_str) if not m: timezone = datetime.timedelta() @@ -830,9 +877,12 @@ def parse_iso8601(date_str, delimiter='T', timezone=None): timezone = datetime.timedelta( hours=sign * int(m.group('hours')), minutes=sign * int(m.group('minutes'))) - date_format = '%Y-%m-%d{0}%H:%M:%S'.format(delimiter) - dt = datetime.datetime.strptime(date_str, date_format) - timezone - return calendar.timegm(dt.timetuple()) + try: + date_format = '%Y-%m-%d{0}%H:%M:%S'.format(delimiter) + dt = datetime.datetime.strptime(date_str, date_format) - timezone + return calendar.timegm(dt.timetuple()) + except ValueError: + pass def unified_strdate(date_str, day_first=True): @@ -897,7 +947,8 @@ def unified_strdate(date_str, day_first=True): timetuple = email.utils.parsedate_tz(date_str) if timetuple: upload_date = datetime.datetime(*timetuple[:6]).strftime('%Y%m%d') - return upload_date + if upload_date is not None: + return compat_str(upload_date) def determine_ext(url, default_ext='unknown_video'): @@ -906,6 +957,9 @@ def determine_ext(url, default_ext='unknown_video'): guess = url.partition('?')[0].rpartition('.')[2] if re.match(r'^[A-Za-z0-9]+$', guess): return guess + # Try extract ext from URLs like http://example.com/foo/bar.mp4/?download + elif guess.rstrip('/') in KNOWN_EXTENSIONS: + return guess.rstrip('/') else: return default_ext @@ -930,7 +984,7 @@ def date_from_str(date_str): if sign == '-': time = -time unit = match.group('unit') - # A bad aproximation? + # A bad approximation? if unit == 'month': unit = 'day' time *= 30 @@ -1253,7 +1307,7 @@ def parse_filesize(s): if s is None: return None - # The lower-case forms are of course incorrect and inofficial, + # The lower-case forms are of course incorrect and unofficial, # but we support those too _UNIT_TABLE = { 'B': 1, @@ -1357,6 +1411,15 @@ def remove_end(s, end): return s +def remove_quotes(s): + if s is None or len(s) < 2: + return s + for quote in ('"', "'", ): + if s[0] == quote and s[-1] == quote: + return s[1:-1] + return s + + def url_basename(url): path = compat_urlparse.urlparse(url).path return path.strip('/').split('/')[-1] @@ -1649,7 +1712,13 @@ def urlencode_postdata(*args, **kargs): def encode_dict(d, encoding='utf-8'): - return dict((k.encode(encoding), v.encode(encoding)) for k, v in d.items()) + def encode(v): + return v.encode(encoding) if isinstance(v, compat_basestring) else v + return dict((encode(k), encode(v)) for k, v in d.items()) + + +def encode_compat_str(string, encoding=preferredencoding(), errors='strict'): + return string if isinstance(string, compat_str) else compat_str(string, encoding, errors) US_RATINGS = { @@ -1746,6 +1815,15 @@ def args_to_str(args): return ' '.join(shlex_quote(a) for a in args) +def error_to_compat_str(err): + err_str = str(err) + # On python 2 error byte string must be decoded with proper + # encoding rather than ascii + if sys.version_info[0] < 3: + err_str = err_str.decode(preferredencoding()) + return err_str + + def mimetype2ext(mt): _, _, res = mt.rpartition('/') @@ -1916,15 +1994,15 @@ def match_filter_func(filter_str): def parse_dfxp_time_expr(time_expr): if not time_expr: - return 0.0 + return mobj = re.match(r'^(?P\d+(?:\.\d+)?)s?$', time_expr) if mobj: return float(mobj.group('time_offset')) - mobj = re.match(r'^(\d+):(\d\d):(\d\d(?:\.\d+)?)$', time_expr) + mobj = re.match(r'^(\d+):(\d\d):(\d\d(?:(?:\.|:)\d+)?)$', time_expr) if mobj: - return 3600 * int(mobj.group(1)) + 60 * int(mobj.group(2)) + float(mobj.group(3)) + return 3600 * int(mobj.group(1)) + 60 * int(mobj.group(2)) + float(mobj.group(3).replace(':', '.')) def srt_subtitles_timecode(seconds): @@ -1960,10 +2038,15 @@ def dfxp2srt(dfxp_data): raise ValueError('Invalid dfxp/TTML subtitle') for para, index in zip(paras, itertools.count(1)): - begin_time = parse_dfxp_time_expr(para.attrib['begin']) + begin_time = parse_dfxp_time_expr(para.attrib.get('begin')) end_time = parse_dfxp_time_expr(para.attrib.get('end')) + dur = parse_dfxp_time_expr(para.attrib.get('dur')) + if begin_time is None: + continue if not end_time: - end_time = begin_time + parse_dfxp_time_expr(para.attrib['dur']) + if not dur: + continue + end_time = begin_time + dur out.append('%d\n%s --> %s\n%s\n\n' % ( index, srt_subtitles_timecode(begin_time),