X-Git-Url: https://git.rapsys.eu/youtubedl/blobdiff_plain/e8cd8c4bd832446f1971215b9fedc4531555dc1a..a8b640ed81fd63ade1773f4aa2f7808bc80175dd:/youtube_dl/utils.py diff --git a/youtube_dl/utils.py b/youtube_dl/utils.py index 73fe1ad..2864e51 100644 --- a/youtube_dl/utils.py +++ b/youtube_dl/utils.py @@ -1,11 +1,16 @@ #!/usr/bin/env python # -*- coding: utf-8 -*- +import calendar +import codecs +import contextlib import ctypes import datetime import email.utils import errno +import getpass import gzip +import itertools import io import json import locale @@ -16,9 +21,12 @@ import platform import re import ssl import socket +import struct import subprocess import sys +import tempfile import traceback +import xml.etree.ElementTree import zlib try: @@ -84,11 +92,9 @@ except ImportError: compat_subprocess_get_DEVNULL = lambda: open(os.path.devnull, 'w') try: - from urllib.parse import parse_qs as compat_parse_qs -except ImportError: # Python 2 - # HACK: The following is the correct parse_qs implementation from cpython 3's stdlib. - # Python 2's version is apparently totally broken - def _unquote(string, encoding='utf-8', errors='replace'): + from urllib.parse import unquote as compat_urllib_parse_unquote +except ImportError: + def compat_urllib_parse_unquote(string, encoding='utf-8', errors='replace'): if string == '': return string res = string.split('%') @@ -123,6 +129,13 @@ except ImportError: # Python 2 string += pct_sequence.decode(encoding, errors) return string + +try: + from urllib.parse import parse_qs as compat_parse_qs +except ImportError: # Python 2 + # HACK: The following is the correct parse_qs implementation from cpython 3's stdlib. + # Python 2's version is apparently totally broken + def _parse_qsl(qs, keep_blank_values=False, strict_parsing=False, encoding='utf-8', errors='replace'): qs, _coerce_result = qs, unicode @@ -142,10 +155,12 @@ except ImportError: # Python 2 continue if len(nv[1]) or keep_blank_values: name = nv[0].replace('+', ' ') - name = _unquote(name, encoding=encoding, errors=errors) + name = compat_urllib_parse_unquote( + name, encoding=encoding, errors=errors) name = _coerce_result(name) value = nv[1].replace('+', ' ') - value = _unquote(value, encoding=encoding, errors=errors) + value = compat_urllib_parse_unquote( + value, encoding=encoding, errors=errors) value = _coerce_result(value) r.append((name, value)) return r @@ -172,10 +187,98 @@ try: except NameError: compat_chr = chr +try: + from xml.etree.ElementTree import ParseError as compat_xml_parse_error +except ImportError: # Python 2.6 + from xml.parsers.expat import ExpatError as compat_xml_parse_error + +try: + from shlex import quote as shlex_quote +except ImportError: # Python < 3.3 + def shlex_quote(s): + return "'" + s.replace("'", "'\"'\"'") + "'" + + def compat_ord(c): if type(c) is int: return c else: return ord(c) + +if sys.version_info >= (3, 0): + compat_getenv = os.getenv + compat_expanduser = os.path.expanduser +else: + # Environment variables should be decoded with filesystem encoding. + # Otherwise it will fail if any non-ASCII characters present (see #3854 #3217 #2918) + + def compat_getenv(key, default=None): + env = os.getenv(key, default) + if env: + env = env.decode(get_filesystem_encoding()) + return env + + # HACK: The default implementations of os.path.expanduser from cpython do not decode + # environment variables with filesystem encoding. We will work around this by + # providing adjusted implementations. + # The following are os.path.expanduser implementations from cpython 2.7.8 stdlib + # for different platforms with correct environment variables decoding. + + if os.name == 'posix': + def compat_expanduser(path): + """Expand ~ and ~user constructions. If user or $HOME is unknown, + do nothing.""" + if not path.startswith('~'): + return path + i = path.find('/', 1) + if i < 0: + i = len(path) + if i == 1: + if 'HOME' not in os.environ: + import pwd + userhome = pwd.getpwuid(os.getuid()).pw_dir + else: + userhome = compat_getenv('HOME') + else: + import pwd + try: + pwent = pwd.getpwnam(path[1:i]) + except KeyError: + return path + userhome = pwent.pw_dir + userhome = userhome.rstrip('/') + return (userhome + path[i:]) or '/' + elif os.name == 'nt' or os.name == 'ce': + def compat_expanduser(path): + """Expand ~ and ~user constructs. + + If user or $HOME is unknown, do nothing.""" + if path[:1] != '~': + return path + i, n = 1, len(path) + while i < n and path[i] not in '/\\': + i = i + 1 + + if 'HOME' in os.environ: + userhome = compat_getenv('HOME') + elif 'USERPROFILE' in os.environ: + userhome = compat_getenv('USERPROFILE') + elif not 'HOMEPATH' in os.environ: + return path + else: + try: + drive = compat_getenv('HOMEDRIVE') + except KeyError: + drive = '' + userhome = os.path.join(drive, compat_getenv('HOMEPATH')) + + if i != 1: #~user + userhome = os.path.join(os.path.dirname(userhome), path[1:i]) + + return userhome + path[i:] + else: + compat_expanduser = os.path.expanduser + + # This is not clearly defined otherwise compiled_regex_type = type(re.compile('')) @@ -209,26 +312,55 @@ else: assert type(s) == type(u'') print(s) -# In Python 2.x, json.dump expects a bytestream. -# In Python 3.x, it writes to a character stream -if sys.version_info < (3,0): - def write_json_file(obj, fn): - with open(fn, 'wb') as f: - json.dump(obj, f) -else: - def write_json_file(obj, fn): - with open(fn, 'w', encoding='utf-8') as f: - json.dump(obj, f) -if sys.version_info >= (2,7): +def write_json_file(obj, fn): + """ Encode obj as JSON and write it to fn, atomically """ + + args = { + 'suffix': '.tmp', + 'prefix': os.path.basename(fn) + '.', + 'dir': os.path.dirname(fn), + 'delete': False, + } + + # In Python 2.x, json.dump expects a bytestream. + # In Python 3.x, it writes to a character stream + if sys.version_info < (3, 0): + args['mode'] = 'wb' + else: + args.update({ + 'mode': 'w', + 'encoding': 'utf-8', + }) + + tf = tempfile.NamedTemporaryFile(**args) + + try: + with tf: + json.dump(obj, tf) + os.rename(tf.name, fn) + except: + try: + os.remove(tf.name) + except OSError: + pass + raise + + +if sys.version_info >= (2, 7): def find_xpath_attr(node, xpath, key, val): """ Find the xpath xpath[@key=val] """ - assert re.match(r'^[a-zA-Z]+$', key) - assert re.match(r'^[a-zA-Z0-9@\s]*$', val) + assert re.match(r'^[a-zA-Z-]+$', key) + assert re.match(r'^[a-zA-Z0-9@\s:._-]*$', val) expr = xpath + u"[@%s='%s']" % (key, val) return node.find(expr) else: def find_xpath_attr(node, xpath, key, val): + # Here comes the crazy part: In 2.6, if the xpath is a unicode, + # .//node does not match if a node is a direct child of . ! + if isinstance(xpath, unicode): + xpath = xpath.encode('ascii') + for f in node.findall(xpath): if f.attrib.get(key) == val: return f @@ -247,30 +379,20 @@ def xpath_with_ns(path, ns_map): replaced.append('{%s}%s' % (ns_map[ns], tag)) return '/'.join(replaced) -def htmlentity_transform(matchobj): - """Transforms an HTML entity to a character. - - This function receives a match object and is intended to be used with - the re.sub() function. - """ - entity = matchobj.group(1) - # Known non-numeric HTML entity - if entity in compat_html_entities.name2codepoint: - return compat_chr(compat_html_entities.name2codepoint[entity]) +def xpath_text(node, xpath, name=None, fatal=False): + if sys.version_info < (2, 7): # Crazy 2.6 + xpath = xpath.encode('ascii') - mobj = re.match(u'(?u)#(x?\\d+)', entity) - if mobj is not None: - numstr = mobj.group(1) - if numstr.startswith(u'x'): - base = 16 - numstr = u'0%s' % numstr + n = node.find(xpath) + if n is None: + if fatal: + name = xpath if name is None else name + raise ExtractorError('Could not find XML element %s' % name) else: - base = 10 - return compat_chr(int(numstr, base)) + return None + return n.text - # Unknown entity in name, return its literal representation - return (u'&%s;' % entity) compat_html_parser.locatestarttagend = re.compile(r"""<[a-zA-Z][-.a-zA-Z0-9:_]*(?:\s+(?:(?<=['"\s])[^\s/>][^\s/=>]*(?:\s*=+\s*(?:'[^']*'|"[^"]*"|(?!['"])[^>\s]*))?\s*)*)?\s*""", re.VERBOSE) # backport bugfix class BaseHTMLParser(compat_html_parser.HTMLParser): @@ -491,14 +613,34 @@ def orderedSet(iterable): res.append(el) return res + +def _htmlentity_transform(entity): + """Transforms an HTML entity to a character.""" + # Known non-numeric HTML entity + if entity in compat_html_entities.name2codepoint: + return compat_chr(compat_html_entities.name2codepoint[entity]) + + mobj = re.match(r'#(x?[0-9]+)', entity) + if mobj is not None: + numstr = mobj.group(1) + if numstr.startswith(u'x'): + base = 16 + numstr = u'0%s' % numstr + else: + base = 10 + return compat_chr(int(numstr, base)) + + # Unknown entity in name, return its literal representation + return (u'&%s;' % entity) + + def unescapeHTML(s): - """ - @param s a string - """ - assert type(s) == type(u'') + if s is None: + return None + assert type(s) == compat_str - result = re.sub(u'(?u)&(.+?);', htmlentity_transform, s) - return result + return re.sub( + r'&([^;]+);', lambda m: _htmlentity_transform(m.group(1)), s) def encodeFilename(s, for_subprocess=False): @@ -529,6 +671,15 @@ def encodeFilename(s, for_subprocess=False): return s.encode(encoding, 'ignore') +def encodeArgument(s): + if not isinstance(s, compat_str): + # Legacy code that uses byte strings + # Uncomment the following line after fixing all post processors + #assert False, 'Internal error: %r should be of type %r, is %r' % (s, compat_str, type(s)) + s = s.decode('ascii') + return encodeFilename(s, True) + + def decodeOption(optval): if optval is None: return optval @@ -561,7 +712,7 @@ def make_HTTPS_handler(opts_no_check_certificate, **kwargs): self.sock = sock self._tunnel() try: - self.sock = ssl.wrap_socket(sock, self.key_file, self.cert_file, ssl_version=ssl.PROTOCOL_SSLv3) + self.sock = ssl.wrap_socket(sock, self.key_file, self.cert_file, ssl_version=ssl.PROTOCOL_TLSv1) except ssl.SSLError: self.sock = ssl.wrap_socket(sock, self.key_file, self.cert_file, ssl_version=ssl.PROTOCOL_SSLv23) @@ -569,8 +720,14 @@ def make_HTTPS_handler(opts_no_check_certificate, **kwargs): def https_open(self, req): return self.do_open(HTTPSConnectionV3, req) return HTTPSHandlerV3(**kwargs) - else: - context = ssl.SSLContext(ssl.PROTOCOL_SSLv3) + elif hasattr(ssl, 'create_default_context'): # Python >= 3.4 + context = ssl.create_default_context(ssl.Purpose.CLIENT_AUTH) + context.options &= ~ssl.OP_NO_SSLv3 # Allow older, not-as-secure SSLv3 + if opts_no_check_certificate: + context.verify_mode = ssl.CERT_NONE + return compat_urllib_request.HTTPSHandler(context=context, **kwargs) + else: # Python < 3.4 + context = ssl.SSLContext(ssl.PROTOCOL_SSLv23) context.verify_mode = (ssl.CERT_NONE if opts_no_check_certificate else ssl.CERT_REQUIRED) @@ -583,13 +740,17 @@ def make_HTTPS_handler(opts_no_check_certificate, **kwargs): class ExtractorError(Exception): """Error during info extraction.""" - def __init__(self, msg, tb=None, expected=False, cause=None): + def __init__(self, msg, tb=None, expected=False, cause=None, video_id=None): """ tb, if given, is the original traceback (so that it can be printed out). If expected is set, this is a normal error message and most likely not a bug in youtube-dl. """ if sys.exc_info()[0] in (compat_urllib_error.URLError, socket.timeout, UnavailableVideoError): expected = True + if video_id is not None: + msg = video_id + ': ' + msg + if cause: + msg += u' (caused by %r)' % cause if not expected: msg = msg + u'; please report this issue on https://yt-dl.org/bug . Be sure to call youtube-dl with the --verbose flag and include its complete output. Make sure you are using the latest version; type youtube-dl -U to update.' super(ExtractorError, self).__init__(msg) @@ -597,6 +758,7 @@ class ExtractorError(Exception): self.traceback = tb self.exc_info = sys.exc_info() # preserve original exception self.cause = cause + self.video_id = video_id def format_traceback(self): if self.traceback is None: @@ -703,10 +865,9 @@ class YoutubeDLHandler(compat_urllib_request.HTTPHandler): return ret def http_request(self, req): - for h,v in std_headers.items(): - if h in req.headers: - del req.headers[h] - req.add_header(h, v) + for h, v in std_headers.items(): + if h 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'] @@ -716,6 +877,12 @@ class YoutubeDLHandler(compat_urllib_request.HTTPHandler): del req.headers['User-agent'] req.headers['User-agent'] = req.headers['Youtubedl-user-agent'] del req.headers['Youtubedl-user-agent'] + + if sys.version_info < (2, 7) and '#' in req.get_full_url(): + # Python 2.6 is brain-dead when it comes to fragments + req._Request__original = req._Request__original.partition('#')[0] + req._Request__r_type = req._Request__r_type.partition('#')[0] + return req def http_response(self, req, resp): @@ -750,31 +917,73 @@ class YoutubeDLHandler(compat_urllib_request.HTTPHandler): https_request = http_request https_response = http_response + +def parse_iso8601(date_str, delimiter='T'): + """ Return a UNIX timestamp from the given date """ + + if date_str is None: + return None + + m = re.search( + r'(\.[0-9]+)?(?:Z$| ?(?P\+|-)(?P[0-9]{2}):?(?P[0-9]{2})$)', + date_str) + if not m: + timezone = datetime.timedelta() + else: + date_str = date_str[:-len(m.group(0))] + if not m.group('sign'): + timezone = datetime.timedelta() + else: + sign = 1 if m.group('sign') == '+' else -1 + timezone = datetime.timedelta( + hours=sign * int(m.group('hours')), + minutes=sign * int(m.group('minutes'))) + date_format = '%Y-%m-%d{0}%H:%M:%S'.format(delimiter) + dt = datetime.datetime.strptime(date_str, date_format) - timezone + return calendar.timegm(dt.timetuple()) + + def unified_strdate(date_str): """Return a string with the date in the format YYYYMMDD""" + + if date_str is None: + return None + upload_date = None #Replace commas - date_str = date_str.replace(',',' ') + date_str = date_str.replace(',', ' ') # %z (UTC offset) is only supported in python>=3.2 - date_str = re.sub(r' (\+|-)[\d]*$', '', date_str) + date_str = re.sub(r' ?(\+|-)[0-9]{2}:?[0-9]{2}$', '', date_str) format_expressions = [ '%d %B %Y', + '%d %b %Y', '%B %d %Y', '%b %d %Y', + '%b %dst %Y %I:%M%p', + '%b %dnd %Y %I:%M%p', + '%b %dth %Y %I:%M%p', '%Y-%m-%d', + '%Y/%m/%d', + '%d.%m.%Y', '%d/%m/%Y', + '%d/%m/%y', '%Y/%m/%d %H:%M:%S', + '%d/%m/%Y %H:%M:%S', '%Y-%m-%d %H:%M:%S', + '%Y-%m-%d %H:%M:%S.%f', '%d.%m.%Y %H:%M', + '%d.%m.%Y %H.%M', '%Y-%m-%dT%H:%M:%SZ', '%Y-%m-%dT%H:%M:%S.%fZ', '%Y-%m-%dT%H:%M:%S.%f0Z', '%Y-%m-%dT%H:%M:%S', + '%Y-%m-%dT%H:%M:%S.%f', + '%Y-%m-%dT%H:%M', ] for expression in format_expressions: try: upload_date = datetime.datetime.strptime(date_str, expression).strftime('%Y%m%d') - except: + except ValueError: pass if upload_date is None: timetuple = email.utils.parsedate_tz(date_str) @@ -783,6 +992,8 @@ def unified_strdate(date_str): return upload_date def determine_ext(url, default_ext=u'unknown_video'): + if url is None: + return default_ext guess = url.partition(u'?')[0].rpartition(u'.')[2] if re.match(r'^[A-Za-z0-9]+$', guess): return guess @@ -864,25 +1075,97 @@ def platform_name(): return res -def write_string(s, out=None): +def _windows_write_string(s, out): + """ Returns True if the string was written using special methods, + False if it has yet to be written out.""" + # Adapted from http://stackoverflow.com/a/3259271/35070 + + import ctypes + import ctypes.wintypes + + WIN_OUTPUT_IDS = { + 1: -11, + 2: -12, + } + + try: + fileno = out.fileno() + except AttributeError: + # If the output stream doesn't have a fileno, it's virtual + return False + if fileno not in WIN_OUTPUT_IDS: + return False + + GetStdHandle = ctypes.WINFUNCTYPE( + ctypes.wintypes.HANDLE, ctypes.wintypes.DWORD)( + ("GetStdHandle", ctypes.windll.kernel32)) + h = GetStdHandle(WIN_OUTPUT_IDS[fileno]) + + WriteConsoleW = ctypes.WINFUNCTYPE( + ctypes.wintypes.BOOL, ctypes.wintypes.HANDLE, ctypes.wintypes.LPWSTR, + ctypes.wintypes.DWORD, ctypes.POINTER(ctypes.wintypes.DWORD), + ctypes.wintypes.LPVOID)(("WriteConsoleW", ctypes.windll.kernel32)) + written = ctypes.wintypes.DWORD(0) + + GetFileType = ctypes.WINFUNCTYPE(ctypes.wintypes.DWORD, ctypes.wintypes.DWORD)(("GetFileType", ctypes.windll.kernel32)) + FILE_TYPE_CHAR = 0x0002 + FILE_TYPE_REMOTE = 0x8000 + GetConsoleMode = ctypes.WINFUNCTYPE( + ctypes.wintypes.BOOL, ctypes.wintypes.HANDLE, + ctypes.POINTER(ctypes.wintypes.DWORD))( + ("GetConsoleMode", ctypes.windll.kernel32)) + INVALID_HANDLE_VALUE = ctypes.wintypes.DWORD(-1).value + + def not_a_console(handle): + if handle == INVALID_HANDLE_VALUE or handle is None: + return True + return ((GetFileType(handle) & ~FILE_TYPE_REMOTE) != FILE_TYPE_CHAR + or GetConsoleMode(handle, ctypes.byref(ctypes.wintypes.DWORD())) == 0) + + if not_a_console(h): + return False + + def next_nonbmp_pos(s): + try: + return next(i for i, c in enumerate(s) if ord(c) > 0xffff) + except StopIteration: + return len(s) + + while s: + count = min(next_nonbmp_pos(s), 1024) + + ret = WriteConsoleW( + h, s, count if count else 2, ctypes.byref(written), None) + if ret == 0: + raise OSError('Failed to write string') + if not count: # We just wrote a non-BMP character + assert written.value == 2 + s = s[1:] + else: + assert written.value > 0 + s = s[written.value:] + return True + + +def write_string(s, out=None, encoding=None): if out is None: out = sys.stderr assert type(s) == compat_str + if sys.platform == 'win32' and encoding is None and hasattr(out, 'fileno'): + if _windows_write_string(s, out): + return + if ('b' in getattr(out, 'mode', '') or sys.version_info[0] < 3): # Python 2 lies about mode of sys.stderr - s = s.encode(preferredencoding(), 'ignore') - try: + byt = s.encode(encoding or preferredencoding(), 'ignore') + out.write(byt) + elif hasattr(out, 'buffer'): + enc = encoding or getattr(out, 'encoding', None) or preferredencoding() + byt = s.encode(enc, 'ignore') + out.buffer.write(byt) + else: out.write(s) - except UnicodeEncodeError: - # In Windows shells, this can fail even when the codec is just charmap!? - # See https://wiki.python.org/moin/PrintFails#Issue - if sys.platform == 'win32' and hasattr(out, 'encoding'): - s = s.encode(out.encoding, 'ignore').decode(out.encoding) - out.write(s) - else: - raise - out.flush() @@ -904,12 +1187,6 @@ def intlist_to_bytes(xs): return bytes(xs) -def get_cachedir(params={}): - cache_root = os.environ.get('XDG_CACHE_HOME', - os.path.expanduser('~/.cache')) - return params.get('cachedir', os.path.join(cache_root, 'youtube-dl')) - - # Cross-platform file locking if sys.platform == 'win32': import ctypes.wintypes @@ -969,10 +1246,10 @@ else: import fcntl def _lock_file(f, exclusive): - fcntl.lockf(f, fcntl.LOCK_EX if exclusive else fcntl.LOCK_SH) + fcntl.flock(f, fcntl.LOCK_EX if exclusive else fcntl.LOCK_SH) def _unlock_file(f): - fcntl.lockf(f, fcntl.LOCK_UN) + fcntl.flock(f, fcntl.LOCK_UN) class locked_file(object): @@ -1006,11 +1283,14 @@ class locked_file(object): return self.f.read(*args) +def get_filesystem_encoding(): + encoding = sys.getfilesystemencoding() + return encoding if encoding is not None else 'utf-8' + + def shell_quote(args): quoted_args = [] - encoding = sys.getfilesystemencoding() - if encoding is None: - encoding = 'utf-8' + encoding = get_filesystem_encoding() for a in args: if isinstance(a, bytes): # We may get a filename encoded with 'encodeFilename' @@ -1059,13 +1339,8 @@ def format_bytes(bytes): return u'%.2f%s' % (converted, suffix) -def str_to_int(int_str): - int_str = re.sub(r'[,\.]', u'', int_str) - return int(int_str) - - def get_term_width(): - columns = os.environ.get('COLUMNS', None) + columns = compat_getenv('COLUMNS', None) if columns: return int(columns) @@ -1092,9 +1367,12 @@ def month_by_name(name): return None -def fix_xml_all_ampersand(xml_str): +def fix_xml_ampersands(xml_str): """Replace all the '&' by '&' in XML""" - return xml_str.replace(u'&', u'&') + return re.sub( + r'&(?!amp;|lt;|gt;|apos;|quot;|#x[0-9a-fA-F]{,4};|#[0-9]{,4};)', + u'&', + xml_str) def setproctitle(title): @@ -1103,11 +1381,11 @@ def setproctitle(title): libc = ctypes.cdll.LoadLibrary("libc.so.6") except OSError: return - title = title - buf = ctypes.create_string_buffer(len(title) + 1) - buf.value = title.encode('utf-8') + title_bytes = title.encode('utf-8') + buf = ctypes.create_string_buffer(len(title_bytes)) + buf.value = title_bytes try: - libc.prctl(15, ctypes.byref(buf), 0, 0, 0) + libc.prctl(15, buf, 0, 0, 0) except AttributeError: return # Strange libc, just skip this @@ -1118,6 +1396,12 @@ def remove_start(s, start): return s +def remove_end(s, end): + if s.endswith(end): + return s[:-len(end)] + return s + + def url_basename(url): path = compat_urlparse.urlparse(url).path return path.strip(u'/').split(u'/')[-1] @@ -1128,16 +1412,39 @@ class HEADRequest(compat_urllib_request.Request): return "HEAD" -def int_or_none(v): - return v if v is None else int(v) +def int_or_none(v, scale=1, default=None, get_attr=None, invscale=1): + if get_attr: + if v is not None: + v = getattr(v, get_attr, None) + if v == '': + v = None + return default if v is None else (int(v) * invscale // scale) + + +def str_or_none(v, default=None): + return default if v is None else compat_str(v) + + +def str_to_int(int_str): + """ A more relaxed version of int_or_none """ + if int_str is None: + return None + int_str = re.sub(r'[,\.\+]', u'', int_str) + return int(int_str) + + +def float_or_none(v, scale=1, invscale=1, default=None): + return default if v is None else (float(v) * invscale / scale) def parse_duration(s): if s is None: return None + s = s.strip() + m = re.match( - r'(?:(?:(?P[0-9]+):)?(?P[0-9]+):)?(?P[0-9]+)$', s) + r'(?i)(?:(?:(?P[0-9]+)\s*(?:[:h]|hours?)\s*)?(?P[0-9]+)\s*(?:[:m]|mins?|minutes?)\s*)?(?P[0-9]+)(?P\.[0-9]+)?\s*(?:s|secs?|seconds?)?$', s) if not m: return None res = int(m.group('secs')) @@ -1145,6 +1452,8 @@ def parse_duration(s): res += int(m.group('mins')) * 60 if m.group('hours'): res += int(m.group('hours')) * 60 * 60 + if m.group('ms'): + res += float(m.group('ms')) return res @@ -1161,3 +1470,269 @@ def check_executable(exe, args=[]): except OSError: return False return exe + + +class PagedList(object): + def __len__(self): + # This is only useful for tests + return len(self.getslice()) + + +class OnDemandPagedList(PagedList): + def __init__(self, pagefunc, pagesize): + self._pagefunc = pagefunc + self._pagesize = pagesize + + def getslice(self, start=0, end=None): + res = [] + for pagenum in itertools.count(start // self._pagesize): + firstid = pagenum * self._pagesize + nextfirstid = pagenum * self._pagesize + self._pagesize + if start >= nextfirstid: + continue + + page_results = list(self._pagefunc(pagenum)) + + startv = ( + start % self._pagesize + if firstid <= start < nextfirstid + else 0) + + endv = ( + ((end - 1) % self._pagesize) + 1 + if (end is not None and firstid <= end <= nextfirstid) + else None) + + if startv != 0 or endv is not None: + page_results = page_results[startv:endv] + res.extend(page_results) + + # A little optimization - if current page is not "full", ie. does + # not contain page_size videos then we can assume that this page + # is the last one - there are no more ids on further pages - + # i.e. no need to query again. + if len(page_results) + startv < self._pagesize: + break + + # If we got the whole page, but the next page is not interesting, + # break out early as well + if end == nextfirstid: + break + return res + + +class InAdvancePagedList(PagedList): + def __init__(self, pagefunc, pagecount, pagesize): + self._pagefunc = pagefunc + self._pagecount = pagecount + self._pagesize = pagesize + + def getslice(self, start=0, end=None): + res = [] + start_page = start // self._pagesize + end_page = ( + self._pagecount if end is None else (end // self._pagesize + 1)) + skip_elems = start - start_page * self._pagesize + only_more = None if end is None else end - start + for pagenum in range(start_page, end_page): + page = list(self._pagefunc(pagenum)) + if skip_elems: + page = page[skip_elems:] + skip_elems = None + if only_more is not None: + if len(page) < only_more: + only_more -= len(page) + else: + page = page[:only_more] + res.extend(page) + break + res.extend(page) + return res + + +def uppercase_escape(s): + unicode_escape = codecs.getdecoder('unicode_escape') + return re.sub( + r'\\U[0-9a-fA-F]{8}', + lambda m: unicode_escape(m.group(0))[0], + s) + + +def escape_rfc3986(s): + """Escape non-ASCII characters as suggested by RFC 3986""" + if sys.version_info < (3, 0) and isinstance(s, unicode): + s = s.encode('utf-8') + return compat_urllib_parse.quote(s, "%/;:@&=+$,!~*'()?#[]") + + +def escape_url(url): + """Escape URL as suggested by RFC 3986""" + url_parsed = compat_urllib_parse_urlparse(url) + return url_parsed._replace( + path=escape_rfc3986(url_parsed.path), + params=escape_rfc3986(url_parsed.params), + query=escape_rfc3986(url_parsed.query), + fragment=escape_rfc3986(url_parsed.fragment) + ).geturl() + +try: + struct.pack(u'!I', 0) +except TypeError: + # In Python 2.6 (and some 2.7 versions), struct requires a bytes argument + def struct_pack(spec, *args): + if isinstance(spec, compat_str): + spec = spec.encode('ascii') + return struct.pack(spec, *args) + + def struct_unpack(spec, *args): + if isinstance(spec, compat_str): + spec = spec.encode('ascii') + return struct.unpack(spec, *args) +else: + struct_pack = struct.pack + struct_unpack = struct.unpack + + +def read_batch_urls(batch_fd): + def fixup(url): + if not isinstance(url, compat_str): + url = url.decode('utf-8', 'replace') + BOM_UTF8 = u'\xef\xbb\xbf' + if url.startswith(BOM_UTF8): + url = url[len(BOM_UTF8):] + url = url.strip() + if url.startswith(('#', ';', ']')): + return False + return url + + with contextlib.closing(batch_fd) as fd: + return [url for url in map(fixup, fd) if url] + + +def urlencode_postdata(*args, **kargs): + return compat_urllib_parse.urlencode(*args, **kargs).encode('ascii') + + +try: + etree_iter = xml.etree.ElementTree.Element.iter +except AttributeError: # Python <=2.6 + etree_iter = lambda n: n.findall('.//*') + + +def parse_xml(s): + class TreeBuilder(xml.etree.ElementTree.TreeBuilder): + def doctype(self, name, pubid, system): + pass # Ignore doctypes + + parser = xml.etree.ElementTree.XMLParser(target=TreeBuilder()) + kwargs = {'parser': parser} if sys.version_info >= (2, 7) else {} + tree = xml.etree.ElementTree.XML(s.encode('utf-8'), **kwargs) + # Fix up XML parser in Python 2.x + if sys.version_info < (3, 0): + for n in etree_iter(tree): + if n.text is not None: + if not isinstance(n.text, compat_str): + n.text = n.text.decode('utf-8') + return tree + + +if sys.version_info < (3, 0) and sys.platform == 'win32': + def compat_getpass(prompt, *args, **kwargs): + if isinstance(prompt, compat_str): + prompt = prompt.encode(preferredencoding()) + return getpass.getpass(prompt, *args, **kwargs) +else: + compat_getpass = getpass.getpass + + +US_RATINGS = { + 'G': 0, + 'PG': 10, + 'PG-13': 13, + 'R': 16, + 'NC': 18, +} + + +def parse_age_limit(s): + if s is None: + return None + m = re.match(r'^(?P\d{1,2})\+?$', s) + return int(m.group('age')) if m else US_RATINGS.get(s, None) + + +def strip_jsonp(code): + return re.sub(r'(?s)^[a-zA-Z0-9_]+\s*\(\s*(.*)\);?\s*?\s*$', r'\1', code) + + +def js_to_json(code): + def fix_kv(m): + v = m.group(0) + if v in ('true', 'false', 'null'): + return v + if v.startswith('"'): + return v + if v.startswith("'"): + v = v[1:-1] + v = re.sub(r"\\\\|\\'|\"", lambda m: { + '\\\\': '\\\\', + "\\'": "'", + '"': '\\"', + }[m.group(0)], v) + return '"%s"' % v + + res = re.sub(r'''(?x) + "(?:[^"\\]*(?:\\\\|\\")?)*"| + '(?:[^'\\]*(?:\\\\|\\')?)*'| + [a-zA-Z_][a-zA-Z_0-9]* + ''', fix_kv, code) + res = re.sub(r',(\s*\])', lambda m: m.group(1), res) + return res + + +def qualities(quality_ids): + """ Get a numeric quality value out of a list of possible values """ + def q(qid): + try: + return quality_ids.index(qid) + except ValueError: + return -1 + return q + + +DEFAULT_OUTTMPL = '%(title)s-%(id)s.%(ext)s' + +try: + subprocess_check_output = subprocess.check_output +except AttributeError: + def subprocess_check_output(*args, **kwargs): + assert 'input' not in kwargs + p = subprocess.Popen(*args, stdout=subprocess.PIPE, **kwargs) + output, _ = p.communicate() + ret = p.poll() + if ret: + raise subprocess.CalledProcessError(ret, p.args, output=output) + return output + + +def limit_length(s, length): + """ Add ellipses to overly long strings """ + if s is None: + return None + ELLIPSES = '...' + if len(s) > length: + return s[:length - len(ELLIPSES)] + ELLIPSES + return s + + +def version_tuple(v): + return [int(e) for e in v.split('.')] + + +def is_outdated_version(version, limit, assume_new=True): + if not version: + return not assume_new + try: + return version_tuple(version) < version_tuple(limit) + except ValueError: + return not assume_new