[utils] Add extract_attributes for extracting html tag attributes

[youtube-dl] / youtube_dl / utils.py

diff --git a/youtube_dl/utils.py b/youtube_dl/utils.py
index 672ce05ea17b9d0ab9c009c383810ff54a97dd10..a0234a3a8eadeb5855cbd622758a58ba1a7e8aee 100644 (file)
--- a/youtube_dl/utils.py
+++ b/youtube_dl/utils.py
@@ -4,6 +4,7 @@
 from __future__ import unicode_literals
 
 import base64
+import binascii
 import calendar
 import codecs
 import contextlib
@@ -34,6 +35,7 @@ import xml.etree.ElementTree
 import zlib
 
 from .compat import (
+    compat_HTMLParser,
     compat_basestring,
     compat_chr,
     compat_etree_fromstring,
@@ -159,8 +161,6 @@ if sys.version_info >= (2, 7):
     def find_xpath_attr(node, xpath, key, val=None):
         """ Find the xpath xpath[@key=val] """
         assert re.match(r'^[a-zA-Z_-]+$', key)
-        if val:
-            assert re.match(r'^[a-zA-Z0-9@\s:._-]*$', val)
         expr = xpath + ('[@%s]' % key if val is None else "[@%s='%s']" % (key, val))
         return node.find(expr)
 else:
@@ -273,6 +273,35 @@ def get_element_by_attribute(attribute, value, html):
 
     return unescapeHTML(res)
 
+class HTMLAttributeParser(compat_HTMLParser):
+    """Trivial HTML parser to gather the attributes for a single element"""
+    def __init__(self):
+        self.attrs = { }
+        compat_HTMLParser.__init__(self)
+
+    def handle_starttag(self, tag, attrs):
+        self.attrs = dict(attrs)
+
+def extract_attributes(html_element):
+    """Given a string for an HTML element such as
+    <el
+         a="foo" B="bar" c="&98;az" d=boz
+         empty= noval entity="&amp;"
+         sq='"' dq="'"
+    >
+    Decode and return a dictionary of attributes.
+    {
+        'a': 'foo', 'b': 'bar', c: 'baz', d: 'boz',
+        'empty': '', 'noval': None, 'entity': '&',
+        'sq': '"', 'dq': '\''
+    }.
+    NB HTMLParser is stricter in Python 2.6 & 3.2 than in later versions,
+    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()
+    return parser.attrs
 
 def clean_html(html):
     """Clean an HTML snippet into a readable string"""
@@ -904,9 +933,9 @@ def unified_strdate(date_str, day_first=True):
         '%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',
+        '%b %dst %Y %I:%M',
+        '%b %dnd %Y %I:%M',
+        '%b %dth %Y %I:%M',
         '%Y %m %d',
         '%Y-%m-%d',
         '%Y/%m/%d',
@@ -1569,9 +1598,12 @@ class PagedList(object):
 
 
 class OnDemandPagedList(PagedList):
-    def __init__(self, pagefunc, pagesize):
+    def __init__(self, pagefunc, pagesize, use_cache=False):
         self._pagefunc = pagefunc
         self._pagesize = pagesize
+        self._use_cache = use_cache
+        if use_cache:
+            self._cache = {}
 
     def getslice(self, start=0, end=None):
         res = []
@@ -1581,7 +1613,13 @@ class OnDemandPagedList(PagedList):
             if start >= nextfirstid:
                 continue
 
-            page_results = list(self._pagefunc(pagenum))
+            page_results = None
+            if self._use_cache:
+                page_results = self._cache.get(pagenum)
+            if page_results is None:
+                page_results = list(self._pagefunc(pagenum))
+            if self._use_cache:
+                self._cache[pagenum] = page_results
 
             startv = (
                 start % self._pagesize
@@ -1835,11 +1873,21 @@ def error_to_compat_str(err):
 
 
 def mimetype2ext(mt):
+    ext = {
+        'audio/mp4': 'm4a',
+    }.get(mt)
+    if ext is not None:
+        return ext
+
     _, _, res = mt.rpartition('/')
 
     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',
@@ -2582,3 +2630,58 @@ class PerRequestProxyHandler(compat_urllib_request.ProxyHandler):
             return None  # No Proxy
         return compat_urllib_request.ProxyHandler.proxy_open(
             self, req, proxy, type)
+
+
+def ohdave_rsa_encrypt(data, exponent, modulus):
+    '''
+    Implement OHDave's RSA algorithm. See http://www.ohdave.com/rsa/
+
+    Input:
+        data: data to encrypt, bytes-like object
+        exponent, modulus: parameter e and N of RSA algorithm, both integer
+    Output: hex string of encrypted data
+
+    Limitation: supports one block encryption only
+    '''
+
+    payload = int(binascii.hexlify(data[::-1]), 16)
+    encrypted = pow(payload, exponent, modulus)
+    return '%x' % encrypted
+
+
+def encode_base_n(num, n, table=None):
+    FULL_TABLE = '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ'
+    if not table:
+        table = FULL_TABLE[:n]
+
+    if n > len(table):
+        raise ValueError('base %d exceeds table length %d' % (n, len(table)))
+
+    if num == 0:
+        return table[0]
+
+    ret = ''
+    while num:
+        ret = table[num % n] + ret
+        num = num // n
+    return ret
+
+
+def decode_packed_codes(code):
+    mobj = re.search(
+        r"}\('(.+)',(\d+),(\d+),'([^']+)'\.split\('\|'\)",
+        code)
+    obfucasted_code, base, count, symbols = mobj.groups()
+    base = int(base)
+    count = int(count)
+    symbols = symbols.split('|')
+    symbol_table = {}
+
+    while count:
+        count -= 1
+        base_n_count = encode_base_n(count, base)
+        symbol_table[base_n_count] = symbols[count] or base_n_count
+
+    return re.sub(
+        r'\b(\w+)\b', lambda mobj: symbol_table[mobj.group(0)],
+        obfucasted_code)