X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;f=youtube_dl%2Futils.py;h=ec186918cd8672ada2da2d5521e0ba8b22eb273d;hb=83548824c29ccdf53a4659260aa3898939833882;hp=1d6ee9964634df197090df0b6f4e9a0b788063dc;hpb=fb47597b09cf8474931a4509bebc510403b30e25;p=youtube-dl

diff --git a/youtube_dl/utils.py b/youtube_dl/utils.py
index 1d6ee9964..ec186918c 100644
--- a/youtube_dl/utils.py
+++ b/youtube_dl/utils.py
@@ -35,6 +35,7 @@ import xml.etree.ElementTree
 import zlib
 
 from .compat import (
+    compat_HTMLParser,
     compat_basestring,
     compat_chr,
     compat_etree_fromstring,
@@ -272,6 +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"""
@@ -1393,6 +1423,7 @@ def parse_count(s):
 
     return lookup_unit_table(_UNIT_TABLE, s)
 
+
 def month_by_name(name):
     """ Return the number of a month by (locale-independently) English name """