/// <summary>
/// Sanitizes the specified HTML.
/// </summary>
/// <param name="html">The HTML to sanitize.</param>
/// <param name="baseUrl">The base URL relative URLs are resolved against. No resolution if empty.</param>
/// <returns>The sanitized HTML.</returns>
public string Sanitize(string html, string baseUrl = "")
{
var dom = CQ.Create(html);
dom["*"].Not(string.Join(",", AllowedTags.ToArray())).Remove();
foreach (var tag in dom["*"])
{
foreach (var attribute in tag.Attributes.Where(a => !AllowedAttributesSet.Contains(a.Key)).ToList())
{
tag.RemoveAttribute(attribute.Key);
}
foreach (var attribute in tag.Attributes.Where(a => UriAttributes.Contains(a.Key)).ToList())
{
var url = SanitizeUrl(attribute.Value, baseUrl);
if (url == null)
{
tag.RemoveAttribute(attribute.Key);
}
else
{
tag.SetAttribute(attribute.Key, url);
}
}
SanitizeStyle(tag.Style, baseUrl);
foreach (var attribute in tag.Attributes.ToList())
{
if (JSInclude.IsMatch(attribute.Value))
{
tag.RemoveAttribute(attribute.Key);
}
var val = attribute.Value;
if (val.Contains('<'))
{
val = val.Replace("<", "<"); tag.SetAttribute(attribute.Key, val);
}
if (val.Contains('>'))
{
val = val.Replace(">", ">"); tag.SetAttribute(attribute.Key, val);
}
}
}
var output = dom.Render(DomRenderingOptions.RemoveComments | DomRenderingOptions.QuoteAllAttributes);
return(output);
}
private void Sanitize(INode node)
{
if (node is IElement htmlElement)
{
if (!IsValidNode(htmlElement.TagName))
{
htmlElement.Remove();
return;
}
for (var i = htmlElement.Attributes.Length - 1; i >= 0; i--)
{
var attribute = htmlElement.Attributes[i];
if (!IsValidAttribute(attribute.Name))
{
htmlElement.RemoveAttribute(attribute.NamespaceUri, attribute.Name);
}
else if (UriAttributes.Contains(attribute.Name))
{
if (!UrlSanitizer.IsSafeUrl(attribute.Value))
{
attribute.Value = "";
}
}
else if (SrcsetAttributes.Contains(attribute.Name))
{
if (!UrlSanitizer.IsSafeSrcset(attribute.Value))
{
attribute.Value = "";
}
}
}
}
for (var i = node.ChildNodes.Length - 1; i >= 0; i--)
{
Sanitize(node.ChildNodes[i]);
}
}
/// <summary>
/// Determines whether the specified attribute can contain a URI.
/// </summary>
/// <param name="attribute">The attribute.</param>
/// <returns><c>true</c> if the attribute can contain a URI; otherwise, <c>false</c>.</returns>
private bool IsUriAttribute(IAttr attribute)
{
return(UriAttributes.Contains(attribute.Name));
}
/// <summary>
/// Determines whether the specified attribute can contain a URI.
/// </summary>
/// <param name="attribute">The attribute.</param>
/// <returns><c>true</c> if the attribute can contain a URI; otherwise, <c>false</c>.</returns>
private bool IsUriAttribute(KeyValuePair <string, string> attribute)
{
return(UriAttributes.Contains(attribute.Key));
}
