protected override void OnBeginTag(BeginTag tag)
{
if (tag.NameEquals(HTMLTokens.Title) && !tag.Complete)
{
_inTitle = true;
}
if (TagsToPreserve.Contains(tag.Name.ToUpper(CultureInfo.InvariantCulture)))
{
EmitTagAndAttributes(tag.Name, tag);
}
else if (ReplaceTags.ContainsKey(tag.Name.ToUpper(CultureInfo.InvariantCulture)))
{
EmitTagAndAttributes((string)ReplaceTags[tag.Name.ToUpper(CultureInfo.InvariantCulture)], tag);
}
}
protected override void OnEndTag(EndTag tag)
{
if (tag.Implicit)
{
return;
}
if (tag.NameEquals(HTMLTokens.Title))
{
_inTitle = false;
}
if (TagsToPreserve.Contains(tag.Name.ToUpper(CultureInfo.InvariantCulture)))
{
EmitTagAndAttributes(tag.Name, tag);
}
else if (ReplaceTags.ContainsKey(tag.Name.ToUpper(CultureInfo.InvariantCulture)))
{
EmitTagAndAttributes((string)ReplaceTags[tag.Name.ToUpper(CultureInfo.InvariantCulture)], tag);
}
}
private void HandleNextNode(Queue <HtmlNode> nodes, IDataMap map, string itemPath, Item importRow)
{
var node = nodes.Dequeue();
var nodeName = node.Name.ToLower();
var parentNode = node.ParentNode;
var childNodes = node.SelectNodes("./*|./text()");
if (childNodes != null)
{
foreach (var child in childNodes)
{
nodes.Enqueue(child);
}
}
if (UnwantedTags.Any(tag => tag == nodeName))
{
// if this node is one to remove
if (childNodes != null)
{
// make sure children are added back
foreach (var child in childNodes)
{
parentNode.InsertBefore(child, node);
}
}
parentNode.RemoveChild(node);
}
else if (node.HasAttributes)
{
// if it's not being removed
foreach (string s in UnwantedAttributes) // remove unwanted attributes
{
node.Attributes.Remove(s);
}
//replace images
if (nodeName.Equals("img"))
{
// see if it exists
string imgSrc = node.Attributes["src"].Value;
DynamicLink dynamicLink;
if (!DynamicLink.TryParse(imgSrc, out dynamicLink))
{
return;
}
MediaItem mediaItem = importRow.Database.GetItem(dynamicLink.ItemId, dynamicLink.Language ?? map.ImportToLanguage);
var mediaParentItem = BuildMediaPath(map.ToDB, mediaItem.InnerItem.Paths.ParentPath);
MediaItem newImg = HandleMediaItem(map, mediaParentItem, itemPath, mediaItem);
if (newImg != null)
{
string newSrc = string.Format("-/media/{0}.ashx", newImg.ID.ToShortID().ToString());
// replace the node with sitecore tag
node.SetAttributeValue("src", newSrc);
}
}
}
else if (ReplaceTags.ContainsKey(nodeName))
{
// Replace tag
node.Name = ReplaceTags[nodeName];
}
else
{
//Keep node as is
}
}
/// <summary>
/// Used as a part of HTML thinning to remove extraneous child nodes from an HTMLDOMNode
/// </summary>
/// <param name="node">The node whose children should be stripped</param>
/// <returns>An HTML string with the DOMNodes cleaned out</returns>
private static void StripChildNodes(IHTMLDOMNode node, StringBuilder escapedText, bool preserveImages, TickableProgressTick progress)
{
// is this a text node? If so, just get the text and return it
if (node.nodeType == HTMLDocumentHelper.HTMLDOMNodeTypes.TextNode)
{
escapedText.Append(HttpUtility.HtmlEncode(node.nodeValue.ToString()));
}
else
{
progress.Tick();
bool tagStillOpen = false;
ArrayList preserveTags = PreserveTags;
if (preserveImages)
{
preserveTags = PreserveTagsWithImages;
}
// if we're in an element node (a tag) and we should preserve the tag,
// append it to the returned text
if (preserveTags.Contains(node.nodeName))
{
// Append the opening tag element, with any extraneous
// attributes stripped
escapedText.Append("<" + node.nodeName);
StripAttributes((IHTMLElement)node, escapedText);
// if the element has no children, we can simply close out the tag
if (!node.hasChildNodes())
{
if (node.nodeName == HTMLTokens.IFrame)
{
escapedText.Append("></" + node.nodeName + ">");
}
else
{
escapedText.Append("/>");
}
}
else // the element has children, leave the tag open
{
escapedText.Append(">");
tagStillOpen = true;
}
}
else if (ReplaceTags.Contains(node.nodeName))
{
// If there are no children, just emit the replacement tag
if (!node.hasChildNodes())
{
// Replace the tag
escapedText.Append("<" + (string)ReplaceTags[node.nodeName] + "/>");
}
else
{
if (!IsChildlessTag((string)ReplaceTags[node.nodeName]))
{
escapedText.Append("<" + (string)ReplaceTags[node.nodeName] + ">");
}
// Since there are children, we're going to emit the replacement
// tag at the end of this node
tagStillOpen = true;
}
}
if (node.firstChild != null)
{
StripChildNodes(node.firstChild, escapedText, preserveImages, progress);
}
// put a closing tag in for the current element (because we left it open in case of children)
if (tagStillOpen)
{
if (PreserveTags.Contains(node.nodeName))
{
escapedText.Append("</" + node.nodeName + ">");
}
else if (ReplaceTags.Contains(node.nodeName))
{
if (!IsChildlessTag((string)ReplaceTags[node.nodeName]))
{
escapedText.Append("</" + (string)ReplaceTags[node.nodeName] + ">");
}
else
{
escapedText.Append("<" + (string)ReplaceTags[node.nodeName] + "/>");
}
}
}
}
if (node.nextSibling != null)
{
StripChildNodes(node.nextSibling, escapedText, preserveImages, progress);
}
}
