I have managed to get it done!
Very much thanks to this post which helped me a lot with the xpath expression:
http://social.msdn.microsoft.com/Forums/en-US/regexp/thread/beae72d6-844f-4a9b-ad56-82869d685037/
My task was to add X keywords to the article using table of keywords and urls on my database.
Once keyword was matched - it won't search for it again, but will try to find the next keyword in the text.
The 'keyword' could have been made of more than one word. That's why i added the Replace(" ", "s+").
Also, i had to give precedence to the longest keywords first. That is if i had:
"good day" and "good" as two different keywords - "good day" always wins.
This is my solution:
static public string AddLinksToArticle(string article, int linksToAdd)
{
try
{
//load keywords and urls
var dt = new DAL().GetArticleLinks();
//sort the it
IEnumerable<ArticlesRow> sortedArticles = dt.OrderBy(row => row.keyword, new StringLengthComparer());
// iterate the dictionary to get keyword to replace with anchor
foreach (var item in sortedArticles)
{
article = FindAndReplaceKeywordWithAnchor(article, item.keyword, item.url, ref linksToAdd);
if (linksToAdd == 0)
{
break;
}
}
return article;
}
catch (Exception ex)
{
Utils.LogErrorAdmin(ex);
return null;
}
}
private static string FindAndReplaceKeywordWithAnchor(string article, string keyword, string url, ref int linksToAdd)
{
//convert text to html
HtmlAgilityPack.HtmlDocument doc = new HtmlAgilityPack.HtmlDocument();
doc.LoadHtml(article);
// w* - means it can start with any alphanumeric charactar
// s+ - was placed to replace all white spaces (when there is more than one word).
// - set bounderies for the keyword
string pattern = @"" + keyword.Trim().Insert(0, "\w*").Replace(" ", "\s+") + @"";
//get all elements text propery except for anchor element
var nodes = doc.DocumentNode.SelectNodes("//text()[not(ancestor::a)]") ?? new HtmlAgilityPack.HtmlNodeCollection(null);
foreach (var node in nodes)
{
if (node.InnerHtml.Contains(keyword))
{
Regex regex = new Regex(pattern);
node.InnerHtml = regex.Replace(node.InnerHtml, "<a href="" + url + "">" + keyword + "</a>", 1);//match only first occurrence
linksToAdd--;
break;
}
}
return doc.DocumentNode.OuterHtml;
}
}
public class StringLengthComparer : IComparer<string>
{
public int Compare(string x, string y)
{
return y.Length.CompareTo(x.Length);
}
}
Hope it will help someone in the future.
与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…