public static async Task GetText(string url)
{
HttpClient http = new HttpClient();
var link = System.Uri.UnescapeDataString(url);
var response = await http.GetByteArrayAsync(link);
String source = Encoding.GetEncoding("utf-8").GetString(response, 0, response.Length - 1);
source = WebUtility.HtmlDecode(source);
HtmlDocument document = new HtmlDocument();
document.LoadHtml(source);
document = HtmlAgilityPackServices.RemoveScripts(document);
var text = document.DocumentNode.SelectNodes("//body//text()").Select(node => node.InnerText);
StringBuilder output = new StringBuilder();
string[] temp = new string[2000];
Count = 0;
foreach (string line in text)
{
if (!line.Contains("\n"))
{
temp[Count] = line;
Count++;
}
}
Text = new string[Count];
for (int i = 0; i < Count; i++)
{
Text[i] = temp[i];
}
// string textOnly = WebUtility.HtmlDecode(output.ToString());
}
public static async Task <string> formatHtml(string url)
{
try
{
HtmlDocument document = new HtmlDocument();
document = await HtmlAgilityPackServices.InitAsync(url);
document = HtmlAgilityPackServices.RemoveStyleAttributes(document);
document = HtmlAgilityPackServices.RemoveScripts(document);
document = HtmlAgilityPackServices.RemoveAllAttributes(document);
// document = HtmlAgilityPackServices.RemoveAllText(document);
//var parser = new HtmlParser();
//var document = parser.Parse(source);
string source = document.DocumentNode.InnerHtml;
return(source);
//remove comment
//var nodes = document.DocumentNode.SelectNodes("//comment()");
//if (nodes != null)
//{
// foreach (var comment in nodes)
// {
// comment.ParentNode.RemoveChild(comment);
// }
//}
//Content content = new Content();
//await ParserExtension.GetText(Search.Link);
//string[] text = new string[ParserExtension.Count];
//text = ParserExtension.Text;
//return new ParserViewModel(Search.Title, Search.Link, text);
}
catch (Exception e)
{
String error = e.StackTrace;
}
return(null);
}