public CatalogNodeViewModel(Catalog catalog)
{
CurrentEntity = catalog;
_articlesObservableList = new ObservableCollection<ArticleViewModel>();
ArticlesCollectionView = new ListCollectionView(_articlesObservableList);
}
private async Task<List<Catalog>> ExtractCatalogsFromWebPage(string url)
{
List<Catalog> reusltCatalogs = new List<Catalog>();
var uri = new Uri(url);
var browser1 = new ScrapingBrowser();
browser1.Encoding = Encoding.UTF8;
var html1 = browser1.DownloadString(uri);
var doc = new HtmlDocument();
doc.LoadHtml(html1);
var html = doc.DocumentNode;
foreach (var script in doc.DocumentNode.Descendants("script").ToArray())
{
script.Remove();
}
foreach (var style in doc.DocumentNode.Descendants("style").ToArray())
{
style.Remove();
}
foreach (var comment in doc.DocumentNode.SelectNodes("//comment()").ToArray())
{
comment.Remove();
}
var days = html.CssSelect("div.day");
if (!days.Any())
{
days = html.CssSelect("div#container >div#wrapper >div#content");
}
foreach (var day in days)
{
var catalog = new Catalog();
catalog.IsChecked = true;
var title = day.CssSelect("div.dayTitle").FirstOrDefault();
if(title!=null)
{
catalog.Title = title.InnerText.ClearNotWords();
var atricles = day.CssSelect("div.postTitle");
foreach (var atricle in atricles)
{
var article = new Article();
article.Title = atricle.InnerText.ClearNotWords();
var articleTitleEl = atricle.CssSelect("a.postTitle2");
article.URL = articleTitleEl.First().Attributes["href"].Value;
catalog.Articles.Add(article);
}
}
else
{
catalog.Title = "CataLog"+DateTime.Now.ToShortTimeString();
var atricles = day.CssSelect("div.post");
foreach (var atricle in atricles)
{
var article = new Article();
var articleTitleEl = atricle.CssSelect("a.PostTitle");
article.Title = articleTitleEl.First().InnerText.ClearNotWords();
article.URL = articleTitleEl.First().Attributes["href"].Value;
catalog.Articles.Add(article);
}
}
reusltCatalogs.Add(catalog);
}
return reusltCatalogs;
}
