public void OnFailed(string processUrl, int processLevel, Exception ex)
{
Console.ForegroundColor = ConsoleColor.Red;
Console.WriteLine($"{processLevel}, {processUrl}, {CollectTools.GetExceptionMessages(ex)}");
Console.ResetColor();
//delete url from 0
//move url to 1
CollectionStore.MoveUrl(rootUrl, new CollectionStore.UrlModel()
{
Url = processUrl,
Level = processLevel,
localFileName = null
});
}
public string LoadHtml()
{
if (string.IsNullOrEmpty(this.url))
{
throw new ArgumentNullException("Url");
}
string htmlContent = CollectTools.GetHtmlContent(url, encoding);
if (string.IsNullOrEmpty(htmlContent))
{
throw new Exception($"Load html content by url = {url} and encoding = {encoding.ToString()} failure.");
}
HtmlDocument document = new HtmlDocument();
document.LoadHtml(htmlContent);
this.htmlNode = document.DocumentNode;
return(htmlContent);
}
public string Save(string rootUrl, ExtractResult result)
{
if (result == null || result.ContentResults == null)
{
return(null);
}
if (result.ContentResults.ContainsKey("title") && !result.ContentResults["title"].HasValue())
{
return(null);
}
if (result.ContentResults.ContainsKey("content") && !result.ContentResults["content"].HasValue())
{
return(null);
}
string saveDir = $"E:\\hwl-collect\\{HttpUtility.UrlEncode(rootUrl)}";
string savePath = $"{saveDir}\\{Guid.NewGuid().ToString()}.txt";
try
{
if (!Directory.Exists(saveDir))
{
Directory.CreateDirectory(saveDir);
}
File.AppendAllLines(savePath, new string[] { result.OriginUrl, result.Level.ToString(), result.ConvertContentToJsonString() }, Encoding.UTF8);
return(savePath);
}
catch (Exception ex)
{
Console.ForegroundColor = ConsoleColor.Red;
Console.WriteLine($"Save html content to path of {savePath} failed, {CollectTools.GetExceptionMessages(ex)}.");
Console.ResetColor();
}
return(null);
}
