/// <summary>
/// Loads a single Html pages and does a microformat parse
/// </summary>
/// <param name="url">The Url of the webpage to be pasred</param>
/// <param name="formatDescriber">A format describer for microformat to be parsed</param>
public void Load(string url, UfFormatDescriber formatDescriber)
{
this.formatDescriber = formatDescriber;
try
{
if (url != string.Empty)
{
// Check for issues with url
url = url.Trim();
url = HttpUtility.UrlDecode(url);
UfWebPage webPage = LoadHtmlDoc(url);
if (webPage != null)
{
Url urlReport = new Url();
urlReport.Address = webPage.Url;
urlReport.Status = webPage.StatusCode;
parsedUrls.Add(urlReport);
DateTime started = DateTime.Now;
if (webPage.StatusCode == 200 && webPage.Html != null)
{
ParseUf(webPage.Html, url, formatDescriber, false, urlReport);
}
if (webPage.StatusCode != 200)
{
throw (new Exception("Could not load url: " + url + " " + webPage.StatusCode));
}
DateTime ended = DateTime.Now;
urlReport.LoadTime = ended.Subtract(started);
Urls.Add(urlReport);
}
}
else
{
throw (new Exception("No Url given"));
}
}
catch (Exception ex)
{
if (ex.Message != string.Empty)
{
throw (new Exception(ex.Message));
}
else
{
throw (new Exception("Could not load Url: " + url));
}
}
}
/// <summary>
/// Converts a UfDataNode structure into JSON
/// </summary>
/// <param name="node">Node</param>
/// <param name="formatArray">Array of microformat format describer to describer data in node</param>
/// <param name="callBack">JSONP callback function name to wrap JSON object</param>
/// <returns>JSON string</returns>
public string Convert(UfDataNode node, ArrayList formatArray, string callBack)
{
this.callBack = callBack;
this.callBack = this.callBack.Replace("(", "").Replace(")", "").Trim();
this.urls = urls;
this.errors = errors;
foreach (UfFormatDescriber formatDescriber in formatArray)
{
foreach (UfDataNode childNode in node.Nodes)
{
foreach (UfDataNode grandChildNode in childNode.Nodes)
{
if (grandChildNode.Name == formatDescriber.BaseElement.Name)
{
UfDataNode xChild = tree.Nodes.Append(grandChildNode.Name, grandChildNode.Value, grandChildNode.SourceUrl, grandChildNode.RepresentativeNode);
if (grandChildNode.Nodes.Count > 0)
AddChildNodes(xChild, grandChildNode, formatDescriber.BaseElement);
}
}
}
}
//string output = "// UfXtract \n";
string output = "";
if (callBack != string.Empty)
output += callBack + "( ";
output += "{";
foreach (UfDataNode childNode in tree.Nodes)
output += BuildDataString(childNode, true, false);
if (tree.Nodes.Count > 0)
output = output.Substring(0, output.Length - 2);
output += AddUfErrors();
output += AddReporting(node);
// End whole block
output += "}";
if (callBack != string.Empty)
output += " )";
//return output.Replace(",", ",\n").Replace("}", "}\n").Replace("{", "{\n").Replace("]", "]\n").Replace("[", "[\n"); ;
return output;
}
