|
public PrependChild ( |
||
| newChild | The node to add. May not be null. | |
| return | ||
//=========================================================================================
// 実体のないファイルの場合、画像やリンクは相対アドレスは不可能なので、絶対アドレスへと置き換える。
//=========================================================================================
private string RelativeToAbsolute(string html)
{
try
{
string fn = strCurFileFullPath;
fn = fn.Replace("\\", "/");
string strBaseUrl = fn;
HtmlAgilityPack.HtmlDocument doc = new HtmlAgilityPack.HtmlDocument();
doc.LoadHtml(html);
bool isExistBaseHref = false;
foreach (var nodeBase in doc.DocumentNode.Descendants("base"))
{
if (nodeBase.Attributes["href"].Value.Length > 0)
{
isExistBaseHref = true;
}
else
{
}
}
// base hrefの指定が無いのであれば、現在の開いているファイルのディレクトリをそこにしておくことで、相対ディレクトリをフォローする。
if (!isExistBaseHref)
{
string basedir = System.IO.Path.GetDirectoryName(strBaseUrl);
HtmlAgilityPack.HtmlNode baseNode = HtmlAgilityPack.HtmlNode.CreateNode("<base href=''>");
baseNode.Attributes["href"].Value = basedir + "\\";
// Headタグがあればそこにたす
HtmlAgilityPack.HtmlNode nodeHead = doc.DocumentNode.SelectSingleNode("/html/head");
HtmlAgilityPack.HtmlNode nodeHtml = doc.DocumentNode.SelectSingleNode("/html");
if (nodeHead != null)
{
nodeHead.PrependChild(baseNode);
}
else if (nodeHtml != null)
{
nodeHtml.PrependChild(baseNode);
}
// Headタグがないなら、トップにたさざるをえないだろう
else
{
doc.DocumentNode.PrependChild(baseNode);
}
}
StringWriter writer = new StringWriter();
doc.Save(writer);
string newHtml = writer.ToString();
return(newHtml);
}
catch (Exception)
{
}
return(html);
}
public override void Parse(Response response)
{
//Create a new HTMLAglityPack document
HtmlDocument ContentDocument = new HtmlDocument();
//load the #content of the page into the document
ContentDocument.LoadHtml(response.Css("#content").First().OuterHtml);
HtmlAgilityPack.HtmlNode BodyNode = ContentDocument.DocumentNode;
patternObject.Title = BodyNode.SelectSingleNode("//*[@id=\"firstHeading\"]").InnerHtml;
HtmlAgilityPack.HtmlNode ContentNode = BodyNode.SelectSingleNode("//*[@id=\"mw-content-text\"]");
//remove the "toc" and "jump" and "siteSub" sections to save space and later client-side processing time
if (ContentNode.SelectSingleNode("//*[@id=\"toc\"]") != null)
{
ContentNode.SelectSingleNode("//*[@id=\"toc\"]").Remove();
}
foreach (var node in ContentNode.SelectNodes("//comment()"))
{
node.Remove();
}
ContentNode.PrependChild(BodyNode.SelectSingleNode("//*[@id=\"firstHeading\"]"));
//set the patternObject's title
patternObject.Title = ContentNode.SelectSingleNode("//*[@id=\"firstHeading\"]").InnerHtml;
foreach (var link in ContentNode.SelectNodes("//a/@href"))
{
//skip if this is a redlink (page doesn't exist).
if (link.Attributes["href"].Value.Contains("redlink=1"))
{
continue;
}
//skip if this links to this page
if (link.Attributes["href"].Value.Split('#').First() == response.FinalUrl)
{
continue;
}
//if any of the links ancestor nodes is the "category links" part of the page
if (link.Ancestors().Any(node => node.Id == "catlinks"))
{
if (link.InnerText != "Categories") //if it is not the "categories" special page
{
//add it to the patterns list of categories
patternObject.Categories.Add(link.InnerText);
}
}
else //assume its a normal text-body link
{
//check if we don't already know about this link
patternObject.CreateOrGetPatternLink(link.InnerText);
}
//add relation info if this is a relation link
if (GetNodeReleventPageHeading(link, "h2") != null &&
GetNodeReleventPageHeading(link, "h2").InnerText == "Relations")
{
//get the relation type of this relation and get its inner text
HtmlAgilityPack.HtmlNode RelationHeadingNode = GetNodeReleventPageHeading(link, "h3");
String RelationName = RelationHeadingNode.InnerText;
//if there is a h4 node before the previous h3 node
if (GetNodeReleventPageHeading(link, "h4") != null &&
RelationHeadingNode.InnerStartIndex < GetNodeReleventPageHeading(link, "h4").InnerStartIndex)
{
//assume it is a "with x" sub-category of relation for the "Can Instantiate" section
RelationName = RelationHeadingNode.InnerText + " " + GetNodeReleventPageHeading(link, "h4").InnerText;
}
//add the relevent relation to this link
patternObject.CreateOrGetPatternLink(link.InnerText).Type.Add(RelationName);
}
}
//get a cleaned copy of the #content HTML for giving in the JSON data
patternObject.Content = ProcessPageContentToString(ContentNode);
string Json = JsonConvert.SerializeObject(patternObject);
File.WriteAllText(Pattern.GetFileName(patternObject.Title), Json);
}
