解析html程序 C# (mshtml)

 
递归枚举IHTMLDocument2的所有元素  
http://blog.csdn.net/fishmai/article/details/52388843 

void  EnumHTMLDocument( MSHTML::IHTMLDocument2* pDoc )  
{  
    if( pDoc == NULL )return;  
  
    //遍历搜索子框架,递归处理子框架的文档  
    CComPtr<MSHTML::IHTMLFramesCollection2>  spFramesCollection;  
  
    pDoc->get_frames( &spFramesCollection );  
  
    long lCount = 0;  
    HRESULT hr = spFramesCollection->get_length( &lCount );  
    if( FAILED( hr ) )return;  
  
    for ( long lIndex = 0; lIndex < lCount; lIndex++ )  
    {  
        CComVariant  vDispWin;  
        vDispWin  = spFramesCollection->item( &CComVariant( lIndex ) );  
  
        CComQIPtr<MSHTML::IHTMLWindow2>  spWin = vDispWin.pdispVal;  
        if( spWin == NULL )continue;  
  
        CComPtr<MSHTML::IHTMLDocument2> spSubDoc;  
        spWin->get_document( &spSubDoc );  
  
        EnumHTMLDocument( spSubDoc );  
    }  
  
    CComQIPtr<MSHTML::IHTMLElementCollection>  spElementCollection;  
    hr  =  pDoc->get_forms( &spElementCollection );  
    if( FAILED( hr ) )return;  
  
    long lFormCount = 0;  
    hr  = spElementCollection->get_length( &lFormCount );  
    if( FAILED( hr ) )return;  
  
    for ( long lIndex = 0; lIndex < lFormCount; lIndex++ )  
    {  
        CComQIPtr<MSHTML::IHTMLFormElement> spFormElement =   spElementCollection->item(  &CComVariant( lIndex ) );  
        if( spFormElement == NULL )continue;  
  
        long lElemCount = 0;  
        hr  = spFormElement->get_length( &lElemCount );  
        if( FAILED( hr ) )continue;  
  
        for ( long lElemIndex = 0; lElemIndex < lElemCount; lElemIndex++ )  
        {  
            CComDispatchDriver  spInputElement;  
            spInputElement  =  spFormElement->item( &CComVariant( lElemIndex ) );  
            if( spInputElement == NULL )continue;  
  
            CComVariant varName, varValue, varType;  
            hr = spInputElement.GetPropertyByName( L"name", &varName );  
            if( SUCCEEDED( hr ) )  
            {  
                LPCTSTR lpszName = varName.bstrVal ? COLE2CT( varName.bstrVal ) : _T("NULL");  
                AtlMessageBox( NULL, lpszName );  
            }  
  
            hr = spInputElement.GetPropertyByName( L"value", &varValue );  
            if( SUCCEEDED( hr ) )  
            {  
                LPCTSTR lpszValue = varValue.bstrVal ? COLE2CT( varValue.bstrVal ) : _T("NULL");  
                AtlMessageBox( NULL, lpszValue );  
            }  
  
            hr = spInputElement.GetPropertyByName( L"type", &varType );  
            if( SUCCEEDED( hr ) )  
            {  
                LPCTSTR lpszType = varType.bstrVal ? COLE2CT( varType.bstrVal ) : _T("NULL");  
                AtlMessageBox( NULL, lpszType );  
            }  
        }  
    }  
  
}  

 
 解析html程序(C#版)——遍历各个节点(mshtml)   .
http://blog.csdn.net/hanjieson/article/details/8576150

/*

在项目里引用了mshtml.dll，并且引用命名空间：using mshtml;

首先，参数html就是html文本内容（里面有markup标记和显示文本等等）

其次，getHtmlDisplayContent这个函数就是获取html里浏览器上可看到的内容，即从源码中取出显示文本。

 最后，traverseNodes是个人写的一个遍历各个节点的一个小小递归程序，没考虑效率什么的，只是想知道怎么使用IHtmlDocument2和IHtmlDocument3接口

Note：当html文档不规范时，比如在<!Document....之前还有别的标记或者符号时，加载工作受到严重影响，此时估计是解析不出来了，我开始还不知道为什么解析有些html时卡住了，原来是因为这些html文档在html标记前有\n\n\n....等。。。

*/  

private static string getHtmlDisplayContent(string html)

        {

            string cont = "";

            mshtml.HTMLDocumentClass oc = new mshtml.HTMLDocumentClass();

            mshtml.IHTMLDocument2 doc2 = oc;

            doc2.write(html);

            

            mshtml.IHTMLDocument3 HTMLDocument = (mshtml.IHTMLDocument3)doc2;

            traverseNodes(HTMLDocument.documentElement, ref cont);

                //mshtml.IHTMLTitleElement title = (mshtml.IHTMLTitleElement)doc2.title;

               /* cont += doc2.title.ToString();

            mshtml.IHTMLBodyElement body = (mshtml.IHTMLBodyElement)doc2.body;

            if (body.text!=null)

                cont += body.text.ToString();

                * */

            doc2.close();

            return cont;

        }

 

private static void traverseNodes(mshtml.IHTMLElement parentNode,ref string cont)

        {

            if (parentNode.innerText!=null)

                cont += parentNode.innerText;

            mshtml.IHTMLElementCollection nodes = (IHTMLElementCollection)parentNode.children;

            IEnumerator ienum= nodes.GetEnumerator();

            while (ienum.MoveNext())

            {

                IHTMLElement node = (IHTMLElement)ienum.Current;

                traverseNodes(node,ref cont);

            }

            

        }