使用mshtml解析html

zoukankan html css js c++ java

使用mshtml解析html

测试用例

测试用例

<html>
<head>
<title>
    Just a Test
</title>
</head>
<body>
gaofeng hello!!
<div>
<table bgcolor="red">

<tr>
<td bgcolor="yellow" border="2">Name</td>
<td id="qualify1" border="1" class="blueBorder" bgcolor=blue></td>
</tr>
<tr>
<td><p id="qualify2" class="blueBorder" bgcolor="blue" border="1">Surname</p></td>
<td></td>
</tr>
<tr>
<td>address</td>
<td></td>
</tr>
</table>
</div>
</body>
</html>

头文件

#include <iostream>
#include <comdef.h>
#include <mshtml.h>
#include <string>
#include <fstream>
#include <vector>
#include <map>
#import <mshtml.tlb> no_auto_exclude

测试代码

代码

// TestMSHTML.cpp : 定义控制台应用程序的入口点。
//

#include "stdafx.h"
#include "TestMSHTML.h"
#ifdef _DEBUG
#define new DEBUG_NEW
#endif

// 唯一的应用程序对象

CWinApp theApp;
FILE * fout;
using namespace std;
//OLECHAR szHTML[] = OLESTR("<HTML><BODY>Hello World!</BODY></HTML>");

typedef  int BorderAttribute;
void FindAllElementHavingBg(IHTMLDocument2 * pNewDoc,map<BorderAttribute,IHTMLElement *>& borderValue2ElementMap)
{
    IHTMLElement * pBody;
    pNewDoc->get_body(&pBody);
    pBody->Release();
}

void PrintTabs(int n)
{
    for (int i = 0;i<n;i++)
    {
        //cout << '\t';
        fwprintf(fout,_T("\t"));
    }
}

void VisitNode(IHTMLElement* pElement,int level)
{
    BSTR strName,strId,strTag;
    PrintTabs(level);
    pElement->get_className(&strName);
    pElement->get_id(&strId);
    pElement->get_tagName(&strTag);
    if (strTag!=NULL)
    {
        fwprintf(fout,_T("TagName:%s "),strTag);
    }
    if (strName!=NULL)
    {
        fwprintf(fout,_T("className:%s "),strName);
    }
    if (strId != NULL)
    {
        fwprintf(fout,_T("Id:%s "),strId);
    }
    SysFreeString(strName);
    SysFreeString(strId);
    SysFreeString(strTag);
    BSTR strAttrName1 = _T("border");
    BSTR strAttrName2 = _T("bgcolor");
    VARIANT val;

    pElement->getAttribute(strAttrName1,2,&val);
    if (val.vt != VT_NULL)
    {
        if (val.bstrVal != NULL)
        {
            fwprintf(fout,_T("border:%s "),val.bstrVal);
        }
    }

    pElement->getAttribute(strAttrName2,2,&val);
    if (val.vt != VT_NULL)
    {
        if (val.bstrVal != NULL)
        {
            fwprintf(fout,_T("bgcolor:%s "),val.bstrVal);
        }
    }


    fwprintf(fout,_T("\n"));
}
//将DOM树打印出来
void Run(IHTMLElement * pElement,int level)
{
    IHTMLElementCollection * children;

    VisitNode(pElement,level);

    IDispatch* pDisp;
    pElement->get_children(&pDisp);
    pDisp->QueryInterface(IID_IHTMLElementCollection,(void**)&children);
    pDisp->Release();

    long len;
    children->get_length(&len);
    VARIANT dummy;
    dummy.vt = VT_I4;
    for (int i = 0;i < len;i++)
    {
        IHTMLElement* child;
        dummy.intVal = i;
        children->item(dummy,dummy,(IDispatch**)&pDisp);
        pDisp->QueryInterface(IID_IHTMLElement,(void**)&child);
        pDisp->Release();
        Run(child,level + 1);
        child->Release();
    }
    children->Release();
}
void TestParse(IHTMLDocument2 * pNewDoc)
{
    BSTR strText;
    IHTMLElement *pBody;
    pNewDoc->get_body(&pBody);
    pBody->get_innerText(&strText);
    wprintf(_T("%s\n"),strText);
    SysFreeString(strText);


    pNewDoc->get_title(&strText);
    wprintf(_T("%s\n"),strText);
    SysFreeString(strText);

    cout << "Run begin...."<<endl;
    Run(pBody,0);
    cout << "Run end...."<<endl;

    pBody->Release();

    //FindAllElementHavingBg(pNewDoc);

}
void TestMSHTML(wchar_t * wcontent)
{
    IHTMLDocument2 *pDoc = NULL;
    CoInitialize(NULL);
    CoCreateInstance(CLSID_HTMLDocument,
                     NULL,
                     CLSCTX_INPROC_SERVER,
                     IID_IHTMLDocument2,
                    (LPVOID *) &pDoc);

    if (pDoc)
    {
        IPersistStreamInit *pPersist = NULL;
        pDoc->QueryInterface(IID_IPersistStreamInit,
                             (LPVOID *) &pPersist);
        if (pPersist)
        {
            IMarkupServices *pMS = NULL;
            pPersist->InitNew();
            pPersist->Release();
            pDoc->QueryInterface(IID_IMarkupServices,
                                (LPVOID *) &pMS);

            if (pMS)
            {
                IMarkupContainer *pMC = NULL;
                IMarkupPointer *pMkStart = NULL;
                IMarkupPointer *pMkFinish = NULL;
                pMS->CreateMarkupPointer(&pMkStart);
                pMS->CreateMarkupPointer(&pMkFinish);
                pMS->ParseString(wcontent,
                    0,
                    &pMC,
                    pMkStart,
                    pMkFinish);

                if (pMC)
                {
                    IHTMLDocument2 *pNewDoc = NULL;

                    pMC->QueryInterface(IID_IHTMLDocument,
                        (LPVOID *) &pNewDoc);

                    if (pNewDoc)
                    {
                        // do anything with pNewDoc, in this case
                        // get the body innerText.
                        TestParse(pNewDoc);

                        pNewDoc->Release();
                    }

                    pMC->Release();
                }

                if (pMkStart)
                    pMkStart->Release();

                if (pMkFinish)
                    pMkFinish->Release();

                pMS->Release();
            }
        }

        pDoc->Release();
    }

    CoUninitialize();

}

inline wchar_t* AnsiToUnicode( const char* szStr )
{
    int nLen = MultiByteToWideChar( CP_ACP, MB_PRECOMPOSED, szStr, -1, NULL, 0 );
    if (nLen == 0)
    {
        return NULL;
    }
    wchar_t* pResult = new wchar_t[nLen+1];
    MultiByteToWideChar( CP_ACP, MB_PRECOMPOSED, szStr, -1, pResult, nLen );
    pResult[nLen] = L'\0';
    return pResult;
}

//调用者负责delete wcontent
wchar_t * ReadFromHtmlFile(string str,string & content)
{
    ifstream fin(str.c_str());
    string line;
    while(getline(fin,line))
    {
    //    cout << line << endl;
        content = content + line;
    }
    //cout << content << endl;
    //cout << content.size() << endl;
    //printf("original html code\n%s\n",content.c_str());
    wchar_t * wcontent = AnsiToUnicode(content.c_str());
    //wprintf(L"after transferred\n%s\n",wcontent);
    //delete[] wcontent;
    fin.close();
    fin.clear();
    return wcontent;
}

int _tmain(int argc, TCHAR* argv[], TCHAR* envp[])
{
    int nRetCode = 0;

    // 初始化 MFC 并在失败时显示错误
    if (!AfxWinInit(::GetModuleHandle(NULL), NULL, ::GetCommandLine(), 0))
    {
        // TODO: 更改错误代码以符合您的需要
        _tprintf(_T("错误: MFC 初始化失败\n"));
        nRetCode = 1;
    }
    else
    {
        fout = fopen("out.txt","w");
        string str = "test.html";
        string content;
        wchar_t * wcontent = ReadFromHtmlFile(str,content);
        int len = wcslen(wcontent);
        //cout << len << endl;

        TestMSHTML(wcontent);
        delete[] wcontent;
        fclose(fout);
    }

    return nRetCode;
}

输入结果

TagName:BODY
TagName:DIV
  TagName:TABLE bgcolor:#ff0000
   TagName:TBODY
    TagName:TR
     TagName:TD border:2 bgcolor:#ffff00
     TagName:TD className:blueBorder Id:qualify1 border:1 bgcolor:#0000ff
    TagName:TR
     TagName:TD
      TagName:P className:blueBorder Id:qualify2 border:1 bgcolor:blue
     TagName:TD
    TagName:TR
     TagName:TD
     TagName:TD

查看全文

相关阅读:
CenterNet-TensorRT错误记录
 NAS研究要点分析
 conda如何安装从源下载的离线安装包
 Xavier上pytorch半精度inference问题
 Xavier 使用便携程序
 Xavier疑问
 Python输入(Leetcode
兴趣爱好
 生活目标
 TX2装机教程

原文地址：https://www.cnblogs.com/speedmancs/p/1797442.html