如何定制一款12306抢票浏览器——实现自动查询和预订功能

检查是否进入订票页面

判断是否进入订票页面，我是确定了两个标准：（转载请指明出于breaksoftware的csdn博客）

1 网址是否为http://www.12306.cn/mormhweb/kyfw/

2 该页面否有查询按钮

BOOL CDeal12306WebPage::IsQueryPage( CComPtr<IHTMLDocument2> & spDoc, CComBSTR & bstrUrl )
{HRESULT hr = E_FAIL;do  {CString cstrUrl = CString((LPWSTR)bstrUrl);if ( 0 == cstrUrl.CompareNoCase(LOGIN12306URL) ) {CComPtr<IHTMLElement> spQueryButton;hr = GetQueryButtonInQueryPage( spDoc, spQueryButton);CHECKHRPOINTER(hr, spQueryButton);}} while (0);return FAILED(hr) ? FALSE : TRUE;
}

URL很好检测，那么我们如何判断是否存在查询按钮呢？我们先看一下订票页面的页面特征。

解决跨域问题

可以见得订票页面内部嵌入了两个Iframe，而我们关心的那块页面恰恰就是最里面一层IFrame。那我们直接通过最外层的Doc获取到最里面的Doc，然后在最里面的Doc执行有关的查询操作即可。然而熟悉javascript的同学可能马上就会想到“跨域”问题。其实在浏览器层面，跨域问题是很好解决的。

HRESULT CDeal12306WebPage::GetIFrameDoc( CComPtr<IHTMLDocument2>& spDoc, const CString& cstrIFrameName, CComPtr<IHTMLDocument2>& spInnerDoc )
{HRESULT hr = E_FAIL;do {CComQIPtr<IHTMLFramesCollection2> spFrameCollection;hr = spDoc->get_frames(&spFrameCollection);CHECKHRPOINTER(hr, spFrameCollection);CComVariant IframeNameReq = CComBSTR(cstrIFrameName.GetString());CComVariant FramePage;hr = spFrameCollection->item(&IframeNameReq, &FramePage);CHECKHRPOINTER(hr,FramePage.pdispVal);CComPtr<IHTMLWindow2> spIFramePage;hr = FramePage.pdispVal->QueryInterface(IID_IHTMLWindow2, (LPVOID*)&spIFramePage);CHECKHRPOINTER(hr, spIFramePage);hr = spIFramePage->get_document(&spInnerDoc);if ( E_ACCESSDENIED == hr ) {CComQIPtr<IServiceProvider> spServiceProvider = spIFramePage;CHECKPOINT(spServiceProvider);CComQIPtr<IWebBrowser2> spInnerWebBrowser;hr = spServiceProvider->QueryService(IID_IWebBrowserApp, IID_IWebBrowser2, (LPVOID*)&spInnerWebBrowser);CHECKHRPOINTER(hr, spInnerWebBrowser);CComPtr<IDispatch> spDisp;hr = spInnerWebBrowser->get_Document(&spDisp);CHECKHRPOINTER(hr, spDisp);hr = spDisp->QueryInterface(IID_IHTMLDocument2, (LPVOID*)&spInnerDoc);CHECKHRPOINTER(hr, spInnerDoc);}} while (0);return hr;
}

上面这个函数试图在spDoc页面中获取其内嵌的名字是cstrIFrameName的IFrame的Doc。于是我们要获取其中最里面一层Iframe的Doc可以如下调用

HRESULT CDeal12306WebPage::GetIFrameNamedIFramePageDoc( CComPtr<IHTMLDocument2> & spDoc, CComPtr<IHTMLDocument2> & spInnerDoc )
{HRESULT hr = E_FAIL;do {hr =  GetIFrameDoc(spDoc, L"iframepage", spInnerDoc);CHECKHRPOINTER(hr, spInnerDoc);} while (0);return hr;
}HRESULT CDeal12306WebPage::GetIFrameNamedMainDoc( CComPtr<IHTMLDocument2> & spIFramPageDoc,CComPtr<IHTMLDocument2> & spMainDoc )
{HRESULT hr = E_FAIL;do {hr =  GetIFrameDoc(spIFramPageDoc, L"main", spMainDoc);CHECKHRPOINTER(hr, spMainDoc);} while (0);return hr;
}HRESULT CDeal12306WebPage::GetMainDoc( CComPtr<IHTMLDocument2> & spDoc,CComPtr<IHTMLDocument2> & spMainDoc )
{HRESULT hr = E_FAIL;do {CComPtr<IHTMLDocument2> spIFramePageDoc;hr = GetIFrameNamedIFramePageDoc(spDoc, spIFramePageDoc);CHECKHRPOINTER(hr, spIFramePageDoc);hr = GetIFrameNamedMainDoc(spIFramePageDoc, spMainDoc);CHECKHRPOINTER(hr, spMainDoc);} while (0);return hr;
}

当我们获得最里层的Doc后，我们将根据页面结构获取Class为cx_from的Table元素。

          获取这个Table的原因是，之后我们会以该Table为节点，执行“查询按钮”查找的操作。

HRESULT CDeal12306WebPage::GetQueryButtonInQueryPage( CComPtr<IHTMLDocument2> & spDoc, CComPtr<IHTMLElement> & spQueryButtonElem )
{HRESULT hr = E_FAIL;do  {CComPtr<IHTMLDocument2> spMainDoc;hr = GetMainDoc( spDoc, spMainDoc);CHECKHRPOINTER(hr, spMainDoc);CComPtr<IHTMLElement> spEnter_wElem;hr = GetEnter_wElement(spMainDoc, spEnter_wElem );CHECKHRPOINTER(hr, spEnter_wElem);CComPtr<IHTMLElement> spQueryTable;hr = GetQueryTable(spEnter_wElem, spQueryTable);CHECKHRPOINTER(hr, spQueryTable);CComPtr<IHTMLButtonElement> spQueryButton;hr = GetQueryButtonInQueryPage(spQueryTable, spQueryButton);CHECKHRPOINTER(hr, spQueryButton);hr = spQueryButton->QueryInterface(IID_IHTMLElement, (LPVOID*)& spQueryButtonElem);CHECKHRPOINTER(hr, spQueryButtonElem);} while (0);return hr;
}

查询按钮在这个table中的位置是

        于是通过该Table查询”查询“按钮的代码是

HRESULT CDeal12306WebPage::GetQueryButtonInQueryPage( CComPtr<IHTMLElement>& spQueryTable, CComPtr<IHTMLButtonElement> & spQueryButton )
{HRESULT hr = E_FAIL;do {CComPtr<IHTMLElement> spTBody;hr = GetElementByIndex(spQueryTable, 0, spTBody);CHECKHRPOINTER(hr, spTBody);CComPtr<IHTMLElement> spFirstTR;hr = GetElementByIndex(spTBody, 0, spFirstTR);CHECKHRPOINTER(hr, spFirstTR);CComPtr<IHTMLElement> spEighthTR;hr = GetElementByIndex(spFirstTR, 8, spEighthTR);CHECKHRPOINTER(hr, spEighthTR);CComPtr<IHTMLElement> spButtonTemp;hr = GetElementByIndex(spEighthTR, 0, spButtonTemp);CHECKHRPOINTER(hr, spButtonTemp);hr = spButtonTemp->QueryInterface(IID_IHTMLButtonElement, (LPVOID*)&spQueryButton);CHECKHRPOINTER(hr, spQueryButton);} while (0);return hr;
}

插入开始和停止自动查询按钮
为了在该页面中提供给用于控制开启和关闭自动查询功能的按钮，我插入了两个按钮。如下图

我们看下”单程“和”返程“按钮的页面结构

我会在Name为querySingleForm的form下的class为cx_tab的Div下插入“开始”和“停止”按钮。

HRESULT CDeal12306WebPage::InsertButtonInQueryPage( CComPtr<IHTMLDocument2> & spDoc )
{HRESULT hr = E_FAIL;do  {CComPtr<IHTMLDocument2> spMainDoc;hr = GetMainDoc( spDoc, spMainDoc);CHECKHRPOINTER(hr, spMainDoc);CComPtr<IHTMLElement> spEnter_wElem;hr = GetEnter_wElement(spMainDoc, spEnter_wElem );CHECKHRPOINTER(hr, spEnter_wElem);CComPtr<IHTMLElement> spForm;hr = GetQuerySingleForm(spEnter_wElem, spForm);CHECKHRPOINTER(hr, spForm);hr = InsertButtons( spForm );} while (0);return hr;
}

HRESULT CDeal12306WebPage::InsertButtons(CComPtr<IHTMLElement> & spEnter_wElem )
{HRESULT hr = E_FAIL;do {CComPtr<IHTMLElement> spDiv;hr  = GetInsertButtonElem(spEnter_wElem, spDiv);if (  FALSE == IsStartButtonExist(spDiv) ) {hr = InsertStartButton(spDiv);CHECKHR(hr);
#ifdef DEBUGif ( FALSE == IsStartButtonExist(spDiv) ) {DebugBreak();}
#endif}if ( FALSE == IsStopButtonExist(spDiv) ) {hr = InsertStopButton(spDiv);CHECKHR(hr);
#ifdef DEBUGif ( FALSE == IsStopButtonExist(spDiv) ) {DebugBreak();}
#endif}} while (0);return hr ;
}

HRESULT CDeal12306WebPage::GetInsertButtonElem( CComPtr<IHTMLElement> & spForm, CComPtr<IHTMLElement> & spDiv )
{HRESULT hr = E_FAIL;do {CComPtr<IHTMLElement> spCx_TabDiv;hr = GetElementByClassName(spForm, L"cx_tab", spCx_TabDiv);CHECKHRPOINTER(hr, spCx_TabDiv);hr = GetElementByIndex(spCx_TabDiv, 0, spDiv);CHECKHRPOINTER(hr, spDiv);} while (0);return hr;
}HRESULT CDeal12306WebPage::InsertStartButton( CComPtr<IHTMLElement> & spElem )
{HRESULT hr = E_FAIL;do {CComBSTR bstrWhere(L"beforeEnd");CString cstrHTML;cstrHTML.Format( BUTTONFORMAT, STARTBUTTONID, STARTCOMD, L"开始" );CComBSTR bstrHTML(cstrHTML.GetString());hr = spElem->insertAdjacentHTML( bstrWhere, bstrHTML );CHECKHR(hr);} while (0);return hr ;
}HRESULT CDeal12306WebPage::InsertStopButton( CComPtr<IHTMLElement> & spElem )
{HRESULT hr = E_FAIL;do {CComBSTR bstrWhere(L"beforeEnd");CString cstrHTML;cstrHTML.Format( BUTTONFORMAT, STOPBUTTONID, STOPCMD, L"停止" );CComBSTR bstrHTML(cstrHTML.GetString());hr = spElem->insertAdjacentHTML( bstrWhere, bstrHTML );CHECKHR(hr);} while (0);return hr ;
}

#define BUTTONFORMAT    L"<li id=\"%s\"><a href=\"%s\" style=\"width:50px;height:30px;\">%s</a></li>"
#define STARTBUTTONID   L"StartButton"
#define STOPBUTTONID    L"StopButton"

#define STARTCOMD       L"http://www.12306.cn/mormhweb/kyfw/StartQuery.fl"
#define STOPCMD         L"http://www.12306.cn/mormhweb/kyfw/StopQuery.fl"

当我们点击开始按钮是，页面将试图跳转到http://www.12306.cn/mormhweb/kyfw/StartQuery.fl，此时，我将终止该跳转，同时将“开启查询”标志设置为TRUE。

void CBrowserHost::BeforeNavigate2(IDispatch *pDisp, VARIANT *url,VARIANT *Flags, VARIANT *TargetFrameName, VARIANT *PostData,VARIANT *Headers, VARIANT_BOOL *Cancel)
{do  {if ( NULL != url ) {CString cstrUrl((LPWSTR)(url->bstrVal));if ( 0 == cstrUrl.CompareNoCase(SETTINGOK) ) {……}else if ( 0 == cstrUrl.CompareNoCase(STARTCOMD) ) {*Cancel = VARIANT_TRUE;m_AutoMan.SetStart(TRUE);break;}else if (  0 == cstrUrl.CompareNoCase(STOPCMD) ) {*Cancel = VARIANT_TRUE;m_AutoMan.SetStart(FALSE);break;}}*Cancel = VARIANT_FALSE;} while (0);
}

点击停止按钮原理同点击开始按钮原理一致。此处不再赘述。
当用户选择好出发地和目的地及时间后，用户点击查询按钮。并点击“开始”按钮。我们的“人”线程就开始了自动查询操作。
查询是否存在票，有票则预订，无票则再次查询

当我们执行完一次查询后，我们要查看下搜索结果列表信息中用户选择的车次是否存在票。我们先看一下页面结构

        其查找该节点的方法如下

HRESULT CDeal12306WebPage::QueryTicketsInfo( CComPtr<IHTMLDocument2> & spDoc )
{HRESULT hr = E_FAIL;do  {CComPtr<IHTMLDocument2> spMainDoc;hr = GetMainDoc( spDoc, spMainDoc);CHECKHRPOINTER(hr, spMainDoc);CComPtr<IHTMLElement> spEnter_wElem;hr = GetEnter_wElement(spMainDoc, spEnter_wElem );CHECKHRPOINTER(hr, spEnter_wElem);CComPtr<IHTMLElement> spIDGridbox;hr = GetElementByID( spEnter_wElem, L"gridbox", spIDGridbox);CHECKHRPOINTER(hr, spIDGridbox);CComPtr<IHTMLElement> spTable;hr = GetElementByIndex( spIDGridbox, 0, spTable);CHECKHRPOINTER(hr, spTable);CComPtr<IHTMLElement> spTbody;hr = GetElementByIndex( spTable, 0, spTbody);CHECKHRPOINTER(hr, spTbody);CComPtr<IHTMLElement> spTr;hr = GetElementByIndex( spTbody, 1, spTr);CHECKHRPOINTER(hr, spTr);CComPtr<IHTMLElement> spTd;hr = GetElementByIndex(spTr, 0, spTd);CHECKHRPOINTER(hr, spTd);CComPtr<IHTMLElement> spDiv;hr = GetElementByIndex(spTd, 0, spDiv);CHECKHRPOINTER(hr, spDiv);CComPtr<IHTMLElement> spDiv2;hr = GetElementByIndex(spDiv, 0, spDiv2);CHECKHRPOINTER(hr, spDiv2);CComPtr<IHTMLElement> spTable2;hr = GetElementByIndex(spDiv2, 0, spTable2);CHECKHRPOINTER(hr, spTable2);CComPtr<IHTMLElement> spTbody2;hr = GetElementByIndex(spTable2, 0, spTbody2);CHECKHRPOINTER(hr, spTbody2);CComPtr<IHTMLElementCollection> spElemCollection;hr = GetElementCollection(spTbody2, spElemCollection );CHECKHRPOINTER(hr, spElemCollection);long lCount = 0;hr = spElemCollection->get_length(&lCount);CHECKHR(hr);for ( long lindex = 0; lindex < lCount; lindex++ ) {if ( 0 == lindex ) {continue;}CComVariant VarIndex = lindex;CComPtr<IDispatch> spDispatchElem;hr = spElemCollection->item( VarIndex, VarIndex, &spDispatchElem );CHECKHRPOINTER(hr,spDispatchElem);CComPtr<IHTMLElement> spChildTr;hr = spDispatchElem->QueryInterface(IID_IHTMLElement, (LPVOID*)& spChildTr);CHECKHRPOINTER(hr, spChildTr);hr = GetQueryInfoInTr( spChildTr );if ( SUCCEEDED(hr) ) {// 点击了订购按钮了break;}}} while (0);return hr;
}

上述代码执行到第57行时，for循环将逐个读取每列车的信息。为了最快速达到点击“预订”按钮，我将判断的操作放在GetQueryInfoInTr中。

HRESULT CDeal12306WebPage::GetQueryInfoInTr( CComPtr<IHTMLElement> & spElem)
{HRESULT hr = E_FAIL;do {CComPtr<IHTMLElementCollection> spElemCollection;hr = GetElementCollection(spElem, spElemCollection );CHECKHRPOINTER(hr, spElemCollection);long lCount = 0;hr = spElemCollection->get_length(&lCount);CHECKHR(hr);StTrainInfo stTraininfoItem;for ( long lindex = 0; lindex < lCount; lindex++ ) {CComVariant VarIndex = lindex;CComPtr<IDispatch> spDispatchElem;hr = spElemCollection->item( VarIndex, VarIndex, &spDispatchElem );CHECKHRPOINTER(hr,spDispatchElem);CComPtr<IHTMLElement> spChildTd;hr = spDispatchElem->QueryInterface(IID_IHTMLElement, (LPVOID*)& spChildTd);CHECKHRPOINTER(hr, spChildTd);hr = GetQueryInfoSubItem( spChildTd, stTraininfoItem, lindex );   CHECKHR(hr);}CHECKHR(hr);CComPtr<IHTMLElement> spTd;hr = GetElementByIndex( spElem, lCount - 1, spTd);CHECKHRPOINTER(hr, spTd);CComPtr<IHTMLElement> spButton;hr = GetElementByIndex( spTd, 0, spButton );CHECKHRPOINTER(hr, spButton);CComBSTR bstrClassName;hr = spButton->get_className(&bstrClassName);CHECKHR(hr);CString cstrClassName = bstrClassName;if ( 0 == cstrClassName.CompareNoCase(HAVETICKETSACLASS) ) {hr = spButton->click();}else {// 还没有票}m_VecTrainInfo.push_back(stTraininfoItem);} while (0);return hr;
}

我这儿做了简化：只要“预订”按钮变成可点击，即点击之。其实这儿应该做更多的判断，比如用户的席别是否有票。上述代码第44行，即是点击“预订”按钮的操作。
如果没有票，则我们点击“查询”按钮。

HRESULT CDeal12306WebPage::StartQueryInQueryPage( CComPtr<IHTMLDocument2> & spDoc )
{HRESULT hr = S_FALSE;do  {CComPtr<IHTMLElement> spQueryButton;hr = GetQueryButtonInQueryPage( spDoc, spQueryButton);CHECKHRPOINTER(hr, spQueryButton);hr = spQueryButton->click();} while (0);return hr;
}

如此，我们便实现了自动查询和自动订票的功能。