IFilter接口，如何提取二进制流中的文本？求助高手！

[zhangxue]

IFilter可以轻松提取磁盘文件中的文本，如office系列文档，但如果不是磁盘文件，例如将文档二进制数据存入了数据库，能否直接对这个二进制流进行文本提取。
下面是我写的一个函数，其中BindIFilterFromStream（）无法定位一个正确的IFilter，请高手帮助解决，非常感谢！！！

/*!
* \brief 通过转换器提取指定字节流的文本
* \param[in] *pStream 字节流，如office文档二进制内容
* \param[in] nStreamSize 字节流的字节数
* \return LPCTSTR 返回的文本内容, NULL 失败
* \author zhangxue
* \date 2005-9-14 16:50:19
* \remarks 获得的文本指针使用完后必须用FreeText()关闭
*/
LPCTSTR CHLFilter::GetText(void *pStream, long nStreamSize)
{
LPTSTR lpszRetval = NULL;
if(!pStream || nStreamSize <= 0)
{
return lpszRetval;
}

_AFX_THREAD_STATE* pState = AfxGetThreadState();
if(!pState->m_bNeedTerm)
{
AfxOleInit();
}

IStream* pIStream;
HRESULT hr;
hr = CreateStreamOnHGlobal(0, TRUE, &pIStream);
if(FAILED(hr))
return lpszRetval;

ULONG lBytesWritten;
hr = pIStream->Write(pStream, nStreamSize, &lBytesWritten);

pIStream->AddRef();

if(FAILED(hr))
return lpszRetval;

LPUNKNOWN pUnkOuter;
LPUNKNOWN ppIUnk;
pUnkOuter = NULL;
ppIUnk = NULL;
hr = BindIFilterFromStream(pIStream, pUnkOuter, (void**)&ppIUnk);
if(FAILED(hr) || !ppIUnk)
{
return lpszRetval;
}

//开始提取文字
LPWSTR lpw = NULL;
IFilter * pFilter;
ULONG uFlag;
ULONG uLen;
SCODE sCode;
STAT_CHUNK sStat;
if(ppIUnk)
{
int iLength = 0;
//2004-12-24 modified by zhangxue 防止堆栈溢出
//WCHAR chunk[MAXLENGTH];
WCHAR *chunk = new WCHAR[MAXLENGTH];
memset(chunk, 0, MAXLENGTH * sizeof(WCHAR));

ppIUnk->QueryInterface(IID_IFilter,(void**)&pFilter);
sCode = pFilter->Init(0, 0, 0, &uFlag);
sCode = pFilter->GetChunk(&sStat);
int iTimes = 0;
while(sCode == S_OK && iTimes < MAXTIMES)
{
while(sCode == S_OK)
{
uLen = MAXLENGTH;
memset(chunk, 0, MAXLENGTH * sizeof(WCHAR));
sCode = pFilter->GetText(&uLen, chunk);
size_t iCheckLen = wcslen(chunk);
if(uLen > 0 && iCheckLen > 0)
{
for(int iIndex = 0; iIndex < (int)iCheckLen; iIndex++)
{
if(chunk[iIndex] == 0x00a0)
chunk[iIndex] = 0x0020;
}
iLength += iCheckLen;
LPWSTR pstrNewString = (LPWSTR)new WCHAR[(iLength + 1) * sizeof(WCHAR)];
memset (pstrNewString, 0, (iLength + 1) * sizeof(WCHAR));
if(lpw)
{
wcscpy(pstrNewString, lpw);
}
wcscat(pstrNewString, chunk);
if(lpw)
{
delete[] lpw;
lpw = NULL;
}
lpw = pstrNewString;
}
}
sCode = pFilter->GetChunk(&sStat);
iTimes ++;
}
if(chunk)
{
delete[] chunk;
chunk = NULL;
}
if(ppIUnk)
{
ppIUnk->Release();
ppIUnk = NULL;
}
if(pFilter)
{
pFilter->Release();
pFilter = NULL;
}
}

if(pIStream)
{
pIStream->Release();
pIStream = NULL;
}

if(lpw)
{
#ifdef _UNICODE
lpszRetval = lpw;
#else
UnicodeToGB(LPCTSTR(lpw), wcslen(lpw) * sizeof(WCHAR), (LPSTR)lpszRetval,0);
delete lpw;
lpw = NULL;
#endif
}
return lpszRetval;
}

[zhangxue]

呜呜~~~每人理我！！！

[wg88888]

关注

[zhangxue]

高手们，救救我吧！