1；apache poi插件鏈接http://poi.apache.org/ 這個插件主要用於office文件文本內以及富文本(表格，圖片)等的提取，還有支持對已知密碼的office文件的提取，
其他格式的文件加密以及解密都有 見鏈接http://poi.apache.org/encryption.html
http://www.openoffice.org/sc/compdocfileformat.pdf
因爲之前做全文檢索的時候需要提取word03裏面的表格進行分析，使用這個插件對加密的文件不支持，所以研究了下doc文件格式以及它的加密方式，另外office07系列的比較簡單，基本都是對xml解析，這裏就不分析了。再說一句，這裏只是對linux平臺，其他調用微軟com的那套就不用說了。
2；首先先說下word97-03文件是怎麼加密的，word加密時並不是所有字節都加密，只是帶有文本，表格，富文本，圖片的stream纔會加密，涉及到的加密算法有RC4，MD5算法。
具體細節參考：
http://wenku.baidu.com/link?url=TrCAFtr1mHXZbh3qnnOlhYmXTS7-ynw-CES0W6KOUzRzzRg7l04Y5LXl0V3W8pQRGO4SxyzXDlXk5zlLOzPphlXOxfC6UglqRzJlnb6439_
這裏只是列出一些代碼實現：
字符串轉utf8
public static String utf8ToUnicode(String inStr) {
char[] myBuffer = inStr.toCharArray();
StringBuffer sb = new StringBuffer();
for (int i = 0; i < inStr.length(); i++) {
UnicodeBlock ub = UnicodeBlock.of(myBuffer[i]);
if(ub == UnicodeBlock.BASIC_LATIN){
sb.append(myBuffer[i]);
}else if(ub == UnicodeBlock.HALFWIDTH_AND_FULLWIDTH_FORMS){
int j = (int) myBuffer[i] - 65248;
sb.append((char)j);
}else{
short s = (short) myBuffer[i];
String hexS = Integer.toHexString(s);
String unicode = "\\u"+hexS;
sb.append(unicode.toLowerCase());
}
}
return sb.toString();
}//驗證passwd
public boolean verifypwd(byte pwarray[], byte docid[], byte salt[], byte hashedsalt[],HWMPFEMD5 valContext)
{
HWPFRC4 tool =new HWPFRC4();
HWPFRC4 key =new HWPFRC4();
int offset=0, keyoffset=0;
int tocopy=5;
HWMPFEMD5 md5=new HWMPFEMD5();
HWMPFEMD5 md51=new HWMPFEMD5();
md5.md5Init();
md5.md5Update(pwarray, 64);
md5.getMD5StoreDigest(md5);
valContext.md5Init();
while (offset != 16)
{
if ((64 - offset) < 5)
tocopy = 64 - offset;
for(int y=0;y<tocopy;y++)
{
pwarray[offset+y]=md5.digest[keyoffset+y];
}
offset += tocopy;
if (offset == 64)
{
valContext.md5Update(pwarray, 64);
keyoffset = tocopy;
tocopy = 5 - tocopy;
offset = 0;
continue;
}
keyoffset = 0;
tocopy = 5;
for(int y=0;y<16;y++)
{
pwarray[offset+y]=docid[y];
}
offset += 16;
}
pwarray[16] = (byte) 0x80;
for(int i=0;i<47;i++)
{
pwarray[17+i]=0;
}
pwarray[56] = (byte)0x80;
pwarray[57] = (byte)0x0A;
valContext.md5Update(pwarray,64);
valContext.getMD5StoreDigest(valContext);
tool.makekey (0, key,valContext);
tool.rc4 (salt, 16, key);
tool.rc4 (hashedsalt, 16, key);
salt[16] = (byte) 0x80;
for(int i=0;i<47;i++) salt[17+i]=0;
salt[56] = (byte) 0x80;
md51.md5Init();
md51.md5Update(salt, 64);
md51.getMD5StoreDigest(md51);
for(int i=0;i<16;i++)
{
if(hashedsalt[i]!=md51.digest[i]) return false;
}
return true;
}
public void preparekey(byte[] key_data_ptr, int key_data_len, HWPFRC4 key)
{
int index1;
int index2;
byte []state=new byte[256];
int counter;
state = key.state;
for (counter = 0; counter < 256; counter++) state[counter] = (byte)counter;
key.x = 0;
key.y = 0;
index1 = 0;
index2 = 0;
for (counter = 0; counter < 256; counter++)
{
index2 = ((key_data_ptr[index1]&0xff) + (state[counter]&0xff )+ index2 ) &0xff;
byte btemp=state[counter];
state[counter]=state[index2];
state[index2]=btemp;
index1 = ((index1 + 1) % key_data_len);
}
}
public void makekey(int block,HWPFRC4 rc4key,HWMPFEMD5 md5)
{
byte[]pwarray=new byte[64];
HWMPFEMD5 temp=new HWMPFEMD5();
for(int i=0;i<64;i++) pwarray[i]=0;
for(int i=0;i<5;i++)
{
pwarray[i]=md5.digest[i];
}
pwarray[5] = (byte) (block & 0xFF);
pwarray[6] = (byte) ((block >> 8) & 0xFF);
pwarray[7] = (byte) ((block >> 16) & 0xFF);
pwarray[8] = (byte) ((block >> 24) & 0xFF);
pwarray[9] = (byte)0x80;
pwarray[56] =(byte)0x48;
temp.md5Init();
temp.md5Update(pwarray, 64);
temp.getMD5StoreDigest(temp);
preparekey(temp.digest, 16, rc4key);
}
void rc4 ( byte[] buffer_ptr, int buffer_len, HWPFRC4 key)
{
int x;
int y;
byte []state=new byte[256];
int xorIndex;
int counter;
x = key.x;
y = key.y;
state = key.state;
for (counter = 0; counter < buffer_len; counter++)
{
x = ((x + 1) & 0xff);
y = (((state[x]&0xff) + y) & 0xff);
byte btemp=state[x];
state[x]=state[y];
state[y]=btemp;
xorIndex = (((state[x]&0xff )+ (state[y]&0xff)) & 0xff);
buffer_ptr[counter]^=(state[xorIndex]);
}
key.x = x;
key.y = y;
}這裏代碼是按照文檔裏面寫的doc文件格式，先對密碼進行unicde解碼，然後取出x，docid，salt，hashedsalt，調用verifypwd進行驗證，密碼正確後會返回HWMPFEMD5對象，然後對_tableStream的每512字節進行RC4操作，在對_mainStream進行RC4操作，然後
_fib = new FileInformationBlock(_mainStream);
_fib.fillVariableFields(_mainStream, _tableStream);
這裏存儲文件頭部信息，然後在對_dataStream進行解碼，這裏就是簡單的針對512字節進行RC4操作，這裏只是對文檔的簡單說明。
3；Java調用方式
POIFSFileSystem pfs;
HWPFDocument hwpf =null;
try {
pfs = new POIFSFileSystem(new FileInputStream("./test/20030523jm.doc"));
hwpf = new HWPFDocument(pfs,"111111");
} catch ( Exception e) {
}
Range range = hwpf.getOverallRange();
TableIterator it = new TableIterator(range);4：githup鏈接
https://github.com/DusonWang/word95-03parse.git