-
由胡正开发的星际译王是Linux平台上很强大的一个开源的翻译软件（也有Windows版本的）支持多种词库、多种语言版本。尤其词库设计比较合理。之前看到一篇博文《星际译王词库应用-自制英汉词典》中用简短的程序就实现了词典的基本功能，不过那个是Linux 下的C/C++版本的，于是决定参考移植一个JAVA版本。
import java.io.ByteArrayOutputStream;
import java.io.IOException;
import java.io.InputStream;
import java.io.*;
/**
* {@docRoot}
*
Java版词典测试版，可以在控制台下输入要查询的单词，回车后会给出单词在词典中的释义
* 词典采用星际译王的词典，本程序主要针对英汉词典
* @author menglongbor
* @updateDate 2012-12-31
* @version v1.0.0
* 相关参考链接：
* http://blog.chinaunix.net/uid-20454005-id-1675913.
html* http://hi.baidu.com/sean_zhu_xiang/item/1581342f88be430e73863eee
* http://blog.csdn.net/ranxiedao/article/details/7787342
* http://www.stardict.cn/
* http://www.huzheng.org/
* http://code.google.com/p/stardict-3/downloads/list
*/
public class testdict
{
final static intMAX_WORD= 256;// 最长输入单词字符数
final static intMAX_KEYS= 27;// 26个字母+"-"开头的后缀
final static intSIZEINT= 4;
final static StringKEY[]= {// 26个字母索引+"-"开头的后缀，不区分大小写
"A", "b", "c", "d", "e", "f",
"g", "h", "i", "j", "k", "l", "m", "n", "o", "p", "q", "r", "s",
"t", "u", "v", "w", "x", "y", "z", "-" };
public static InputStreamisidx= null;// 读取idx文件时所要的流
public static InputStreamisdict= null;// 读取dict文件时所要的流
public static longSTREAM_LOCAL= 0;// 记录单词索引在文件流中的位置
public static StringidxfileString= "oxford-gb.idx";// idx文件路径
public static StringdictfileString= "oxford-gb.dict";// dict文件路径
* 从idx文件中获取当前目标单词
* @param word_buf 保存的是c/c++字符串数组转换为JAVA字符串
* @param data_poffset 用来保存单词的data偏移位置信息
* @param data_plength 用来保存单词的data长度信息
* @param len
* @return
*/
public static boolean get_word(String[] word_buf, int[] data_poffset,
int[] data_plength, int[] len)
// int len = 0;
boolean flag = true;
len[0] = 0;
int index = -1;
byte wd[] = new byte[MAX_WORD];
int value = 0;
try
// 读取单词，对每个字母开头的单词都进行搜索，最多考虑256个字符的单词，
// 读到单词结束符\0时赋值表达式的值就不满足while条件而退出
while (true)
index = isidx.read();
STREAM_LOCAL++;// 每读取一次，位置标识加一以记录下单词在文件流中的起始位置
if (index == -1)
// isidx.reset();
flag = false;
break;
}
if ((index != 0) && (len[0] < MAX_WORD))
wd[len[0]] = (byte) index;// 将int转换为byte
len[0]++;
} else
// 转换为JAVA字符串
// 此处不用再需要像c/c++那样去掉了最后那个结束符了
byte wd2[] = new byte[len[0]];
for (int i = 0; i < len[0]; i++)
wd2[i] = wd[i];
word_buf[0] = new String(wd2);
// System.out.println("get_word:"+word_buf[0]+" len:"+len[0]);
// wd = null;// 释放内存
// wd2 = null;
// 读取偏移量值
for (int i = 0; i < SIZEINT; i++)
// 将4个byte转换为int
int shift = (4 - 1 - i) * 8;
return flag;
value += (index & 0x00FF) << shift;
data_poffset[0] = value;
// 读取区块大小值
value = 0;
data_plength[0] = value;
catch (Exception e)
System.out.println("idx file read error!");
// System.out.println("Now local is:"+STREAM_LOCAL);
// 得到单词字符长度
* 通过偏移位置offset和长度length 来从dict文件中获取data内容UTF-8编码的字符
* @param offset 要读取的内容的起始偏移，为字节数
* @param length 要读取的内容的数据块大小，为字节数
* @return 字节数组的data int
public static byte[] get_data(int[] offset, int[] length)
long oft = offset[0];
long len = length[0];
long skip;
byte data_buf[] = new byte[length[0]];
System.out.println("This word's" + "offset:" + offset[0] + "len:"
+ length[0]);
isdict.reset();
long valuedata = isdict.available();
if (valuedata < oft + len)
System.out.println("No so much value data! " + valuedata);
// skip=isdict.skip(oft);
skip = skipBytesFromStream(isdict, oft);
if (skip != oft)
System.out.println("Skip" + skip + " dict file error!");
if (isdict.read(data_buf) == -1)
System.out.println("Arrive at the end of file!");
// // Unicode
// StringBuffer sb = new StringBuffer();
//
// int size =isdict.read(data_buf);
// for (int j = 0; j < size;)
// {
// int l = data_buf[j++];
// int h = data_buf[j++];
// char c = (char) ((l & 0xff) | ((h << 8) & 0xff00));
// sb.append(c);
// }
// // return sb.toString();
data_buf = null;
System.out.println("dict file read error!");
e.printStackTrace();
if (data_buf == null)
return null;
return data_buf;
* utf8解码 参考自http://hi.baidu.com/leo10086/item/d6853813373b19001994ec24 用法:
* 假如 newContent 为UTF8编码的字符串 byte[] b = newContent.getBytes(); newContent =
* URLEncoder.UTF8Decode( b, 0, b.length );
* @param in 要进行解码的UTF8编码的字节数组
* @param offset
* @param length
public static String UTF8Decode(byte in[], int offset, int length)
StringBuffer buff = new StringBuffer();
int max = offset + length;
for (int i = offset; i < max; i++)
char c = 0;
if ((in[i] & 0x80) == 0)
c = (char) in[i];
} else if ((in[i] & 0xe0) == 0xc0) // 11100000
c |= ((in[i] & 0x1f) << 6); // 00011111
i++;
c |= ((in[i] & 0x3f) << 0); // 00111111
} else if ((in[i] & 0xf0) == 0xe0) // 11110000
c |= ((in[i] & 0x0f) << 12); // 00001111
c |= ((in[i] & 0x3f) << 6); // 00111111
} else if ((in[i] & 0xf8) == 0xf0) // 11111000
c |= ((in[i] & 0x07) << 18); // 00000111 (move 18, not 16?)
c |= ((in[i] & 0x3f) << 12); // 00111111
c = ' ';
buff.append(c);
return buff.toString();
public static byte[] UTF8Encode(String str)
ByteArrayOutputStream bos = new ByteArrayOutputStream();
int strlen = str.length();
for (int i = 0; i < strlen; i++)
char t = str.charAt(i);
int c = 0;
c |= (t & 0xffff);
if (c >= 0 && c < 0x80)
bos.write((byte) (c & 0xff));
} else if (c > 0x7f && c < 0x800)
bos.write((byte) (((c >>> 6) & 0x1f) | 0xc0));
bos.write((byte) (((c >>> 0) & 0x3f) | 0x80));
} else if (c > 0x7ff && c < 0x10000)
bos.write((byte) (((c >>> 12) & 0x0f) | 0xe0)); // <--
// correction
// (mb)
bos.write((byte) (((c >>> 6) & 0x3f) | 0x80));
} else if (c > 0x00ffff && c < 0xfffff)
bos.write((byte) (((c >>> 18) & 0x07) | 0xf0));
bos.write((byte) (((c >>> 12) & 0x3f) | 0x80));
bos.flush();
return bos.toByteArray();
* 将UTF-8字节数据转化为Unicode字符串
*
* @param utf_data
* byte[] - UTF-8编码字节数组
* int - 字节数组长度
* @return String - 变换后的Unicode编码字符串
public static String UTF2Uni(byte[] utf_data, int len)
StringBuffer unis = new StringBuffer();
char unic = 0;
int ptr = 0;
int cntBits = 0;
for (; ptr < len;)
cntBits = getCntBits(utf_data[ptr]);
if (cntBits == -1)
++ptr;
continue;
} else if (cntBits == 0)
unic = UTFC2UniC(utf_data, ptr, cntBits);
ptr += cntBits;
unis.append(unic);
return unis.toString();
* 将指定的UTF-8字节组合成一个Unicode编码字符
* @param utf byte[] - UTF-8字节数组
* @param sptr int - 编码字节起始位置
* @param cntBits int - 编码字节数
* @return char - 变换后的Unicode字符
public static char UTFC2UniC(byte[] utf, int sptr, int cntBits)
/*
* Unicode <-> UTF-8 U-00000000 - U-0000007F: 0xxxxxxx U-00000080 -
* U-000007FF: 110xxxxx 10xxxxxx U-00000800 - U-0000FFFF: 1110xxxx
* 10xxxxxx 10xxxxxx U-00010000 - U-001FFFFF: 11110xxx 10xxxxxx 10xxxxxx
* 10xxxxxx U-00200000 - U-03FFFFFF: 111110xx 10xxxxxx 10xxxxxx 10xxxxxx
* 10xxxxxx U-04000000 - U-7FFFFFFF: 1111110x 10xxxxxx 10xxxxxx 10xxxxxx
* 10xxxxxx 10xxxxxx
int uniC = 0; // represent the unicode char
byte firstByte = utf[sptr];
int ptr = 0; // pointer 0 ~ 15
// resolve single byte UTF-8 encoding char
if (cntBits == 0)
return (char) firstByte;
// resolve the first byte
firstByte &= (1 << (7 - cntBits)) - 1;
// resolve multiple bytes UTF-8 encoding char(except the first byte)
for (int i = sptr + cntBits - 1; i > sptr; --i)
byte utfb = utf[i];
uniC |= (utfb & 0x3f) << ptr;
ptr += 6;
uniC |= firstByte << ptr;
return (char) uniC;
* 根据给定字节计算UTF-8编码的一个字符所占字节数，UTF-8规则定义，字节标记只能为0或2~6
* @param b
private static int getCntBits(byte b)
int cnt = 0;
if (b == 0)
return -1;
for (int i = 7; i >= 0; --i)
if (((b >> i) & 0x1) == 1)
++cnt;
else
return (cnt > 6 || cnt == 1) ? -1 : cnt;
* 显示data内容
* @param data_buf UTF-8的单词释义数组
* @param data_length UTF-8的单词释义数组长度
public static void display_data(byte[] data_buf, int data_length[])
// 将UTF-8byte字节数组转为当前环境字符并显示
// String tempString = UTF8Decode(data_buf, 0, data_length[0]);
String tempString = UTF2Uni(data_buf, data_length[0]);
// String tempString = new String(data_buf);
System.out.println(tempString);
* 从idx文件中搜索由word指定的单词，并保存相应的偏移和长度信息
* @param word
* @param data_poffset
* @param data_plength
* @return 是否搜索成功
public static boolean search_word(String word, int[] data_poffset,
int[] data_plength)
String wd[] = new String[1];
boolean temp = false;
int len[] = new int[1];
// 从idx文件中获取当前目标单词
// for (get_word(wd, data_poffset, data_plength); end; get_word(wd,
// data_poffset, data_plength))
while (get_word(wd, data_poffset, data_plength, len))
// System.out.println("compared_word:"+wd[0]);
// if (wd[0].compareToIgnoreCase(word) == 0) //
// 比较字符串s1和s2，但不区分字母的大小写
if (strsEqualsIgnoreCase(wd[0], word) == 0)
System.out.println("compared_word:" + word + " " + wd[0]);
temp = true;
return temp;
* 从标准输入获取待查询的单词，控制台下为GBK字符，字典索引中的英文单词字母也是如此
* @param max_len
* @param count
* @return
public static String get_input(int max_len, int[] count)
byte input_buf[] = new byte[max_len];
count[0] = 0;
String tempString[] = new String[1];
count[0] = System.in.read(input_buf) - 2;// 返回实际读取到的字符数，减去2个控制字符
byte temp_buf[] = new byte[count[0]];
for (int i = 0; i < count[0]; i++)
temp_buf[i] = input_buf[i];
tempString[0] = new String(temp_buf);
System.out.println("Input error!");
System.out.println("Your input is:" + tempString[0]);
return tempString[0];
* @param input_buf
public static byte[] get_input(byte[] input_buf, int[] count)
input_buf = null;
return input_buf;
* 缓存KEYS在idx中的偏移信息，以便加快search_word的搜索速度
* @param idx_cache 保存每个单字母单词对应的起始位置
public static void cache_idx(long[] idx_cache)
int i;
long[] p = idx_cache;
int unused1[] = new int[1];
int unused2[] = new int[1];
// 将文件内部的位置指针重新指向一个流（数据流/文件）的开头返回FILE指针当前位置，
// 然后重新遍历整个文件搜寻下一个字母开头的单词
isidx.reset();
STREAM_LOCAL = 0;
for (i = 0; i < MAX_KEYS; i++)
// System.out.println("Start search_word：" + KEY[i]);
if (search_word(KEY[i], unused1, unused2))// 从idx文件中搜索由word指定的单词，并保存相应的偏移和长度信息
p[i] = STREAM_LOCAL; // 返回当前文件位置
// String tempString = Long.toString(STREAM_LOCAL);
// System.out.println(KEY[i] + "'s local is:" + tempString);
System.out.println(KEY[i] + "'s local is:" + STREAM_LOCAL
+ " offset:" + unused1[0] + "length:" + unused2[0]);
p[i] = 0;
// TODO: handle exception
* 定位由word指定的单词在idx文件中的大概偏移位置
* @param idx_cache
public static long locate_idx(String word, long[] idx_cache)
int i = 0;
int pre = 0;
String tempString = word.toLowerCase();
while (i < MAX_KEYS && KEY[i].charAt(0) < tempString.charAt(0))
pre = i;
++i;
if (tempString.charAt(0) == '-')
pre = 0;
System.out.println("Now word's locate is:" + idx_cache[pre]);
return idx_cache[pre];
* 主要查询函数
public static void consult()
byte data[] = null;// 释义数据，UTF-8数据
long idx[] = new long[MAX_KEYS];// 26个字母孤立单词+"-"开头的后缀对应的索引缓冲
int offset[] = new int[1];
int length[] = new int[1];
System.out.println("Start cache_idx....！");
System.out.println("Open files....!");
// 读取字典索引文件
isidx = new BufferedInputStream(new FileInputStream(
idxfileString));
isidx.mark(isidx.available() + 1);
if (!isidx.markSupported())
System.out.println("This stream do not support mark....!");
System.out.println("Open files error!");
cache_idx(idx);// 缓存KEYS在idx中的偏移信息，以便加快search_word的搜索速度
isdict = new BufferedInputStream(new FileInputStream(
dictfileString));
isdict.mark(isdict.available() + 1);
if (!isdict.markSupported())
System.out.println("INPUT A WORD OR PHRASE: ");
int count[] = new int[1];
String word = get_input(MAX_WORD, count);
long skips1, skips2;
if (count[0] > 0)// 从控制台得到输入单词字符
// 从文件开头跳到单词大致索引所在位置
// isidx.mark(0);
skips1 = locate_idx(word, idx);
// skips2 = isidx.skip(skips1);
skips2 = skipBytesFromStream(isidx, skips1);
System.out
.println("skips1:" + skips1 + " skips2:" + skips2);
System.out.println("locate_idx run error");
if (search_word(word, offset, length))
data = get_data(offset, length);
display_data(data, length);
data = null;
System.out.println("SORRY " + word + " CANNOT BE FOUND!\n");
.println("\n----------------------------------------\n\n");
* 不区分大小写比较两个字符串
* @param s1
* @param s2
public static int strsEqualsIgnoreCase(String s1, String s2)
int n1 = s1.length(), n2 = s2.length();
for (int i1 = 0, i2 = 0; i1 < n1 && i2 < n2; i1++, i2++)
char c1 = s1.charAt(i1);
char c2 = s2.charAt(i2);
if (c1 != c2)
// 源字符串全部都转为大写字符串
c1 = Character.toUpperCase(c1);
c2 = Character.toUpperCase(c2);
// 源字符串全部都转为小写字符串
c1 = Character.toLowerCase(c1);
c2 = Character.toLowerCase(c2);
return c1 - c2;
return n1 - n2;// 如果其中一个或者两个String都比较完了还没有同样的char的话，那就return两个String的长度差距
* 重写了Inpustream 中的skip(long n) 方法，将数据流中起始的n 个字节跳过
* 参考：http://blog.csdn.net/ranxiedao/article/details/7787342
* @param inputStream
* @param n
private static long skipBytesFromStream(InputStream inputStream, long n)
long remaining = n; // SKIP_BUFFER_SIZE is used to determine the size of
// skipBuffer
int SKIP_BUFFER_SIZE = 2048; // skipBuffer is initialized in
// skip(long), if needed.
byte[] skipBuffer = null;
int nr = 0;
if (skipBuffer == null)
skipBuffer = new byte[SKIP_BUFFER_SIZE];
byte[] localSkipBuffer = skipBuffer;
if (n <= 0)
return 0;
while (remaining > 0)
nr = inputStream.read(localSkipBuffer, 0, (int) Math.min(
SKIP_BUFFER_SIZE, remaining));
catch (IOException e)
if (nr < 0)
remaining -= nr;
return n - remaining;
* 主函数
* @param args
public static void main(String args[])
consult();
isidx.close();
isdict.close();
System.out.println("Close files error!");
如果要在
windows平台下编译http://blog.chinaunix.net/uid-20454005-id-1675913.html文章中的程序代码最好保存为cpp文件以C++项目编译执行，而且strcasecmp函数应该换为stricmp函数，并且上面作者原来的程序是在linux平台下的，字符编码本身就是UTF8的不需要进行编码转换，但在windows平台下中文为gb232编码，就需要进行编码的转换，下面为需要添加修改上的字符编码转换后的程序。
//UTF-8到GB2312的转换
char* U2G(const char* utf8)
{
int len = MultiByteToWideChar(CP_UTF8, 0, utf8, -1, NULL, 0);
wchar_t* wstr = new wchar_t[len+1];
memset(wstr, 0, len+1);
MultiByteToWideChar(CP_UTF8, 0, utf8, -1, wstr, len);
len = WideCharToMultiByte(CP_ACP, 0, wstr, -1, NULL, 0, NULL, NULL);
char* str = new char[len+1];
memset(str, 0, len+1);
WideCharToMultiByte(CP_ACP, 0, wstr, -1, str, len, NULL, NULL);
if(wstr) delete[] wstr;
return str;
}
//GB2312到UTF-8的转换
char* G2U(const char* gb2312)
int len = MultiByteToWideChar(CP_ACP, 0, gb2312, -1, NULL, 0);
MultiByteToWideChar(CP_ACP, 0, gb2312, -1, wstr, len);
len = WideCharToMultiByte(CP_UTF8, 0, wstr, -1, NULL, 0, NULL, NULL);
WideCharToMultiByte(CP_UTF8, 0, wstr, -1, str, len, NULL, NULL);
/*
* 显示data内容
*/
void display_data(char *data_buf, unsigned int data_length)
fwrite(data_buf,data_length,1,stdout);
char *data=(char *)malloc(data_length);
memcpy(data,data_buf,data_length);
char *p=U2G(data_buf);
printf("%s\n",p);
free(data);
delete p;
以星际译王所支持的牛津英汉词典oxford-gb作为测试，词典格式为UTF8编码的单词字符串，然后是四个字节的int型数据表示该单词在dict释义文件中的起始偏移量，再后四个字节的int型数据表示dict文件中该单词释义总共的长度，如下图所示：
结果显示能够正确得到单词的释义，只是音标未能正确解码，如下图所示：
JAVA版StarDict星际译王简单实现 
JAVA版StarDict星际译王简单实现