/// <summary>
/// 文本相似度比较
/// </summary>
/// <param name="baseVersion"></param>
/// <param name="newVersion"></param>
/// <returns>返回值为 (baseVersion.length + newVersion.length - different.length)/(baseVersion.length + newVersion.length)</returns>
public static decimal Similar(string baseVersion, string newVersion)
{
if (baseVersion == newVersion)
{
return(1);
}
if (string.IsNullOrEmpty(baseVersion) || string.IsNullOrEmpty(newVersion))
{
return(0);
}
int diffLength = 0;
DiffResult result = DiffResult.Diff(baseVersion, newVersion, 1);
List <ChunkResult> chunks = result.ChunkList;
foreach (ChunkResult chunk in chunks)
{
if (!chunk.sameChunk)
{
if (chunk.xString.Length < 1)
{
diffLength += DiffResult.CopyShaow(chunk.yString).Length;
diffLength += chunk.yString.Length;
}
else if (chunk.yString.Length < 1)
{
diffLength += chunk.xString.Length;
diffLength += DiffResult.CopyShaow(chunk.xString).Length;
}
else
{
diffLength += chunk.xString.Length;
diffLength += chunk.yString.Length;
}
}
}
return(Decimal.Divide(baseVersion.Length + newVersion.Length - diffLength, baseVersion.Length + newVersion.Length));
}
/// <summary>
/// 对比两个版本的字符串
/// </summary>
/// <param name="baseVersion">旧版</param>
/// <param name="newVersion">新版</param>
/// <param name="minSegments">最小相同区间偏移量(取值范围 (1-minStringLength))</param>
/// <returns>对比结果对象</returns>
public static unsafe DiffResult Diff(string baseVersion, string newVersion, int minSegments)
{
//比较文件
fixed(char *xptr = baseVersion)
{
fixed(char *yptr = newVersion)
{
//int minSegments = 100;
//x,y 对应的字节流长度
int xlen = baseVersion.Length;
int ylen = newVersion.Length;
//文档最后一次相同的地方
int xdiff = -1, ydiff = -1;
//遍历开始，和结束
int xi = 0, yi = 0;
//回溯索引
int byi, bxi;
//交叉缓冲区索引
int xbuf, ybuf;
//x，y 字节流对应的字节码
char xbyte, ybyte, inBackByte = '\0';
SameChunk root = null, prev = null, curr;
/*
*
* 不匹配缓冲区，用来存储不匹配字节索引
* 大小为：256*4 占用1K内存
* xbuffer[x不匹配字节]=x_index;
* ybuffer[y不匹配字节]=y_index;
*/
int[] xbuffer = new int[UInt16.MaxValue + 1];
int[] ybuffer = new int[UInt16.MaxValue + 1];
//一些临时变量
int xtmp = 0, ytmp = 0, elen = 0;
//下列变量标识当前状态是否为回溯
bool xInback = false;
int xInbackTmp = 0;
bool yInback = false;
int yInBackTmp = 0;
//在回溯失败重检有没有发现同项标识
bool find = false;
//确保不会溢出
while (xi < xlen && yi < ylen)
{
xbyte = xptr[xi];
ybyte = yptr[yi];
LABEL_NEQ:
if (xbyte != ybyte)
{
//记录缓冲区数据
bxi = xbuffer[xbyte];
byi = ybuffer[ybyte];
//交叉索引缓冲区数据，看是否可以回溯
xbuf = xbuffer[ybyte];
ybuf = ybuffer[xbyte];
//检查两个缓冲区，如果都存在，回溯最小的对象
if (xbuf > xdiff)
{
if (ybuf > ydiff)
{
if (ybuf < xbuf)
{
goto LABLE_BACK_Y;
}
}
goto LABLE_BACK_X;
}
if (ybuf > ydiff)
{
goto LABLE_BACK_Y;
}
goto LABLE_REC_BUF;
LABLE_BACK_Y:
//开始按XByte回溯Y
yInBackTmp = yi;
xInbackTmp = xi;
yi = ybuffer[xbyte];
yInback = true;
inBackByte = xbyte;
//避免覆盖重复序列
if (bxi < xdiff)
{
xbuffer[xbyte] = xi;
}
if (byi < ydiff)
{
ybuffer[ybyte] = yi;
}
goto LABLE_LPEQ;
LABLE_BACK_X:
//开始按YByte回溯X
xInbackTmp = xi;
yInBackTmp = yi;
xi = xbuffer[ybyte];
inBackByte = ybyte;
xInback = true;
//避免覆盖重复序列
if (bxi < xdiff)
{
xbuffer[xbyte] = xi;
}
if (byi < ydiff)
{
ybuffer[ybyte] = yi;
}
goto LABLE_LPEQ;
LABLE_REC_BUF:
//避免覆盖重复序列
if (bxi < xdiff)
{
xbuffer[xbyte] = xi;
}
if (byi < ydiff)
{
ybuffer[ybyte] = yi;
}
goto LABLE_LPED;
}
//当前字节码相等
LABLE_LPEQ:
//往上回溯匹配看看是不是有漏网的
xtmp = xi - 1;
ytmp = yi - 1;
while (xtmp > -1 && xtmp > xdiff && ytmp > -1 && ytmp > ydiff)
{
if (xptr[xtmp] != yptr[ytmp])
{
break;
}
xtmp--;
ytmp--;
}
xi = xtmp + 1;
yi = ytmp + 1;
//往下匹配
xtmp = xi + 1;
ytmp = yi + 1;
while (xtmp < xlen && ytmp < ylen)
{
xbyte = xptr[xtmp];
ybyte = yptr[ytmp];
if (xbyte != ybyte)
{
break;
}
xtmp++;
ytmp++;
}
elen = xtmp - xi;
if (elen > minSegments)
{
xdiff = xtmp - 1;
ydiff = ytmp - 1;
//切分相同项
curr = new SameChunk();
curr.xStartIndex = xi;
curr.xEndIndex = xdiff;
curr.yStartIndex = yi;
curr.yEndIndex = ydiff;
if (prev != null)
{
prev.next = curr;
curr.prev = prev;
prev = curr;
}
else
{
prev = curr;
if (root == null)
{
root = prev;
}
}
//InternalAddChuncks(xdoc, xi, xtmp);
xi = xtmp;
yi = ytmp;
xInback = false;
yInback = false;
goto LABEL_NEQ;
}
else
{
//说明太短，不视为是相同区间（重复区间）
if (xInback)
{
//回溯失败，尝试向下找同项覆盖当前失败项
while (xi < xInbackTmp)
{
if (xptr[xi] == inBackByte)
{
xbuffer[inBackByte] = xi;
find = true;
break;
}
xi++;
}
if (!find)
{
xbuffer[inBackByte] = 0;
}
find = false;
xi = xInbackTmp;
yi = yInBackTmp;
xInback = false;
goto LABLE_LPED;
}
if (yInback)
{
while (yi < yInBackTmp)
{
if (yptr[yi] == inBackByte)
{
ybuffer[inBackByte] = yi;
find = true;
break;
}
yi++;
}
if (!find)
{
ybuffer[inBackByte] = 0;
}
find = false;
yi = yInBackTmp;
xi = xInbackTmp;
yInback = false;
goto LABLE_LPED;
}
xi = xtmp;
yi = ytmp;
continue;
}
LABLE_LPED:
xi++;
yi++;
}
DiffResult diff = new DiffResult();
diff.baseVersion = baseVersion;
diff.newVersion = newVersion;
diff.chunksRoot = root;
return(diff);
}
}
}
