lzari.cpp

这是一个基于LZARI算法的数据压缩的类.Haruhiko Okumura 于1989年7月4日用C语言写实现了这个算法.但是上面用到了一些全局或静态的变量,在MFC下用起来很不方便.作者把它改写成了

{
	int i, j, k;
	
	i = 1;  j = N;
	while (i < j)
	{
		k = (i + j) / 2;
		if (position_cum[k] > x) i = k + 1;  else j = k;
	}
	return i - 1;
}

void LZARI::StartDecode(void)
{
	int i;

	for (i = 0; i < M + 2; i++)
		value = 2 * value + GetBit();
}

int LZARI::DecodeChar(void)
{
	int	 sym, ch;
	unsigned long int  range;
	
	range = high - low;
	sym = BinarySearchSym((unsigned int)
		(((value - low + 1) * sym_cum[0] - 1) / range));
	high = low + (range * sym_cum[sym - 1]) / sym_cum[0];
	low +=       (range * sym_cum[sym    ]) / sym_cum[0];
	for ( ; ; ) {
		if (low >= Q2) {
			value -= Q2;  low -= Q2;  high -= Q2;
		} else if (low >= Q1 && high <= Q3) {
			value -= Q1;  low -= Q1;  high -= Q1;
		} else if (high > Q2) break;
		low += low;  high += high;
		value = 2 * value + GetBit();
	}
	ch = sym_to_char[sym];
	UpdateModel(sym);
	return ch;
}

int LZARI::DecodePosition(void)
{
	int position;
	unsigned long int  range;
	
	range = high - low;
	position = BinarySearchPos((unsigned int)
		(((value - low + 1) * position_cum[0] - 1) / range));
	high = low + (range * position_cum[position    ]) / position_cum[0];
	low +=       (range * position_cum[position + 1]) / position_cum[0];
	for ( ; ; ) {
		if (low >= Q2) {
			value -= Q2;  low -= Q2;  high -= Q2;
		} else if (low >= Q1 && high <= Q3) {
			value -= Q1;  low -= Q1;  high -= Q1;
		} else if (high > Q2) break;
		low += low;  high += high;
		value = 2 * value + GetBit();
	}
	return position;
}

/********** Encode and Decode **********/

void LZARI::Encode(void)
{
	int  i, c, len, r, s, last_match_length;
	
	if(!m_bMem)
	{
		fseek(infile, 0L, SEEK_END);
		textsize = ftell(infile);
		if (fwrite(&textsize, sizeof textsize, 1, outfile) < 1)
			Error("Write Error");  /* output size of text */
		codesize += sizeof textsize;
		if (textsize == 0) return;
		rewind(infile);
		textsize = 0;
	}
	else
	{
		textsize = m_nInLength;
		m_OutBuffer.resize(sizeof textsize);
		memcpy(&m_OutBuffer[0],&textsize,sizeof textsize);
		//m_nOutCur += sizeof textsize;
		codesize += sizeof textsize;
		if(textsize == 0) return;
		m_nInCur = 0;
		textsize = 0;
	}
	
	StartModel(); 
	InitTree();
	s = 0;  r = N - F;
	for (i = s; i < r; i++) text_buf[i] = ' ';
	if(!m_bMem)
		for (len = 0; len < F && (c = getc(infile)) != EOF; len++) text_buf[r + len] = c;
	else
		for (len = 0; len < F && m_nInCur < m_nInLength ; len++)
		{
			c = m_pInBuffer[m_nInCur++];
			text_buf[r + len] = c;
		}
	
	textsize = len;
	for (i = 1; i <= F; i++) InsertNode(r - i);
	
	InsertNode(r);
	
	do {
		if (match_length > len) match_length = len;
		if (match_length <= THRESHOLD) 
		{
			match_length = 1;  EncodeChar(text_buf[r]);
		} 
		else 
		{
			EncodeChar(255 - THRESHOLD + match_length);
			EncodePosition(match_position - 1);
		}
		last_match_length = match_length;
		if(!m_bMem)
		{
			for (i = 0; i < last_match_length && (c = getc(infile)) != EOF; i++) 
			{
				DeleteNode(s);  text_buf[s] = c;
				if (s < F - 1) text_buf[s + N] = c;
				s = (s + 1) & (N - 1);
				r = (r + 1) & (N - 1);
				InsertNode(r);
			}
		}
		else
		{			
			for (i = 0; i < last_match_length && m_nInCur < m_nInLength ; i++) 
			{
				c = m_pInBuffer[m_nInCur++];
				DeleteNode(s);  
				text_buf[s] = c;
				if (s < F - 1) text_buf[s + N] = c;
				s = (s + 1) & (N - 1);
				r = (r + 1) & (N - 1);
				InsertNode(r);
			}
		}
		if ((textsize += i) > printcount) 
		{
#ifdef _OUTPUT_STATUS
			printf("%12ld\r", textsize);
#endif	
			printcount += 1024;
		}
		while (i++ < last_match_length) 
		{
			DeleteNode(s);
			s = (s + 1) & (N - 1);
			r = (r + 1) & (N - 1);
			if (--len) InsertNode(r);
		}
	} while (len > 0);
	
	EncodeEnd();

#ifdef _OUTPUT_STATUS	
	printf("In : %lu bytes\n", textsize);
	printf("Out: %lu bytes\n", codesize);
	printf("Out/In: %.3f\n", (double)codesize / textsize);
#endif
}

void LZARI::Decode(void)
{
	int  i, j, k, r, c;
	unsigned long int  count;

	if (!m_bMem)
	{
		if (fread(&textsize, sizeof textsize, 1, infile) < 1)
			Error("Read Error");  /* read size of text */
	}
	else
	{
		if(m_nInLength < sizeof textsize)
			Error("Read Error");
		memcpy(&textsize,m_pInBuffer + m_nInCur,sizeof textsize);
		
		//m_OutBuffer.reserve(textsize);
		m_nOutLength = textsize;
		//m_nOutCur = 0;
		
		m_nInCur += sizeof textsize;
	}
	
	if (textsize == 0) return;
	
	StartDecode();
	StartModel();
	
	for (i = 0; i < N - F; i++) text_buf[i] = ' ';
	r = N - F;
	for (count = 0; count < textsize; ) 
	{
		c = DecodeChar();
		if (c < 256) 
		{
			if(!m_bMem)
				putc(c, outfile);
			else
			{
				//m_OutBuffer[m_nOutCur++] = c;
				m_OutBuffer.push_back(c);
				//m_nOutCur++;
			}
			text_buf[r++] = c;
			r &= (N - 1);
			count++;
		}
		else
		{
			i = (r - DecodePosition() - 1) & (N - 1);
			j = c - 255 + THRESHOLD;
			for (k = 0; k < j; k++) 
			{
				c = text_buf[(i + k) & (N - 1)];
				if(!m_bMem)
					putc(c, outfile);
				else
				{
				//	m_pOutBuffer[m_nOutCur++] = c;
					m_OutBuffer.push_back(c);
					//m_nOutCur ++;
				}
				text_buf[r++] = c;
				r &= (N - 1);
				count++;
			}
		}
		if (count > printcount) 
		{
#ifdef _OUTPUT_STATUS
			printf("%12lu\r", count);
#endif			
			printcount += 1024;
		}
	}

#ifdef _OUTPUT_STATUS
	printf("%12lu\n", count);
#endif
}

void LZARI::Compress(const char *lpszInfile,const char *lpszOutfile)
{
	m_bMem = FALSE;
	
	infile = fopen(lpszInfile,"rb");
	outfile = fopen(lpszOutfile,"wb");
	if(infile && outfile)
	{
		Encode();
		fclose(infile);
		fclose(outfile);
		infile = NULL;
		outfile = NULL;
	}
}

void LZARI::UnCompress(const char *lpszInfile,const char *lpszOutfile)
{
	m_bMem = FALSE;

	infile = fopen(lpszInfile,"rb");
	outfile = fopen(lpszOutfile,"wb");
	if(infile && outfile)
	{
		Decode();
		fclose(infile);
		fclose(outfile);
		infile = NULL;
		outfile = NULL;
	}
}

void LZARI::Compress(const BYTE *pInBuffer,int nInLength,const BYTE *&pOutBuffer ,int &nOutLength)
{
	m_pInBuffer = pInBuffer;
	m_nInLength = nInLength;
	m_nInCur = 0;

//	m_nOutCur = 0;

	m_bMem = TRUE;
	Encode();
	pOutBuffer = &m_OutBuffer[0];
	nOutLength = m_OutBuffer.size();
}

void LZARI::UnCompress(const BYTE *pInBuffer,int nInLength,const BYTE *&pOutBuffer ,int &nOutLength)
{
	m_pInBuffer = pInBuffer;
	m_nInLength = nInLength;
	m_nInCur = 0;

	m_bMem = TRUE;
	Decode();
	pOutBuffer = &m_OutBuffer[0];
	nOutLength = m_OutBuffer.size();
	m_OutBuffer.push_back(0);
}

void LZARI::Release()
{
	if(!m_OutBuffer.empty())
	{
		infile = NULL;
		outfile = NULL;
		
		textsize = 0;
		codesize = 0;
		printcount = 0;
		
		low = 0;
		high = Q4;
		value = 0;
		shifts = 0;
		
		m_bMem = FALSE;
		
		m_pInBuffer = NULL;
		m_nInLength = 0;
		m_nInCur = 0;
		
		m_OutBuffer.clear();
		m_nOutLength = 0;

		buffer_putbit = 0;
		mask_putbit = 128;
		
		buffer_getbit = 0;
		mask_getbit = 0;
	}
}