tbxi/speedups/fast_lzss.c
2019-05-05 21:06:14 +08:00

371 lines
11 KiB
C

#define PY_SSIZE_T_CLEAN 1
#include <Python.h>
#define LARGE_BUFFER 0x1000000
#include <stdlib.h>
#include <stdint.h>
#include <string.h>
/**************************************************************
LZSS.C -- A Data Compression Program
***************************************************************
4/6/1989 Haruhiko Okumura
Use, distribute, and modify this program freely.
Please send me your improved versions.
PC-VAN SCIENCE
NIFTY-Serve PAF01022
CompuServe 74050,1022
**************************************************************/
#define N 4096 /* size of ring buffer - must be power of 2 */
#define F 18 /* upper limit for match_length */
#define THRESHOLD 2 /* encode string into position and length
if match_length is greater than this */
#define NIL N /* index for root of binary search trees */
struct encode_state {
/*
* left & right children & parent. These constitute binary search trees.
*/
int lchild[N + 1], rchild[N + 257], parent[N + 1];
/* ring buffer of size N, with extra F-1 bytes to aid string comparison */
uint8_t text_buf[N + F - 1];
/*
* match_length of longest match.
* These are set by the insert_node() procedure.
*/
int match_position, match_length;
};
int
decompress_lzss(uint8_t *dst, uint8_t *src, uint32_t srclen)
{
/* ring buffer of size N, with extra F-1 bytes to aid string comparison */
uint8_t text_buf[N + F - 1];
uint8_t *dststart = dst;
uint8_t *srcend = src + srclen;
int i, j, k, r, c;
unsigned int flags;
dst = dststart;
srcend = src + srclen;
for (i = 0; i < N - F; i++)
text_buf[i] = ' ';
r = N - F;
flags = 0;
for ( ; ; ) {
if (((flags >>= 1) & 0x100) == 0) {
if (src < srcend) c = *src++; else break;
flags = c | 0xFF00; /* uses higher byte cleverly */
} /* to count eight */
if (flags & 1) {
if (src < srcend) c = *src++; else break;
*dst++ = c;
text_buf[r++] = c;
r &= (N - 1);
} else {
if (src < srcend) i = *src++; else break;
if (src < srcend) j = *src++; else break;
i |= ((j & 0xF0) << 4);
j = (j & 0x0F) + THRESHOLD;
for (k = 0; k <= j; k++) {
c = text_buf[(i + k) & (N - 1)];
*dst++ = c;
text_buf[r++] = c;
r &= (N - 1);
}
}
}
return dst - dststart;
}
/*
* initialize state, mostly the trees
*
* For i = 0 to N - 1, rchild[i] and lchild[i] will be the right and left
* children of node i. These nodes need not be initialized. Also, parent[i]
* is the parent of node i. These are initialized to NIL (= N), which stands
* for 'not used.' For i = 0 to 255, rchild[N + i + 1] is the root of the
* tree for strings that begin with character i. These are initialized to NIL.
* Note there are 256 trees. */
static void init_state(struct encode_state *sp)
{
int i;
memset(sp, 0, sizeof(*sp));
for (i = 0; i < N - F; i++)
sp->text_buf[i] = ' ';
for (i = N + 1; i <= N + 256; i++)
sp->rchild[i] = NIL;
for (i = 0; i < N; i++)
sp->parent[i] = NIL;
}
/*
* Inserts string of length F, text_buf[r..r+F-1], into one of the trees
* (text_buf[r]'th tree) and returns the longest-match position and length
* via the global variables match_position and match_length.
* If match_length = F, then removes the old node in favor of the new one,
* because the old one will be deleted sooner. Note r plays double role,
* as tree node and position in buffer.
*/
static void insert_node(struct encode_state *sp, int r)
{
int i, p, cmp;
uint8_t *key;
cmp = 1;
key = &sp->text_buf[r];
p = N + 1 + key[0];
sp->rchild[r] = sp->lchild[r] = NIL;
sp->match_length = 0;
for ( ; ; ) {
if (cmp >= 0) {
if (sp->rchild[p] != NIL)
p = sp->rchild[p];
else {
sp->rchild[p] = r;
sp->parent[r] = p;
return;
}
} else {
if (sp->lchild[p] != NIL)
p = sp->lchild[p];
else {
sp->lchild[p] = r;
sp->parent[r] = p;
return;
}
}
for (i = 1; i < F; i++) {
if ((cmp = key[i] - sp->text_buf[p + i]) != 0)
break;
}
if (i > sp->match_length) {
sp->match_position = p;
if ((sp->match_length = i) >= F)
break;
}
}
sp->parent[r] = sp->parent[p];
sp->lchild[r] = sp->lchild[p];
sp->rchild[r] = sp->rchild[p];
sp->parent[sp->lchild[p]] = r;
sp->parent[sp->rchild[p]] = r;
if (sp->rchild[sp->parent[p]] == p)
sp->rchild[sp->parent[p]] = r;
else
sp->lchild[sp->parent[p]] = r;
sp->parent[p] = NIL; /* remove p */
}
/* deletes node p from tree */
static void delete_node(struct encode_state *sp, int p)
{
int q;
if (sp->parent[p] == NIL)
return; /* not in tree */
if (sp->rchild[p] == NIL)
q = sp->lchild[p];
else if (sp->lchild[p] == NIL)
q = sp->rchild[p];
else {
q = sp->lchild[p];
if (sp->rchild[q] != NIL) {
do {
q = sp->rchild[q];
} while (sp->rchild[q] != NIL);
sp->rchild[sp->parent[q]] = sp->lchild[q];
sp->parent[sp->lchild[q]] = sp->parent[q];
sp->lchild[q] = sp->lchild[p];
sp->parent[sp->lchild[p]] = q;
}
sp->rchild[q] = sp->rchild[p];
sp->parent[sp->rchild[p]] = q;
}
sp->parent[q] = sp->parent[p];
if (sp->rchild[sp->parent[p]] == p)
sp->rchild[sp->parent[p]] = q;
else
sp->lchild[sp->parent[p]] = q;
sp->parent[p] = NIL;
}
uint8_t *
compress_lzss(uint8_t *dst, size_t dstlen, uint8_t *src, size_t srcLen)
{
/* Encoding state, mostly tree but some current match stuff */
struct encode_state *sp;
int i, c, len, r, s, last_match_length, code_buf_ptr;
uint8_t code_buf[17], mask;
uint8_t *srcend = src + srcLen;
uint8_t *dstend = dst + dstlen;
/* initialize trees */
sp = (struct encode_state *) malloc(sizeof(*sp));
init_state(sp);
/*
* code_buf[1..16] saves eight units of code, and code_buf[0] works
* as eight flags, "1" representing that the unit is an unencoded
* letter (1 byte), "" a position-and-length pair (2 bytes).
* Thus, eight units require at most 16 bytes of code.
*/
code_buf[0] = 0;
code_buf_ptr = mask = 1;
/* Clear the buffer with any character that will appear often. */
s = 0; r = N - F;
/* Read F bytes into the last F bytes of the buffer */
for (len = 0; len < F && src < srcend; len++)
sp->text_buf[r + len] = *src++;
if (!len) {
free(sp);
return (void *) 0; /* text of size zero */
}
/*
* Insert the F strings, each of which begins with one or more
* 'space' characters. Note the order in which these strings are
* inserted. This way, degenerate trees will be less likely to occur.
*/
for (i = 1; i <= F; i++)
insert_node(sp, r - i);
/*
* Finally, insert the whole string just read.
* The global variables match_length and match_position are set.
*/
insert_node(sp, r);
do {
/* match_length may be spuriously long near the end of text. */
if (sp->match_length > len)
sp->match_length = len;
if (sp->match_length <= THRESHOLD) {
sp->match_length = 1; /* Not long enough match. Send one byte. */
code_buf[0] |= mask; /* 'send one byte' flag */
code_buf[code_buf_ptr++] = sp->text_buf[r]; /* Send uncoded. */
} else {
/* Send position and length pair. Note match_length > THRESHOLD. */
code_buf[code_buf_ptr++] = (uint8_t) sp->match_position;
code_buf[code_buf_ptr++] = (uint8_t)
( ((sp->match_position >> 4) & 0xF0)
| (sp->match_length - (THRESHOLD + 1)) );
}
if ((mask <<= 1) == 0) { /* Shift mask left one bit. */
/* Send at most 8 units of code together */
for (i = 0; i < code_buf_ptr; i++)
if (dst < dstend)
*dst++ = code_buf[i];
else {
free(sp);
return (void *) 0;
}
code_buf[0] = 0;
code_buf_ptr = mask = 1;
}
last_match_length = sp->match_length;
for (i = 0; i < last_match_length && src < srcend; i++) {
delete_node(sp, s); /* Delete old strings and */
c = *src++;
sp->text_buf[s] = c; /* read new bytes */
/*
* If the position is near the end of buffer, extend the buffer
* to make string comparison easier.
*/
if (s < F - 1)
sp->text_buf[s + N] = c;
/* Since this is a ring buffer, increment the position modulo N. */
s = (s + 1) & (N - 1);
r = (r + 1) & (N - 1);
/* Register the string in text_buf[r..r+F-1] */
insert_node(sp, r);
}
while (i++ < last_match_length) {
delete_node(sp, s);
/* After the end of text, no need to read, */
s = (s + 1) & (N - 1);
r = (r + 1) & (N - 1);
/* but buffer may not be empty. */
if (--len)
insert_node(sp, r);
}
} while (len > 0); /* until length of string to be processed is zero */
if (code_buf_ptr > 1) { /* Send remaining code. */
for (i = 0; i < code_buf_ptr; i++)
if (dst < dstend)
*dst++ = code_buf[i];
else {
free(sp);
return (void *) 0;
}
}
free(sp);
return dst;
}
/* Python wrapper stuff happens here */
static PyObject *wrap_compress(PyObject *self, PyObject *args)
{
uint8_t *src, *dst, *returned;
Py_ssize_t src_len, dst_len;
if(!PyArg_ParseTuple(args, "y#", &src, &src_len)) {
PyErr_SetString(PyExc_ValueError, "bad args"); return NULL;
}
/* Now, we guess how long the object goes (naughty!) */
dst = malloc(LARGE_BUFFER);
if (dst == NULL) {
return PyErr_NoMemory();
}
returned = compress_lzss(dst, (size_t)LARGE_BUFFER, src, (size_t)src_len);
if (returned == NULL) {
free(dst);
return PyErr_NoMemory();
}
dst_len = returned - dst;
PyObject *retval = PyBytes_FromStringAndSize((const char *)dst, dst_len);
free(dst);
return retval;
}
static PyMethodDef module_methods[] = {
{"compress", wrap_compress, METH_VARARGS, NULL},
{NULL, NULL, 0, NULL}
};
static struct PyModuleDef this_module = {
PyModuleDef_HEAD_INIT,
"fast_lzss",
"Fast Lempel-Ziv-Storer-Szymanski compression",
-1,
module_methods
};
PyMODINIT_FUNC PyInit_fast_lzss(void)
{
return PyModule_Create(&this_module);
}