void csa_text_lf(uchar *p,CSA *csa, pos_t s, pos_t t)
{
i64 i;
i64 j,c;
#if DEBUG
if (s < 1 || t > csa->n || s > t) {
printf("csa_text: out of range [%d,%d] n=%d\n",s,t,csa->n);
}
#endif
// lf = (lf_dna *)csa->psi_struc;
i = csa->inverse(csa,(t+1));
for (j=t-s; j>=0; j--) {
// printf("rank %ld\n",i);
c = csa->BW(csa,i);
if (c == -1) {
printf("csa_text_lf: ??? j = %ld i = %ld c = %ld s=%ld t=%ld \n",
j,i,c,s,t);
exit(1);
}
// p[j] = c;
putuint(p, j, c, csa->k2);
i = csa->LF(csa,i);
}
}
static PyObject *sendPacket(PyObject *self, PyObject *args)
{
int goodcn = -1;
packetbuf p(MAXTRANS,ENET_PACKET_FLAG_RELIABLE);
putint(p, N_CLIENT);
std::cout<<"N_CLIENT ";
loopv(server::clients) {
int cn = server::clients[i]->clientnum;
if(goodcn < 0) goodcn = cn; // find a good cn
}
putint(p, goodcn);
std::cout<<goodcn<<" ";
putuint(p, (((unsigned int)PyTuple_Size(args)))*2);
std::cout<<(((unsigned int)PyTuple_Size(args))*2)<<" [ ";
for(Py_ssize_t i=0;i<PyTuple_Size(args);i++)
{
int data=PyInt_AsLong(PyTuple_GetItem(args,i));
std::cout<<data<<" ";
putint(p, data);
}
sendpacket(-1, 1, p.finalize(), -1);
std::cout<<"- sent\n";
Py_INCREF(Py_None);
return Py_None;
}
///////////////////////////////////////////
// i64 csa_substring(uchar *p,CSA *csa,rank_t r,i64 len)
// extracts substring [s,s+len-1] into the buffer p using psi
// and returns its length
// length is shorter than len if (s+len-1 > n)
// where s = SA[r]
///////////////////////////////////////////
i64 csa_substring(uchar *p,CSA *csa,rank_t r,i64 len)
{
i64 i;
i = 0;
while (i < len) {
if (r == 0) {
// *p++ = 0; // $
putuint(p, i, 0, csa->k2);
break;
}
// *p++ = csa->T(csa,r);
putuint(p, i, csa->T(csa,r), csa->k2);
r = csa->psi(csa,r);
i++;
}
return i;
}
i64 csa_substring_lf_naive(uchar *p,CSA *csa,rank_t r,i64 len)
{
i64 i;
int c;
// p += len;
i = 0;
while (i < len) {
// r = csa_BW_LF_by_psi(csa,r, &c);
r = csa->BW_LF(csa,r, &c);
if (c == -1) {
// *--p = 0; // $
putuint(p, len-1-i, 0, csa->k2);
break;
}
// *--p = c;
putuint(p, len-1-i, c, csa->k2);
i++;
}
return i;
}
///////////////////////////////////////////
// i64 csa_substring_lf(uchar *p,CSA *csa,rank_t r,i64 len)
// extracts substring [s-len,s-1] into the buffer p[0,len-1] using BW
// where s = SA[r]
///////////////////////////////////////////
i64 csa_substring_lf(uchar *p,CSA *csa,rank_t r,i64 len)
{
i64 i,r2;
int c;
// p += len;
i = 0;
while (i < len) {
c = csa->BW_rank(csa,r,&r2);
if (c == -1) {
// *--p = 0; // $
putuint(p, len-1-i, 0, csa->k2);
break;
}
// *--p = c;
putuint(p, len-1-i, c, csa->k2);
r = csa->K[csa->CtoA[c]+1]-1 + r2;
i++;
}
return i;
}
///////////////////////////////////////////
// i64 csa_search_r(i64 keylen,int c, CSA *csa,rank_t *li,rank_t *ri)
// returns the length of the longest suffix of key:c
// where key is the substring of length keylen represented by range [*li, *ri]
// returned [*li, *ri] is the range of key:c
///////////////////////////////////////////
i64 csa_search_r(i64 keylen, int c, CSA *csa,rank_t *li,rank_t *ri)
{
uchar *substr;
i64 ret;
substr = mymalloc((keylen+1)*csa->k2);
csa->substring(substr, csa, *li, keylen);
// substr[keylen] = c;
putuint(substr, keylen, c, csa->k2);
ret = csa->search(substr, keylen+1, csa, li, ri);
myfree(substr,(keylen+1)*csa->k2);
return ret;
}
///////////////////////////////////////////
// int csa_child_r(CSA *csa,i64 len,rank_t l,rank_t r,uchar *tail,rank_t *ll,rank_t *rr)
// returns the number of characters following suffixes [l,r]
// tail stores the list of following characters
// ll and rr store ranges for the following characters
// lengths of head, ll, and rr must be at least csa->m (#distinct characters)
///////////////////////////////////////////
int csa_child_r(CSA *csa, i64 len, rank_t l, rank_t r, uchar *tail, rank_t *ll, rank_t *rr)
{
int i,c,num;
int sigma;
uchar *substr;
rank_t *ltmp, *rtmp;
uchar *tailtmp;
int *hc;
sigma = csa->sigma;
substr = mymalloc((len+1)*csa->k2);
csa->substring(substr, csa, l, len);
tailtmp = mymalloc(sigma * sizeof(*tailtmp) * csa->k2);
ltmp = mymalloc(sigma * sizeof(*ltmp));
rtmp = mymalloc(sigma * sizeof(*ltmp));
hc = mymalloc(sigma * sizeof(*hc));
num = csa_child_r_sub(csa, len, l, r, 0, substr, tailtmp, ltmp, rtmp);
for (c=0; c<sigma; c++) hc[c] = -1;
// for (i=0; i<num; i++) hc[tailtmp[i]] = i;
for (i=0; i<num; i++) hc[getuint(tailtmp,i,csa->k2)] = i;
i = 0;
for (c=0; c<sigma; c++) {
if (hc[c] >= 0) {
// tail[i] = tailtmp[hc[c]];
putuint(tail,i, getuint(tailtmp,hc[c],csa->k2), csa->k2);
ll[i] = ltmp[hc[c]];
rr[i] = rtmp[hc[c]];
i++;
}
}
free(hc); free(rtmp); free(ltmp); free(tailtmp);
myfree(substr,(len+1)*csa->k2);
return num;
}
int csa_child_rs(CSA *csa, uchar *substr, i64 len, uchar *tail, rank_t *ll, rank_t *rr)
{
int i,c,num;
rank_t lll, rrr;
i64 ret;
num = 0;
for (i=0; i<csa->m; i++) {
c = csa->AtoC[i];
// substr[len] = c;
putuint(substr, len, c, csa->k2);
ret = csa->search(substr, len+1, csa, &lll, &rrr);
if (ret == len+1) {
tail[num] = c;
ll[num] = lll;
rr[num] = rrr;
num++;
}
}
return num;
}
void csa_text_lf(uchar *p,CSA *csa, pos_t s, pos_t t)
{
i64 i,j;
int c;
#if DEBUG
if (s < 1 || t > csa->n || s > t) {
printf("csa_text: out of range [%d,%d] n=%d\n",s,t,csa->n);
}
#endif
// lf = (lf_dna *)csa->psi_struc;
i = csa->inverse(csa,(t+1));
for (j=t-s; j>=0; j--) {
// printf("rank %ld\n",i);
i = csa->BW_LF(csa,i, &c);
#if DEBUG
if (c == -1) {
printf("csa_text_lf: ??? j = %ld i = %ld c = %ld s=%ld t=%ld \n",
j,i,c,s,t);
exit(1);
}
#endif
// p[j] = c;
#if 0
{
int x,j;
x = c;
printf("(");
for (j=0; j<csa->k2; j++) {
printf("%c", x & 0xff);
x >>= 8;
}
printf(")");
}
#endif
putuint(p, j, c, csa->k2);
}
}
void csa_new_from_bwt(CSA csa, char *fname, char *fidx, int psi_id, int idx_id, bool coded)
{
int k;
i64 i,j,v,m;
FILE *f2;
i64 psize,isize;
i64 n;
psi_id = csa.id;
if (psi_id >= 0) {
printf("create psi: id=%d\n",psi_id);
}
if (idx_id >= 0) {
printf("create idx: id=%d D=%d D2=%d\n",idx_id,csa.D,csa.D2);
}
psize = 0;
if (psi_id >= 0) {
switch (psi_id & 0x3f) {
case ID_DIFF_GAMMA:
case ID_DIFF_GAMMA_RL:
case ID_DIFF_GAMMA_SPARSE:
case ID_DIFF_GAMMA_RL_SPARSE:
psize = psi1_makeindex(&csa, fname);
printf("n %ld\n",csa.n);
printf("Psi %ld bytes (%1.3f bpc)\n",
psize,(double)psize*8/csa.n);
break;
case ID_DIFF_GAMMA_RR:
psize = psi12_makeindex(&csa, fname);
printf("n %ld\n",csa.n);
printf("Psi %ld bytes (%1.3f bpc)\n",
psize,(double)psize*8/csa.n);
break;
case ID_BWT_DNA:
psize = lf_dna_makeindex(&csa, fname, coded);
printf("n %ld\n",csa.n);
printf("BW %ld bytes (%1.3f bpc)\n",
psize,(double)psize*8/csa.n);
break;
case ID_BWT_BIT:
psize = lf_bit_makeindex(&csa, fname);
printf("n %ld\n",csa.n);
printf("BW %ld bytes (%1.3f bpc)\n",
psize,(double)psize*8/csa.n);
break;
case ID_BWT_WT:
case ID_BWT_WT_HUF:
case ID_BWT_WT_DENSE:
case ID_BWT_WT_SPARSE4:
case ID_BWT_WT_RR:
psize = lf_wt_makeindex(&csa, fname);
printf("n %ld\n",csa.n);
printf("BW %ld bytes (%1.3f bpc)\n",
psize,(double)psize*8/csa.n);
break;
#if 0
case ID_BWT_HUF:
psize = lf_bwt_makeindex(&csa, fname);
printf("n %ld\n",csa.n);
printf("BW %ld bytes (%1.3f bpc)\n",
psize,(double)psize*8/csa.n);
break;
#endif
case ID_SPARSE4:
psize = psi2_makeindex(&csa, fname);
printf("n %ld\n",csa.n);
printf("Psi %ld bytes (%1.3f bpc)\n",
psize,(double)psize*8/csa.n);
break;
default:
printf("psi_id = %d\n",psi_id);
exit(1);
}
}
csa.k = (blog(csa.n+1)+1+8-1)/8;
for (i=0; i<SIGMA; i++) csa.CtoA[i] = -1;
// csa.K[-1+1] = 0;
csa.K[-1+1] = 1;
for (m=0,v=1,i=0; i<SIGMA; i++) {
if (csa.C[i]>0) {
csa.AtoC[m] = i;
csa.CtoA[i] = m;
csa.K[m+1] = v;
// printf("i=%ld v = %ld C[i] = %ld\n",i,v,csa.C[i]);
v += csa.C[i];
m++;
}
}
csa.K[m+1] = v;
csa.m = m;
if (csa.D >= csa.n) {
printf("D=%d >= n=%ld\n",csa.D,csa.n);
exit(0);
}
if (csa.D2 >= csa.n) {
printf("D2=%d >= n=%ld\n",csa.D2,csa.n);
exit(0);
}
if (idx_id >= 0) {
n = csa.n;
k = csa.k;
//// compute SA and ISA
if (csa.D > 0) csa.SA = (uchar *) mymalloc(((n-1)/csa.D+1+1)*k);
if (csa.D2 > 0) csa.ISA = (uchar *) mymalloc(((n-1)/csa.D2+1+1)*k);
if (csa.D == 0 && csa.D2 == 0) goto brk;
switch (psi_id & 0x3f) {
case ID_DIFF_GAMMA:
case ID_DIFF_GAMMA_RL:
case ID_DIFF_GAMMA_SPARSE:
case ID_DIFF_GAMMA_RL_SPARSE:
case ID_SPARSE4:
case ID_DIFF_GAMMA_RR:
j = 0;
for (i=0; i<=n; i++) {
display_progressbar("making sa ",i,n);
j = csa.psi(&csa,j);
// sa[j] = i;
if (csa.D > 0 && j % csa.D == 0) {
putuint(csa.SA,j / csa.D,i,k);
}
if (csa.D2 > 0 && i % csa.D2 == 0) {
putuint(csa.ISA,i / csa.D2,j,k);
}
}
// putuint(csa.SA,0,n,k);
break;
case ID_BWT_DNA:
case ID_BWT_BIT:
case ID_BWT_WT:
case ID_BWT_WT_HUF:
case ID_BWT_WT_DENSE:
case ID_BWT_WT_SPARSE4:
case ID_BWT_WT_RR:
case ID_BWT_HUF:
j = 0;
for (i=n-1; i>=0; i--) {
display_progressbar("making sa ",i,n);
v = csa.LF(&csa,j);
// printf("LF[%ld] = %ld\n",j,v);
j = v;
if (csa.D > 0 && j % csa.D == 0) putuint(csa.SA, j/csa.D , i, k);
if (csa.D2 > 0 && i % csa.D2 == 0) putuint(csa.ISA, i/csa.D2, j, k);
}
putuint(csa.SA,0,n,k);
break;
default:
break;
}
brk:
//// write idx
f2 = fopen(fidx,"wb"); /* directory */
if (f2 == NULL) {
perror("csa2_new1: ");
exit(1);
}
isize = 0;
writeint(4,VERSION,f2); /* version */
isize += 4;
writeint(1,ID_HEADER,f2); // header ID
isize += 1;
isize = write_header(&csa, f2, isize);
if (csa.D > 0) {
writeint(1,ID_SA,f2);
isize += 1;
isize = write_sa(&csa, f2, isize);
}
if (csa.D2 > 0) {
writeint(1,ID_ISA,f2);
isize += 1;
isize = write_isa(&csa, f2, isize);
}
fclose(f2);
if (csa.D > 0) free(csa.SA);
if (csa.D2 > 0) free(csa.ISA);
printf("Total %ld bytes (%1.3f bpc)\n",(psize+isize),
(double)(psize+isize)*8/csa.n);
}
free(fidx);
}
static void
bvprintf(struct putcinfo *action, const char *fmt, va_list args)
{
const char *s = fmt;
for (;; s++) {
char c = GET_GLOBAL(*(u8*)s);
if (!c)
break;
if (c != '%') {
putc(action, c);
continue;
}
const char *n = s+1;
int field_width = 0;
char padchar = ' ';
u8 is64 = 0;
for (;;) {
c = GET_GLOBAL(*(u8*)n);
if (!isdigit(c))
break;
if (!field_width && (c == '0'))
padchar = '0';
else
field_width = field_width * 10 + c - '0';
n++;
}
if (c == 'l') {
// Ignore long format indicator
n++;
c = GET_GLOBAL(*(u8*)n);
}
if (c == 'l') {
is64 = 1;
n++;
c = GET_GLOBAL(*(u8*)n);
}
s32 val;
const char *sarg;
switch (c) {
case '%':
putc(action, '%');
break;
case 'd':
val = va_arg(args, s32);
if (is64)
va_arg(args, s32);
if (val < 0) {
putc(action, '-');
val = -val;
}
putuint(action, val);
break;
case 'u':
val = va_arg(args, s32);
if (is64)
va_arg(args, s32);
putuint(action, val);
break;
case 'p':
val = va_arg(args, s32);
putc(action, '0');
putc(action, 'x');
puthex(action, val, 8);
break;
case 'x':
val = va_arg(args, s32);
if (is64) {
u32 upper = va_arg(args, s32);
if (upper) {
putprettyhex(action, upper, field_width - 8, padchar);
puthex(action, val, 8);
break;
}
}
putprettyhex(action, val, field_width, padchar);
break;
case 'c':
val = va_arg(args, int);
putc(action, val);
break;
case '.':
// Hack to support "%.s" - meaning string on stack.
if (GET_GLOBAL(*(u8*)(n+1)) != 's')
break;
n++;
sarg = va_arg(args, const char *);
puts(action, sarg);
break;
case 's':
sarg = va_arg(args, const char *);
puts_cs(action, sarg);
break;
default:
putc(action, '%');
n = s;
}
s = n;
}
}
/**
* @brief This is an minimally useful subset of the POSIX printf() function.
*
* To reduce code size, this pico_printf() implementation only supports a few
* conversion specifiers:
* - @a 'd', @a 'u': for signed and unsigned decimal numbers, respectively;
* - @a 'x', @a 'X': for hexadecimal numbers, downcase and upcase;
* - @a 's': for NULL terminated strings.
*
* Other limitations:
* - No flag specifier is implemented;
* - No field width specifier is implemented;
* - The only supported length modifier is 'l', which is parsed and ignored,
* on supported archictetures 'int' and 'long int' are both 32 bits long.
* - 32 digits maximum length for formatted numbers.
*/
int pico_printf(const char *format, ...)
{
const char *s = format;
int len = 0;
va_list ap;
va_start(ap, format);
while (*s) {
char c = *s++;
if (c == '%') {
c = *s++;
/*
* Ignore 'l' length sub-specifier.
* This has no effect in ILP32 (4/4/4).
* In i586, 'int' and 'long int' are both 4 bytes long.
*/
if (c == 'l') {
c = *s++;
}
switch (c) {
#if (PICO_PRINTF_D)
case 'd': {
int n;
n = va_arg(ap, int);
if (n < 0) {
len += pico_putchar('-');
n = -n;
}
len += putuint(n, 10, false);
break;
}
#endif
#if (PICO_PRINTF_U)
case 'u': {
unsigned int u;
u = va_arg(ap, unsigned int);
len += putuint(u, 10, false);
break;
}
#endif
#if (PICO_PRINTF_X)
case 'X':
case 'x': {
unsigned int u;
u = va_arg(ap, unsigned int);
len += putuint(u, 16, c == 'X');
break;
}
#endif
#if (PICO_PRINTF_S)
case 's': {
const char *str;
str = va_arg(ap, const char *);
len += pico_putchars(str);
break;
}
#endif
#if (PICO_PRINTF_ESC)
default:
len += pico_putchar('%');
case '%':
len += pico_putchar(c);
break;
#endif
}
continue;
}
len += pico_putchar(c);
}
void csa_new_from_bwt(int argc, char *argv[])
{
i64 i,j,v,m;
FILE *f2;
i64 psize,isize;
i64 n;
int k;
char *fname,*fidx;
char *p;
int psi_id, idx_id;
CSA csa;
int sigma;
csa.sigma = 256; /* default alphabet size */
csa.k2 = 1;
// for (i=0; i<SIGMA+2; i++) csa.C[i] = 0;
// for (i=0; i<SIGMA; i++) csa.C[i] = 0;
fname = NULL; fidx = NULL;
psi_id = idx_id = -1;
for (i=1; i<argc; i++) {
p = argv[i];
if (p[0] == '-') {
p++;
switch (toupper(p[0])) {
case 'I':
// -I[n]:[D]:[D2]
p++;
idx_id = 0;
csa_options(&csa, p);
break;
case 'P':
// -P[n]:[L]
p++;
psi_id = 0;
psi_options(&csa, p);
break;
case 'C':
// -C[s]
p++;
sigma_options(&csa, p);
break;
default:
printf("??? no such option %s\n",argv[i]);
exit(1);
}
} else {
fname = argv[i];
k = strlen(fname);
fidx = mymalloc(k+5);
sprintf(fidx,"%s.idx",fname);
}
}
if (fname == NULL) {
printf("no input file.\n");
exit(0);
}
printf("sigma = %d k2 = %d\n", csa.sigma, csa.k2);
sigma = csa.sigma;
csa.C = mymalloc(sizeof(*csa.C)*sigma); //
csa.CtoA = mymalloc(sizeof(*csa.CtoA)*sigma); //
csa.AtoC = mymalloc(sizeof(*csa.AtoC)*sigma); //
csa.K = mymalloc(sizeof(*csa.K)*(sigma+2)); //
for (i=0; i<sigma; i++) csa.C[i] = 0;
psi_id = csa.id;
if (psi_id >= 0) {
printf("create psi: id=%d\n",psi_id);
}
if (idx_id >= 0) {
printf("create idx: id=%d D=%d D2=%d\n",idx_id,csa.D,csa.D2);
}
psize = 0;
if (psi_id >= 0) {
switch (psi_id & 0x3f) {
case ID_DIFF_GAMMA:
case ID_DIFF_GAMMA_RL:
case ID_DIFF_GAMMA_SPARSE:
case ID_DIFF_GAMMA_RL_SPARSE:
psize = psi1_makeindex(&csa, fname);
printf("n %ld\n",csa.n);
printf("Psi %ld bytes (%1.3f bpc)\n",
psize,(double)psize*8/csa.n);
break;
case ID_DIFF_GAMMA_RR:
psize = psi12_makeindex(&csa, fname);
printf("n %ld\n",csa.n);
printf("Psi %ld bytes (%1.3f bpc)\n",
psize,(double)psize*8/csa.n);
break;
case ID_BWT_DNA:
psize = lf_dna_makeindex(&csa, fname);
printf("n %ld\n",csa.n);
printf("BW %ld bytes (%1.3f bpc)\n",
psize,(double)psize*8/csa.n);
break;
case ID_BWT_DNA2:
psize = lf_dna2_makeindex(&csa, fname);
printf("n %ld\n",csa.n);
printf("BW %ld bytes (%1.3f bpc)\n",
psize,(double)psize*8/csa.n);
break;
case ID_BWT_BIT:
psize = lf_bit_makeindex(&csa, fname);
printf("n %ld\n",csa.n);
printf("BW %ld bytes (%1.3f bpc)\n",
psize,(double)psize*8/csa.n);
break;
case ID_BWT_WT:
case ID_BWT_WT_HUF:
case ID_BWT_WT_DENSE:
case ID_BWT_WT_SPARSE4:
case ID_BWT_WT_RR:
psize = lf_wt_makeindex(&csa, fname);
printf("n %ld\n",csa.n);
printf("BW %ld bytes (%1.3f bpc)\n",
psize,(double)psize*8/csa.n);
break;
#if 0
case ID_BWT_HUF:
psize = lf_bwt_makeindex(&csa, fname);
printf("n %ld\n",csa.n);
printf("BW %ld bytes (%1.3f bpc)\n",
psize,(double)psize*8/csa.n);
break;
#endif
case ID_SPARSE4:
psize = psi2_makeindex(&csa, fname);
printf("n %ld\n",csa.n);
printf("Psi %ld bytes (%1.3f bpc)\n",
psize,(double)psize*8/csa.n);
break;
default:
printf("psi_id = %d\n",psi_id);
exit(1);
}
}
csa.k = (blog(csa.n+1)+1+8-1)/8;
for (i=0; i<sigma; i++) csa.CtoA[i] = -1;
csa.K[-1+1] = 1;
for (m=0,v=1,i=0; i<sigma; i++) {
if (csa.C[i]>0) {
csa.AtoC[m] = i;
csa.CtoA[i] = m;
csa.K[m+1] = v;
// printf("i=%ld v = %ld C[i] = %ld\n",i,v,csa.C[i]);
v += csa.C[i];
m++;
}
}
csa.K[m+1] = v;
csa.m = m;
if (csa.D >= csa.n) {
printf("D=%d >= n=%ld\n",csa.D,csa.n);
exit(0);
}
if (csa.D2 >= csa.n) {
printf("D2=%d >= n=%ld\n",csa.D2,csa.n);
exit(0);
}
if (idx_id >= 0) {
n = csa.n;
k = csa.k;
//// compute SA and ISA
if (csa.D > 0) csa.SA = mymalloc(((n-1)/csa.D+1+1)*k);
if (csa.D2 > 0) csa.ISA = mymalloc(((n-1)/csa.D2+1+1)*k);
if (csa.D == 0 && csa.D2 == 0) goto brk;
switch (psi_id & 0x3f) {
case ID_DIFF_GAMMA:
case ID_DIFF_GAMMA_RL:
case ID_DIFF_GAMMA_SPARSE:
case ID_DIFF_GAMMA_RL_SPARSE:
case ID_SPARSE4:
case ID_DIFF_GAMMA_RR:
j = 0;
for (i=0; i<=n; i++) {
display_progressbar("making sa ",i,n);
j = csa.psi(&csa,j);
// sa[j] = i;
if (csa.D > 0 && j % csa.D == 0) {
putuint(csa.SA,j / csa.D,i,k);
}
if (csa.D2 > 0 && i % csa.D2 == 0) {
putuint(csa.ISA,i / csa.D2,j,k);
}
}
// putuint(csa.SA,0,n,k);
break;
case ID_BWT_DNA:
case ID_BWT_DNA2:
case ID_BWT_BIT:
case ID_BWT_WT:
case ID_BWT_WT_HUF:
case ID_BWT_WT_DENSE:
case ID_BWT_WT_SPARSE4:
case ID_BWT_WT_RR:
case ID_BWT_HUF:
j = 0;
for (i=n-1; i>=0; i--) {
display_progressbar("making sa ",i,n);
v = csa.LF(&csa,j);
// printf("LF[%ld] = %ld\n",j,v);
j = v;
if (csa.D > 0 && j % csa.D == 0) putuint(csa.SA, j/csa.D , i, k);
if (csa.D2 > 0 && i % csa.D2 == 0) putuint(csa.ISA, i/csa.D2, j, k);
}
// putuint(csa.SA,0,n,k);
if (csa.D > 0) putuint(csa.SA,0,n,k); // 2011-12-20
break;
default:
break;
}
brk:
//// write idx
f2 = fopen(fidx,"wb"); /* directory */
if (f2 == NULL) {
perror("csa2_new1: ");
exit(1);
}
isize = 0;
writeint(4,VERSION,f2); /* version */
isize += 4;
writeint(1,ID_HEADER,f2); // header ID
isize += 1;
isize = write_header(&csa, f2, isize);
if (csa.D > 0) {
writeint(1,ID_SA,f2);
isize += 1;
isize = write_sa(&csa, f2, isize);
}
if (csa.D2 > 0) {
writeint(1,ID_ISA,f2);
isize += 1;
isize = write_isa(&csa, f2, isize);
}
fclose(f2);
if (csa.D > 0) free(csa.SA);
if (csa.D2 > 0) free(csa.ISA);
printf("Total %ld bytes (%1.3f bpc)\n",(psize+isize),
(double)(psize+isize)*8/csa.n);
}
free(fidx);
}
int csa_child_r_sub(CSA *csa, i64 len, rank_t l, rank_t r, int num, uchar *substr, uchar *tail, rank_t *ll, rank_t *rr)
{
rank_t lll, rrr;
i64 j, ret;
int c1,c2;
// printf("child_r_sub [%ld,%ld]\n",l,r);
lll = l; rrr = r;
for (j=0; j<len; j++) {
lll = csa->psi(csa, lll);
rrr = csa->psi(csa, rrr);
}
c1 = csa->T(csa, lll);
c2 = csa->T(csa, rrr);
if (c1 == c2) {
// printf("num=%d c=%d [%ld,%ld]\n",num,c1,l,r);
// tail[num] = c1;
putuint(tail,num,c1, csa->k2);
ll[num] = l;
rr[num] = r;
num++;
} else {
i64 l2,r2;
if (c1 == -1) {
l2 = l+1;
} else {
lll = l; rrr = r;
// substr[len] = c1;
putuint(substr, len, c1, csa->k2);
ret = csa->search(substr, len+1, csa, &lll, &rrr);
if (ret == len+1) {
// printf("num=%d c=%d [%ld,%ld]\n",num,c1,lll,rrr);
// tail[num] = c1;
putuint(tail,num,c1, csa->k2);
ll[num] = lll;
rr[num] = rrr;
num++;
l2 = rrr+1;
// } else {
// printf("??? c1\n");
}
}
if (c2 == -1) {
r2 = r-1;
} else {
lll = l; rrr = r;
// substr[len] = c2;
putuint(substr, len, c2, csa->k2);
ret = csa->search(substr, len+1, csa, &lll, &rrr);
if (ret == len+1) {
// printf("num=%d c=%d [%ld,%ld]\n",num,c2,lll,rrr);
// tail[num] = c2;
putuint(tail,num,c2, csa->k2);
ll[num] = lll;
rr[num] = rrr;
num++;
r2 = lll-1;
// } else {
// printf("??? c2\n");
}
}
if (l2 <= r2) {
num = csa_child_r_sub(csa, len, l2, r2, num, substr, tail, ll, rr);
}
}
return num;
}
