static void
populate_lrdiph(dict2pid_t *d2p, s3ssid_t ***rdiph_rc, s3cipid_t b)
{
bin_mdef_t *mdef = d2p->mdef;
s3cipid_t l, r;
for (l = 0; l < bin_mdef_n_ciphone(mdef); l++) {
for (r = 0; r < bin_mdef_n_ciphone(mdef); r++) {
s3pid_t p;
p = bin_mdef_phone_id_nearest(mdef, (s3cipid_t) b,
(s3cipid_t) l,
(s3cipid_t) r,
WORD_POSN_SINGLE);
d2p->lrdiph_rc[b][l][r]
= bin_mdef_pid2ssid(mdef, p);
if (r == bin_mdef_silphone(mdef))
d2p->ldiph_lc[b][r][l]
= bin_mdef_pid2ssid(mdef, p);
if (rdiph_rc && l == bin_mdef_silphone(mdef))
rdiph_rc[b][l][r]
= bin_mdef_pid2ssid(mdef, p);
assert(IS_S3SSID(bin_mdef_pid2ssid(mdef, p)));
E_DEBUG(2,("%s(%s,%s) => %d / %d\n",
bin_mdef_ciphone_str(mdef, b),
bin_mdef_ciphone_str(mdef, l),
bin_mdef_ciphone_str(mdef, r),
p, bin_mdef_pid2ssid(mdef, p)));
}
}
}
void fsg_psubtree_dump_node(fsg_lextree_t *tree, fsg_pnode_t *node, FILE *fp)
{
int32 i;
fsg_link_t *tl;
/* Indentation */
for (i = 0; i <= node->ppos; i++)
fprintf(fp, " ");
fprintf(fp, "%p.@", node); /* Pointer used as node
* ID */
fprintf(fp, " %5d.SS", hmm_nonmpx_ssid(&node->hmm));
fprintf(fp, " %10d.LP", node->logs2prob);
fprintf(fp, " %p.SIB", node->sibling);
fprintf(fp, " %s.%d", bin_mdef_ciphone_str(tree->mdef, node->ci_ext), node->ppos);
if ((node->ppos == 0) || node->leaf) {
fprintf(fp, " [");
for (i = 0; i < FSG_PNODE_CTXT_BVSZ; i++)
fprintf(fp, "%08x", node->ctxt.bv[i]);
fprintf(fp, "]");
}
if (node->leaf) {
tl = node->next.fsglink;
fprintf(fp, " {%s[%d->%d](%d)}",
fsg_model_word_str(tree->fsg, tl->wid),
tl->from_state, tl->to_state, tl->logs2prob);
} else {
fprintf(fp, " %p.NXT", node->next.succ);
}
fprintf(fp, "\n");
return;
}
const char *
dict_ciphone_str(dict_t * d, s3wid_t wid, int32 pos)
{
assert(d != NULL);
assert((wid >= 0) && (wid < d->n_word));
assert((pos >= 0) && (pos < d->word[wid].pronlen));
return bin_mdef_ciphone_str(d->mdef, d->word[wid].ciphone[pos]);
}
static void
allphone_search_fill_iter(ps_seg_t *seg, phseg_t *phseg)
{
seg->sf = phseg->sf;
seg->ef = phseg->ef;
seg->ascr = phseg->score;
seg->lscr = phseg->tscore;
seg->word = bin_mdef_ciphone_str(ps_search_acmod(seg->search)->mdef, phseg->ci);
}
int
bin_mdef_phone_str(bin_mdef_t * m, int pid, char *buf)
{
char *wpos_name;
assert(m);
assert((pid >= 0) && (pid < m->n_phone));
wpos_name = WPOS_NAME;
buf[0] = '\0';
if (pid < m->n_ciphone)
sprintf(buf, "%s", bin_mdef_ciphone_str(m, pid));
else {
sprintf(buf, "%s %s %s %c",
bin_mdef_ciphone_str(m, m->phone[pid].info.cd.ctx[0]),
bin_mdef_ciphone_str(m, m->phone[pid].info.cd.ctx[1]),
bin_mdef_ciphone_str(m, m->phone[pid].info.cd.ctx[2]),
wpos_name[m->phone[pid].info.cd.wpos]);
}
return 0;
}
void
dict2pid_dump(FILE * fp, dict2pid_t * d2p)
{
int32 w, p, pronlen;
int32 i, j, b, l, r;
bin_mdef_t *mdef = d2p->mdef;
dict_t *dict = d2p->dict;
fprintf(fp, "# INTERNAL (wd comssid ssid ssid ... ssid comssid)\n");
for (w = 0; w < dict_size(dict); w++) {
fprintf(fp, "%30s ", dict_wordstr(dict, w));
pronlen = dict_pronlen(dict, w);
for (p = 0; p < pronlen; p++)
fprintf(fp, " %5d", dict2pid_internal(d2p, w, p));
fprintf(fp, "\n");
}
fprintf(fp, "#\n");
fprintf(fp, "# LDIPH_LC (b r l ssid)\n");
for (b = 0; b < bin_mdef_n_ciphone(mdef); b++) {
for (r = 0; r < bin_mdef_n_ciphone(mdef); r++) {
for (l = 0; l < bin_mdef_n_ciphone(mdef); l++) {
if (IS_S3SSID(d2p->ldiph_lc[b][r][l]))
fprintf(fp, "%6s %6s %6s %5d\n", bin_mdef_ciphone_str(mdef, (s3cipid_t) b), bin_mdef_ciphone_str(mdef, (s3cipid_t) r), bin_mdef_ciphone_str(mdef, (s3cipid_t) l), d2p->ldiph_lc[b][r][l]); /* RAH, ldiph_lc is returning an int32, %d expects an int16 */
}
}
}
fprintf(fp, "#\n");
fprintf(fp, "# SSEQ %d (senid senid ...)\n", mdef->n_sseq);
for (i = 0; i < mdef->n_sseq; i++) {
fprintf(fp, "%5d ", i);
for (j = 0; j < bin_mdef_n_emit_state(mdef); j++)
fprintf(fp, " %5d", mdef->sseq[i][j]);
fprintf(fp, "\n");
}
fprintf(fp, "#\n");
fprintf(fp, "# END\n");
fflush(fp);
}
int
bin_mdef_phone_str(bin_mdef_t * m, int pid, char *buf)
{
char *wpos_name;
assert(m);
assert((pid >= 0) && (pid < m->n_phone));
wpos_name = WPOS_NAME;
buf[0] = '\0';
if (pid < m->n_ciphone)
{
#ifndef POCKETSPHINX_NET
sprintf(buf, "%s", bin_mdef_ciphone_str(m, pid));
#else
strcpy(buf, bin_mdef_ciphone_str(m, pid));
#endif
}
else
{
#ifndef POCKETSPHINX_NET
sprintf(buf, "%s %s %s %c",
bin_mdef_ciphone_str(m, m->phone[pid].info.cd.ctx[0]),
bin_mdef_ciphone_str(m, m->phone[pid].info.cd.ctx[1]),
bin_mdef_ciphone_str(m, m->phone[pid].info.cd.ctx[2]),
wpos_name[m->phone[pid].info.cd.wpos]);
#else
strcpy(buf, bin_mdef_ciphone_str(m, m->phone[pid].info.cd.ctx[0]));
strcat(buf, bin_mdef_ciphone_str(m, m->phone[pid].info.cd.ctx[1]));
strcat(buf, bin_mdef_ciphone_str(m, m->phone[pid].info.cd.ctx[2]));
char tbf [2] = {wpos_name[m->phone[pid].info.cd.wpos]};
strcat(buf, tbf);
#endif
}
return 0;
}
dict2pid_t *
dict2pid_build(bin_mdef_t * mdef, dict_t * dict)
{
dict2pid_t *dict2pid;
s3ssid_t ***rdiph_rc;
bitvec_t *ldiph, *rdiph, *single;
int32 pronlen;
int32 b, l, r, w, p;
E_INFO("Building PID tables for dictionary\n");
assert(mdef);
assert(dict);
dict2pid = (dict2pid_t *) ckd_calloc(1, sizeof(dict2pid_t));
dict2pid->refcount = 1;
dict2pid->mdef = bin_mdef_retain(mdef);
dict2pid->dict = dict_retain(dict);
E_INFO("Allocating %d^3 * %d bytes (%d KiB) for word-initial triphones\n",
mdef->n_ciphone, sizeof(s3ssid_t),
mdef->n_ciphone * mdef->n_ciphone * mdef->n_ciphone * sizeof(s3ssid_t) / 1024);
dict2pid->ldiph_lc =
(s3ssid_t ***) ckd_calloc_3d(mdef->n_ciphone, mdef->n_ciphone,
mdef->n_ciphone, sizeof(s3ssid_t));
/* Only used internally to generate rssid */
rdiph_rc =
(s3ssid_t ***) ckd_calloc_3d(mdef->n_ciphone, mdef->n_ciphone,
mdef->n_ciphone, sizeof(s3ssid_t));
dict2pid->lrdiph_rc = (s3ssid_t ***) ckd_calloc_3d(mdef->n_ciphone,
mdef->n_ciphone,
mdef->n_ciphone,
sizeof
(s3ssid_t));
/* Actually could use memset for this, if BAD_S3SSID is guaranteed
* to be 65535... */
for (b = 0; b < mdef->n_ciphone; ++b) {
for (r = 0; r < mdef->n_ciphone; ++r) {
for (l = 0; l < mdef->n_ciphone; ++l) {
dict2pid->ldiph_lc[b][r][l] = BAD_S3SSID;
dict2pid->lrdiph_rc[b][l][r] = BAD_S3SSID;
rdiph_rc[b][l][r] = BAD_S3SSID;
}
}
}
/* Track which diphones / ciphones have been seen. */
ldiph = bitvec_alloc(mdef->n_ciphone * mdef->n_ciphone);
rdiph = bitvec_alloc(mdef->n_ciphone * mdef->n_ciphone);
single = bitvec_alloc(mdef->n_ciphone);
for (w = 0; w < dict_size(dict2pid->dict); w++) {
pronlen = dict_pronlen(dict, w);
if (pronlen >= 2) {
b = dict_first_phone(dict, w);
r = dict_second_phone(dict, w);
/* Populate ldiph_lc */
if (bitvec_is_clear(ldiph, b * mdef->n_ciphone + r)) {
/* Mark this diphone as done */
bitvec_set(ldiph, b * mdef->n_ciphone + r);
/* Record all possible ssids for b(?,r) */
for (l = 0; l < bin_mdef_n_ciphone(mdef); l++) {
p = bin_mdef_phone_id_nearest(mdef, (s3cipid_t) b,
(s3cipid_t) l, (s3cipid_t) r,
WORD_POSN_BEGIN);
dict2pid->ldiph_lc[b][r][l] = bin_mdef_pid2ssid(mdef, p);
}
}
/* Populate rdiph_rc */
l = dict_second_last_phone(dict, w);
b = dict_last_phone(dict, w);
if (bitvec_is_clear(rdiph, b * mdef->n_ciphone + l)) {
/* Mark this diphone as done */
bitvec_set(rdiph, b * mdef->n_ciphone + l);
for (r = 0; r < bin_mdef_n_ciphone(mdef); r++) {
p = bin_mdef_phone_id_nearest(mdef, (s3cipid_t) b,
(s3cipid_t) l, (s3cipid_t) r,
WORD_POSN_END);
rdiph_rc[b][l][r] = bin_mdef_pid2ssid(mdef, p);
}
}
}
else if (pronlen == 1) {
b = dict_pron(dict, w, 0);
E_DEBUG(1,("Building tables for single phone word %s phone %d = %s\n",
dict_wordstr(dict, w), b, bin_mdef_ciphone_str(mdef, b)));
/* Populate lrdiph_rc (and also ldiph_lc, rdiph_rc if needed) */
if (bitvec_is_clear(single, b)) {
populate_lrdiph(dict2pid, rdiph_rc, b);
bitvec_set(single, b);
}
}
}
bitvec_free(ldiph);
bitvec_free(rdiph);
bitvec_free(single);
/* Try to compress rdiph_rc into rdiph_rc_compressed */
compress_right_context_tree(dict2pid, rdiph_rc);
compress_left_right_context_tree(dict2pid);
ckd_free_3d(rdiph_rc);
dict2pid_report(dict2pid);
return dict2pid;
}
int
dict2pid_add_word(dict2pid_t *d2p,
int32 wid)
{
bin_mdef_t *mdef = d2p->mdef;
dict_t *d = d2p->dict;
if (dict_pronlen(d, wid) > 1) {
s3cipid_t l;
/* Make sure we have left and right context diphones for this
* word. */
if (d2p->ldiph_lc[dict_first_phone(d, wid)][dict_second_phone(d, wid)][0]
== BAD_S3SSID) {
E_INFO("Filling in left-context diphones for %s(?,%s)\n",
bin_mdef_ciphone_str(mdef, dict_first_phone(d, wid)),
bin_mdef_ciphone_str(mdef, dict_second_phone(d, wid)));
for (l = 0; l < bin_mdef_n_ciphone(mdef); l++) {
s3ssid_t p
= bin_mdef_phone_id_nearest(mdef,
dict_first_phone(d, wid), l,
dict_second_phone(d, wid),
WORD_POSN_BEGIN);
d2p->ldiph_lc[dict_first_phone(d, wid)][dict_second_phone(d, wid)][l]
= bin_mdef_pid2ssid(mdef, p);
}
}
if (d2p->rssid[dict_last_phone(d, wid)][dict_second_last_phone(d, wid)].n_ssid
== 0) {
s3ssid_t *rmap;
s3ssid_t *tmpssid;
s3cipid_t *tmpcimap;
s3cipid_t r;
E_INFO("Filling in right-context diphones for %s(%s,?)\n",
bin_mdef_ciphone_str(mdef, dict_last_phone(d, wid)),
bin_mdef_ciphone_str(mdef, dict_second_last_phone(d, wid)));
rmap = ckd_calloc(bin_mdef_n_ciphone(mdef), sizeof(*rmap));
for (r = 0; r < bin_mdef_n_ciphone(mdef); r++) {
s3ssid_t p
= bin_mdef_phone_id_nearest(mdef,
dict_last_phone(d, wid),
dict_second_last_phone(d, wid), r,
WORD_POSN_END);
rmap[r] = bin_mdef_pid2ssid(mdef, p);
}
tmpssid = ckd_calloc(bin_mdef_n_ciphone(mdef), sizeof(*tmpssid));
tmpcimap = ckd_calloc(bin_mdef_n_ciphone(mdef), sizeof(*tmpcimap));
compress_table(rmap, tmpssid, tmpcimap, bin_mdef_n_ciphone(mdef));
for (r = 0; r < mdef->n_ciphone && tmpssid[r] != BAD_S3SSID; r++)
;
d2p->rssid[dict_last_phone(d, wid)][dict_second_last_phone(d, wid)].ssid = tmpssid;
d2p->rssid[dict_last_phone(d, wid)][dict_second_last_phone(d, wid)].cimap = tmpcimap;
d2p->rssid[dict_last_phone(d, wid)][dict_second_last_phone(d, wid)].n_ssid = r;
ckd_free(rmap);
}
}
else {
/* Make sure we have a left-right context triphone entry for
* this word. */
E_INFO("Filling in context triphones for %s(?,?)\n",
bin_mdef_ciphone_str(mdef, dict_first_phone(d, wid)));
if (d2p->lrdiph_rc[dict_first_phone(d, wid)][0][0] == BAD_S3SSID) {
populate_lrdiph(d2p, NULL, dict_first_phone(d, wid));
}
}
return 0;
}