int
ps_add_word(ps_decoder_t *ps,
char const *word,
char const *phones,
int update)
{
int32 wid, lmwid;
ngram_model_t *lmset;
s3cipid_t *pron;
char **phonestr, *tmp;
int np, i, rv;
/* Parse phones into an array of phone IDs. */
tmp = ckd_salloc(phones);
np = str2words(tmp, NULL, 0);
phonestr = ckd_calloc(np, sizeof(*phonestr));
str2words(tmp, phonestr, np);
pron = ckd_calloc(np, sizeof(*pron));
for (i = 0; i < np; ++i) {
pron[i] = bin_mdef_ciphone_id(ps->acmod->mdef, phonestr[i]);
if (pron[i] == -1) {
E_ERROR("Unknown phone %s in phone string %s\n",
phonestr[i], tmp);
ckd_free(phonestr);
ckd_free(tmp);
ckd_free(pron);
return -1;
}
}
/* No longer needed. */
ckd_free(phonestr);
ckd_free(tmp);
/* Add it to the dictionary. */
if ((wid = dict_add_word(ps->dict, word, pron, np)) == -1) {
ckd_free(pron);
return -1;
}
/* No longer needed. */
ckd_free(pron);
/* Now we also have to add it to dict2pid. */
dict2pid_add_word(ps->d2p, wid);
if ((lmset = ps_get_lmset(ps)) != NULL) {
/* Add it to the LM set (meaning, the current LM). In a perfect
* world, this would result in the same WID, but because of the
* weird way that word IDs are handled, it doesn't. */
if ((lmwid = ngram_model_add_word(lmset, word, 1.0))
== NGRAM_INVALID_WID)
return -1;
}
/* Rebuild the widmap and search tree if requested. */
if (update) {
if ((rv = ps_search_reinit(ps->search, ps->dict, ps->d2p) < 0))
return rv;
}
return wid;
}
ngram_model_t *
ps_update_lmset(ps_decoder_t *ps, ngram_model_t *lmset)
{
ngram_search_t *ngs;
ps_search_t *search;
/* Look for N-Gram search. */
search = ps_find_search(ps, "ngram");
if (search == NULL) {
/* Initialize N-Gram search. */
search = ngram_search_init(ps->config, ps->acmod, ps->dict, ps->d2p);
if (search == NULL)
return NULL;
search->pls = ps->phone_loop;
ps->searches = glist_add_ptr(ps->searches, search);
ngs = (ngram_search_t *)search;
}
else {
ngs = (ngram_search_t *)search;
/* Free any previous lmset if this is a new one. */
if (ngs->lmset != NULL && ngs->lmset != lmset)
ngram_model_free(ngs->lmset);
ngs->lmset = lmset;
/* Tell N-Gram search to update its view of the world. */
if (ps_search_reinit(search, ps->dict, ps->d2p) < 0)
return NULL;
}
ps->search = search;
return ngs->lmset;
}
int
ps_load_dict(ps_decoder_t *ps, char const *dictfile,
char const *fdictfile, char const *format)
{
cmd_ln_t *newconfig;
dict2pid_t *d2p;
dict_t *dict;
hash_iter_t *search_it;
/* Create a new scratch config to load this dict (so existing one
* won't be affected if it fails) */
newconfig = cmd_ln_init(NULL, ps_args(), TRUE, NULL);
cmd_ln_set_boolean_r(newconfig, "-dictcase",
cmd_ln_boolean_r(ps->config, "-dictcase"));
cmd_ln_set_str_r(newconfig, "-dict", dictfile);
if (fdictfile)
cmd_ln_set_str_r(newconfig, "-fdict", fdictfile);
else
cmd_ln_set_str_r(newconfig, "-fdict",
cmd_ln_str_r(ps->config, "-fdict"));
/* Try to load it. */
if ((dict = dict_init(newconfig, ps->acmod->mdef, ps->acmod->lmath)) == NULL) {
cmd_ln_free_r(newconfig);
return -1;
}
/* Reinit the dict2pid. */
if ((d2p = dict2pid_build(ps->acmod->mdef, dict)) == NULL) {
cmd_ln_free_r(newconfig);
return -1;
}
/* Success! Update the existing config to reflect new dicts and
* drop everything into place. */
cmd_ln_free_r(newconfig);
cmd_ln_set_str_r(ps->config, "-dict", dictfile);
if (fdictfile)
cmd_ln_set_str_r(ps->config, "-fdict", fdictfile);
dict_free(ps->dict);
ps->dict = dict;
dict2pid_free(ps->d2p);
ps->d2p = d2p;
/* And tell all searches to reconfigure themselves. */
for (search_it = hash_table_iter(ps->searches); search_it;
search_it = hash_table_iter_next(search_it)) {
if (ps_search_reinit(hash_entry_val(search_it->ent), dict, d2p) < 0) {
hash_table_iter_free(search_it);
return -1;
}
}
return 0;
}
int
ps_load_dict(ps_decoder_t *ps, char const *dictfile,
char const *fdictfile, char const *format)
{
cmd_ln_t *newconfig;
dict2pid_t *d2p;
dict_t *dict;
gnode_t *gn;
int rv;
/* Create a new scratch config to load this dict (so existing one
* won't be affected if it fails) */
newconfig = cmd_ln_init(NULL, ps_args(), TRUE, NULL);
cmd_ln_set_boolean_r(newconfig, "-dictcase",
cmd_ln_boolean_r(ps->config, "-dictcase"));
cmd_ln_set_str_r(newconfig, "-dict", dictfile);
if (fdictfile)
cmd_ln_set_str_r(newconfig, "-fdict", fdictfile);
else
cmd_ln_set_str_r(newconfig, "-fdict",
cmd_ln_str_r(ps->config, "-fdict"));
/* Try to load it. */
if ((dict = dict_init(newconfig, ps->acmod->mdef)) == NULL) {
cmd_ln_free_r(newconfig);
return -1;
}
/* Reinit the dict2pid. */
if ((d2p = dict2pid_build(ps->acmod->mdef, dict)) == NULL) {
cmd_ln_free_r(newconfig);
return -1;
}
/* Success! Update the existing config to reflect new dicts and
* drop everything into place. */
cmd_ln_free_r(newconfig);
cmd_ln_set_str_r(ps->config, "-dict", dictfile);
if (fdictfile)
cmd_ln_set_str_r(ps->config, "-fdict", fdictfile);
dict_free(ps->dict);
ps->dict = dict;
dict2pid_free(ps->d2p);
ps->d2p = d2p;
/* And tell all searches to reconfigure themselves. */
for (gn = ps->searches; gn; gn = gnode_next(gn)) {
ps_search_t *search = gnode_ptr(gn);
if ((rv = ps_search_reinit(search, dict, d2p)) < 0)
return rv;
}
return 0;
}
fsg_set_t *
ps_update_fsgset(ps_decoder_t *ps)
{
ps_search_t *search;
/* Look for FSG search. */
search = ps_find_search(ps, "fsg");
if (search == NULL) {
/* Initialize FSG search. */
search = fsg_search_init(ps->config,
ps->acmod, ps->dict, ps->d2p);
search->pls = ps->phone_loop;
ps->searches = glist_add_ptr(ps->searches, search);
}
else {
/* Tell FSG search to update its view of the world. */
if (ps_search_reinit(search, ps->dict, ps->d2p) < 0)
return NULL;
}
ps->search = search;
return (fsg_set_t *)search;
}
int
ps_add_word(ps_decoder_t *ps,
char const *word,
char const *phones,
int update)
{
int32 wid;
s3cipid_t *pron;
hash_iter_t *search_it;
char **phonestr, *tmp;
int np, i, rv;
/* Parse phones into an array of phone IDs. */
tmp = ckd_salloc(phones);
np = str2words(tmp, NULL, 0);
phonestr = ckd_calloc(np, sizeof(*phonestr));
str2words(tmp, phonestr, np);
pron = ckd_calloc(np, sizeof(*pron));
for (i = 0; i < np; ++i) {
pron[i] = bin_mdef_ciphone_id(ps->acmod->mdef, phonestr[i]);
if (pron[i] == -1) {
E_ERROR("Unknown phone %s in phone string %s\n",
phonestr[i], tmp);
ckd_free(phonestr);
ckd_free(tmp);
ckd_free(pron);
return -1;
}
}
/* No longer needed. */
ckd_free(phonestr);
ckd_free(tmp);
/* Add it to the dictionary. */
if ((wid = dict_add_word(ps->dict, word, pron, np)) == -1) {
ckd_free(pron);
return -1;
}
/* No longer needed. */
ckd_free(pron);
/* Now we also have to add it to dict2pid. */
dict2pid_add_word(ps->d2p, wid);
/* TODO: we definitely need to refactor this */
for (search_it = hash_table_iter(ps->searches); search_it;
search_it = hash_table_iter_next(search_it)) {
ps_search_t *search = hash_entry_val(search_it->ent);
if (!strcmp(PS_SEARCH_NGRAM, ps_search_name(search))) {
ngram_model_t *lmset = ((ngram_search_t *) search)->lmset;
if (ngram_model_add_word(lmset, word, 1.0) == NGRAM_INVALID_WID) {
hash_table_iter_free(search_it);
return -1;
}
}
if (update) {
if ((rv = ps_search_reinit(search, ps->dict, ps->d2p) < 0)) {
hash_table_iter_free(search_it);
return rv;
}
}
}
/* Rebuild the widmap and search tree if requested. */
return wid;
}