static void
serialize_psus(FILE *f_xcl,Hash_table*psus)
{
static_f_xcl = f_xcl;
fputs("<psus>",f_xcl);
hash_exec2(psus, (hash_exec2_func*)serialize_psu_list);
fputs("</psus>",f_xcl);
static_f_xcl = NULL;
}
int
main(int argc, char **argv)
{
setlocale(LC_ALL,LOCALE);
sig_pool = npool_init();
signiary = hash_create(1000);
pertext = hash_create(1000);
psl_init();
runexpatNS(i_stdin,NULL,sH,eH,":");
hash_exec2(signiary, printsigns);
npool_term(sig_pool);
psl_term();
return 1;
}
static void
serialize_one_node(FILE*f_xcl,void *vp)
{
enum xcl_node_types t = ((struct xcl_c *)vp)->node_type;
/* int need_ll = 0; */
int i, unwrapping = 0, phrase_index = 0;
#define cp ((struct xcl_c*)vp)
#define dp ((struct xcl_d*)vp)
#define lp ((struct xcl_l*)vp)
switch (t)
{
case xcl_node_c:
if (cp->type == xcl_c_sentence)
{
if (no_words(cp))
unwrapping = 1;
else
{
int skipped_any_phrases = 0;
int set_cp_subtype = 0;
while (cp->nchildren && singleton_phrase_children(cp, &phrase_index))
{
const char *phrase_subtype = cp->children[phrase_index].c->subtype;
int skip_phrase = 0;
if (!phrase_subtype)
phrase_subtype = (const char *)ilem_para_head_label(cp,1);
if (phrase_subtype)
{
if ((set_cp_subtype || !cp->subtype)
&& (!phrase_subtype || 'S' == *phrase_subtype))
{
cp->subtype = phrase_subtype;
set_cp_subtype = 1;
skip_phrase = 1;
/* either this was already NULL or we've used it as
phrase_subtype */
cp->children[phrase_index].c->subtype = NULL;
}
}
else
skip_phrase = 1;
if (skip_phrase)
{
union xcl_u tmp = cp->children[phrase_index];
union xcl_u *free_me = cp->children;
union xcl_u *new_c = NULL;
int n_new_c = 0, src, dest;
n_new_c = cp->nchildren + tmp.c->nchildren - 1;
new_c = malloc(n_new_c * sizeof(union xcl_u));
for (src = dest = 0; src < phrase_index; )
new_c[dest++].c = cp->children[src++].c;
for (src = 0; src < tmp.c->nchildren; ++src)
new_c[dest++].c = tmp.c->children[src].c;
for (src = phrase_index + 1; src < cp->nchildren; )
new_c[dest++].c = cp->children[src++].c;
cp->nchildren = cp->children_alloced = n_new_c;
cp->children = new_c;
#if 0
cp->children = tmp.c->children;
cp->nchildren = tmp.c->nchildren;
cp->children_alloced = tmp.c->children_alloced;
#endif
free(free_me);
tmp.c->nchildren = tmp.c->children_alloced = 0;
tmp.c->children = NULL;
++skipped_any_phrases;
}
else
break;
}
/* if (skipped_any_phrases) */
goto skipping_phrase;
}
}
else if (cp->type == xcl_c_phrase && singleton_phrase_children(cp, &phrase_index))
{
unwrapping = 1;
if (cp->subtype && !cp->children[phrase_index].c->subtype)
cp->children[phrase_index].c->subtype = cp->subtype;
}
else
{
skipping_phrase:
fprintf(f_xcl,"<c");
x2_attr(f_xcl,"type",xcl_c_type_str[cp->type]);
if (cp->type == xcl_c_sentence)
x2_attr(f_xcl,"implicit","yes");
if (cp->subtype && *cp->subtype)
{
if (cp->type == xcl_c_sentence
|| cp->type == xcl_c_clause
|| cp->type == xcl_c_phrase)
x2_attr(f_xcl,"tag",cp->subtype);
else
x2_attr(f_xcl,"subtype",cp->subtype);
}
if (cp->id)
x2_attr(f_xcl,"xml:id",cp->id);
x2_attr(f_xcl,"label",cp->label);
x2_attr(f_xcl,"ref",cp->ref);
x2_attr_i(f_xcl,"level",cp->level);
fprintf(f_xcl," bracketing_level=\"%d\"",cp->bracketing_level);
fprintf(f_xcl,">");
if (cp->meta)
{
fprintf(f_xcl,"<mds xml:id=\"%s\">",(char*)hash_find(cp->meta,
(unsigned char *)"#xml:id"));
static_f_xcl = f_xcl;
hash_exec2(cp->meta,serialize_m);
fputs("</mds>",f_xcl);
}
}
for (i = 0; i < cp->nchildren; ++i)
{
switch (cp->children[i].c->node_type)
{
case xcl_node_c:
serialize_one_node(f_xcl,cp->children[i].c);
break;
case xcl_node_d:
serialize_one_node(f_xcl,cp->children[i].d);
break;
case xcl_node_l:
serialize_one_node(f_xcl,cp->children[i].l);
break;
}
}
if (!unwrapping)
fprintf(f_xcl,"</c>");
break;
case xcl_node_d:
fprintf(f_xcl,"<d type=\"%s\"",xcl_d_type_str[dp->type]);
if (dp->ref)
fprintf(f_xcl, " ref=\"%s\"", dp->ref);
if (dp->subtype)
fprintf(f_xcl, " subtype=\"%s\"", dp->subtype);
fputs("/>", f_xcl);
break;
case xcl_node_l:
#if 0
if (lp->f && lp->f->acount)
{
fprintf(f_xcl,"<ll type=\"ambig\" xml:id=\"%s\" ref=\"%s\">",lp->xml_id, lp->ref);
need_ll = 1;
}
#endif
serialize_one_l(f_xcl,lp);
#if 0
while (need_ll)
{
fprintf(f_xcl,"</ll>");
--need_ll;
}
#endif
break;
}
#undef cp
#undef dp
#undef lp
}
