void ThetaRoleCheckSubcat(char *features)
{
char *p, *save;
for (p = features; *p; p++) {
if (StringIn(*p, FT_SUBCAT)) {
save = p;
p--;
if (*p == F_OPTIONAL) {
p--;
}
if (p <= features || !StringIn(*p, FT_OBJLOC)) {
Dbg(DBGLEX, DBGBAD,
"<%s> FT_SUBCAT doesn't follow FT_OBJLOC [F_OPTIONAL]", features);
}
p = save;
}
}
for (p = features; *p; p++) {
if (*p == F_OPTIONAL) {
if (p <= features || !StringIn(*(p-1), FT_OBJLOC)) {
Dbg(DBGLEX, DBGBAD,
"<%s> F_OPTIONAL doesn't follow FT_OBJLOC", features);
}
}
}
}
Bool CharPhoneMatch(int ptn, int c)
{
switch (ptn) {
case 'X': return(StringIn(c, "0123456789"));
case 'N': return(StringIn(c, "23456789"));
case '=': return(StringIn(c, PHONE_WHITESPACE));
default:
return(ptn == c);
}
}
Bool FeatureTaskOK(char *features, Obj *task)
{
if (task == N("parse")) {
return(!StringIn(F_TRANS_ONLY, features));
} else if (task == N("generate")) {
return(!StringIn(F_TRANS_ONLY, features));
} else if (task == N("translate")) {
return(1);
}
return(1);
}
Float Syn_ParseFilterZZ_Z(PNode *z1, PNode *z2, int lang)
{
PNode *pn;
if (z1->pn1 && z1->pn1->feature == F_INTERJECTION &&
z1->pn2 == NULL) {
/* "What, are you thinking?" ACCEPT THIS PARSE vs.
* "What are you thinking" REJECT THIS PARSE
*/
if (StringIn(',', PNodeGetEndPunc(z1))) return(1.0);
else return(0.0);
}
if ((pn = PNodeLeftmost(z1)) &&
(pn->feature == F_CONJUNCTION)) {
if (!Syn_ParseZZ_ZConjunctionOK(pn)) {
return(0.0);
} else if (pn->lexitem && pn->lexitem->le &&
LexEntryAllConceptsFeat(F_CLAUSE2_ONLY, pn->lexitem->le)) {
return(0.0);
} else {
return(1.0);
}
}
if ((pn = PNodeLeftmost(z2)) &&
(pn->feature == F_CONJUNCTION)) {
if (!Syn_ParseZZ_ZConjunctionOK(pn)) {
return(0.0);
} else {
return(1.0);
}
}
if (StringIn(',', PNodeGetEndPunc(z1))) {
/* Allow without conjunction if comma is used. In French these are frequent.
* In English these are frowned upon "run-on sentences":
* "The air was cold and humid, I hate that."
* However if there is an elipsis, this is not considered a run-on sentence:
* "The air was cold and humid, the sky overcast."
*/
if (PNodeNumberOfWords(z1) >= 5 &&
PNodeNumberOfWords(z2) <= 1) {
/* Disallow if too lopsided:
* "The American who is Jim Garnier's sister, ate."
* todoSCORE
*/
} else {
return(1.0);
}
}
/* Disallow. */
return(0.0);
}
void TA_LELAdd(LexEntry *le, char *features, int maxlen,
/* RESULTS */ LexEntry **le_le, char **le_feat, int *le_len)
{
int i, len;
char buf[FEATLEN];
len = *le_len;
if (!StringIn(F_VERB, features)) {
/* This condition allows "Qui a peint Le Déjeuner sur l herbe?" to parse.
* But is collapsing needed for verbs for other reasons?
* It would be nice to collapse verbs to reduce number of parses, but we
* can't reduce them because tense (and the resulting aspect)
* information is significant.
*/
for (i = 0; i < len; i++) {
if (le == le_le[i]) {
StringIntersect(le_feat[i], features, buf);
StringCpy(le_feat[i], buf, FEATLEN);
return;
}
}
}
if (len < maxlen) {
le_le[len] = le;
le_feat[len] = MemAlloc(FEATLEN, "char TA_LELAdd");
StringCpy(le_feat[len], features, FEATLEN);
len++;
*le_len = len;
} else {
Dbg(DBGGEN, DBGBAD, "TA_LELAdd: increase maxlen <%s.%s>",
le->srcphrase, le->features);
}
}
void RptTextPrint1(Text *text, char *s, int fieldlen, int just)
{
int i, slen, total;
if (StringIn(NEWLINE, s)) {
if (0) {
/* if (strlen(s) < fieldlen) { */
StringElimChar(s, NEWLINE); /* Destructive. Watch out. */
TextPuts(s, text);
} else {
TextPutsIndented(s, text, " ");
return;
}
}
slen = strlen(s);
if (slen > fieldlen) {
for (i = 0; i < fieldlen; i++) TextPutc(s[i], text);
return;
}
if (just == RPT_JUST_RIGHT) {
total = fieldlen - slen;
TextPrintSpaces(text, total);
} else if (just == RPT_JUST_CENTER) {
total = (fieldlen - slen)/2;
TextPrintSpaces(text, total);
} else total = 0;
TextPuts(s, text);
total += slen;
TextPrintSpaces(text, fieldlen-total);
}
void StringMapLeWhitespaceToDash(char *to)
{
while (*to) {
if (StringIn(*to, LE_WHITESPACE)) *to = '-';
to++;
}
}
char *StringReadToNons(char *in, int including, char *out, int maxlen,
char *seps)
{
char *p;
p = out;
maxlen--;
while (*in && StringIn(*in, seps)) {
if (*in == TREE_ESCAPE) {
*p = *in;
p++; in++;
if (in[0] == TERM) {
Dbg(DBGGEN, DBGBAD, "StringReadToNons: empty escape");
break;
}
}
if ((p - out) >= maxlen) {
out[maxlen] = TERM;
Dbg(DBGGEN, DBGBAD, "StringReadToNons: overflow reading up to <%s>: <%s>",
seps, out);
return(in);
}
*p = *in;
p++; in++;
}
*p = TERM;
if (including && (*in != TERM)) in++;
return(in);
}
void StringMapLeWhitespaceToSpace(char *to)
{
while (*to) {
if (StringIn(*to, LE_WHITESPACE)) *to = SPACE;
to++;
}
}
PNode *TransformGenFrenchContractions(PNode *pn, Discourse *dc, /* RESULTS */
int *change)
{
int det_gender, det_number;
PNode *pnword2;
Obj *noun_con;
pnword2 = PNodeLeftmost(pn->pn2);
if (pn->pn1 && pn->pn1->feature == F_PREPOSITION &&
pnword2 && pnword2->feature == F_DETERMINER &&
pnword2->lexitem &&
LexEntryConceptIsAncestor(N("definite-article"), pnword2->lexitem->le)) {
det_number = FeatureGet(pnword2->lexitem->features, FT_NUMBER);
det_gender = FeatureGet(pnword2->lexitem->features, FT_GENDER);
if ((!StringIn(F_ELISION, pnword2->lexitem->features)) &&
(det_number == F_PLURAL || det_gender == F_MASCULINE)) {
if (LexEntryConceptIsAncestor(N("prep-to"), pn->pn1->lexitem->le)) {
/* à le -> au
* à les -> aux
*/
pn->pn1 = GenMakePrep(N("prep-au"), det_gender, det_number, dc);
LexitemMakeTrace(pnword2->lexitem);
*change = 1;
return(pn);
} else if (LexEntryConceptIsAncestor(N("prep-of"),
pn->pn1->lexitem->le)) {
/* de le -> du
* de les -> des
*/
pn->pn1 = GenMakePrep(N("prep-du"), det_gender, det_number, dc);
LexitemMakeTrace(pnword2->lexitem);
*change = 1;
return(pn);
}
}
if ((noun_con = PNodeNounConcept(pn->pn2)) &&
ISA(N("polity"), noun_con) &&
LexEntryConceptIsAncestor(N("prep-en"), pn->pn1->lexitem->le)) {
/* If noun is "polity".
* en la -> en
* en le + vowel -> en
* en le + nonvowel -> au
* en les -> aux
*/
if (det_gender == F_FEMININE ||
(F_VOCALIC ==
LexitemInitialSound(PNodeLeftmostLexitem(pn->pn2), dc))) {
LexitemMakeTrace(pnword2->lexitem);
*change = 1;
return(pn);
} else {
pn->pn1 = GenMakePrep(N("prep-au"), det_gender, det_number, dc);
LexitemMakeTrace(pnword2->lexitem);
*change = 1;
return(pn);
}
}
}
return(pn);
}
Bool StringIsDigitOr(char *s, char *set)
{
while (*s) {
if ((!Char_isdigit(*s)) && (!StringIn(*s, set))) return(0);
s++;
}
return(1);
}
Bool StringAllIn(char *s1, char *s2)
{
while (*s1) {
if (!StringIn(*s1, s2)) return(0);
s1++;
}
return(1);
}
Bool StringIsAlphaOr(char *s, char *set)
{
while (*s) {
if ((!CharIsAlpha(*s)) && (!StringIn(*s, set))) return(0);
s++;
}
return(1);
}
/* Used during parsing when parent is not available. */
Bool XBarValidE_MAX(PNode *e)
{
LexEntry *le;
if ((le = PNodeLeftmostLexEntry(e)) &&
StringIn(F_CONJUNCTION, le->features)) {
return(0);
}
return(1);
}
void StringUnion(char *in1, char *in2, int maxlen, /* RESULTS */ char *out)
{
int i;
char *orig_out,
orig_out = out;
maxlen--;
for (i = 1; i < 256; i++) {
if (StringIn(i, in1) || StringIn(i, in2)) {
if ((out - orig_out) >= maxlen) {
orig_out[maxlen] = TERM;
return;
}
*out = i;
out++;
}
}
*out = TERM;
}
Bool StringAnyIn(char *s1, char *s2)
{
if (s1[0] == TERM) return(1);
while (*s1) {
if (StringIn(*s1, s2)) return(1);
s1++;
}
return(0);
}
void StringAppendIfNotAlreadyIn(int in, int maxlen, /* RESULTS */ char *out)
{
char *p;
maxlen--;
p = StringEndOf(out);
if (!StringIn(in, out)) {
if ((p - out) >= maxlen) { out[maxlen] = TERM; return; }
p[0] = in;
p[1] = TERM;
}
}
void StringIntersect(char *in1, char *in2, /* RESULTS */ char *out)
{
while (*in1) {
if (StringIn(*in1, in2)) {
*out = *in1;
out++;
}
in1++;
}
*out = TERM;
}
void Load()
{
Ts ts1, ts2;
Dur d;
TsSetNow(&ts1);
if (StringIn(F_FRENCH, StdDiscourse->langs)) {
LexEntryReadInflFile("db/frinfl.txt");
}
if (StringIn(F_ENGLISH, StdDiscourse->langs)) {
LexEntryReadInflFile("db/eninfl.txt");
}
DbFileRead("db/name.txt", DBFILETYPE_ISA);
DbFileRead("db/fooddrug.txt", DBFILETYPE_ISA);
DbFileRead("db/geog.txt", DBFILETYPE_POLITY);
DbFileRead("db/absobj.txt", DBFILETYPE_ISA);
DbFileRead("db/street.txt", DBFILETYPE_ISA);
DbFileRead("db/grid.txt", DBFILETYPE_ISA);
DbFileRead("db/all.txt", DBFILETYPE_ISA);
DbFileRead("db/physics.txt", DBFILETYPE_ISA);
DbFileRead("db/chem.txt", DBFILETYPE_ISA);
DbFileRead("db/trans.txt", DBFILETYPE_ISA);
DbFileRead("db/celest.txt", DBFILETYPE_ISA);
DbFileRead("db/physobj.txt", DBFILETYPE_ISA);
DbFileRead("db/ling.txt", DBFILETYPE_ISA);
/* FeatPrintUnused(stdout); */
DbFileRead("db/relation.txt", DBFILETYPE_ISA);
DbFileRead("db/action.txt", DBFILETYPE_ISA);
DbFileRead("db/attr.txt", DBFILETYPE_ISA);
DbFileRead("db/enum.txt", DBFILETYPE_ISA);
DbFileRead("db/living.txt", DBFILETYPE_ISA);
DbFileRead("db/human.txt", DBFILETYPE_ISA);
DbFileRead("db/clothing.txt", DBFILETYPE_ISA);
DbFileRead("db/mediaobj.txt", DBFILETYPE_ISA);
DbFileRead("db/music.txt", DBFILETYPE_ISA);
DbFileRead("db/tv.txt", DBFILETYPE_ISA);
DbFileRead("db/company.txt", DBFILETYPE_ISA);
DbFileRead("db/elec.txt", DBFILETYPE_ISA);
TsSetNow(&ts2);
d = TsMinus(&ts2, &ts1);
Dbg(DBGGEN, DBGBAD, "Load time = %.2ld:%.2ld.", d/60, d%60);
}
Bool StringPhoneMatch(char *ptn, char *s, /* RESULTS */ char *digits,
char **nextp)
{
while (*ptn) {
if (!CharPhoneMatch(*ptn, *s)) return(0);
if (StringIn(*s, "0123456789")) *digits++ = *s;
ptn++;
s++;
}
*digits = TERM;
*nextp = s;
return(1);
}
void FeatSubstPOS(char *in, int newpos, /* RESULTS */ char *out)
{
while (*in) {
if (StringIn(*in, FT_POS)) {
*out = newpos;
} else {
*out = *in;
}
in++;
out++;
}
*out = TERM;
}
Bool StringLineIsAll(char *s, char *set, /* RESULTS */ char **next_line)
{
while (*s) {
if (*s == NEWLINE) {
*next_line = s+1;
return(1);
}
if (!StringIn(*s, set)) return(0);
s++;
}
/* Hit EOF; we'll accept. */
*next_line = s;
return(1);
}
/* Used during parsing when parent is not available. */
Bool XBarValidX_MAX(PNode *x)
{
LexEntry *le;
if ((le = PNodeLeftmostLexEntry(x)) &&
StringIn(F_CONJUNCTION, le->features)) {
return(0);
}
if (x->pn1 && x->pn1->feature == F_PRONOUN && x->pn2 == NULL) {
/* Reject [Z [X [H <à peine de quoi.Hy>]] [W [W [V <remplir.fVy>]]]] */
if (!LexEntryConceptIsAncestor(N("noun-phrase-pronoun"),
PNodeLeftmostLexEntry(x))) {
return(0);
}
}
return(1);
}
void StringElims(char *to, char *set, /* RESULTS */ int *mods)
{
char *from;
from = to;
if (mods) *mods = 0;
while (*from) {
if (!StringIn(*from, set)) {
*to = *from;
to++;
from++;
} else {
if (mods) *mods = 1;
from++;
}
}
*to = TERM;
}
/* Returns pointer to first non separator character. */
char *StringSkipToNons(char *in, char *seps)
{
while (*in && StringIn(*in, seps)) {
if (*in == TREE_ESCAPE) {
in++;
if (*in == TERM) {
Dbg(DBGGEN, DBGBAD, "StringSkipNons");
return(in);
}
}
if (*in == TERM) {
Dbg(DBGGEN, DBGBAD, "StringSkipNons: hit end skipping to <%s>", seps);
return(in);
}
in++;
}
return(in);
}
void StringAppendIfNotAlreadyIns(char *in, int maxlen, /* RESULTS */ char *out)
{
char *p;
maxlen--;
p = StringEndOf(out);
while (*in) {
if (!StringIn(*in, out)) {
if ((p - out) >= maxlen) {
out[maxlen] = TERM;
return;
}
*p = *in;
p++;
*p = TERM;
}
in++;
}
*p = TERM;
}
PNode *TranslateAWord(PNode *pn, PNode *pnp, Obj *max, int pos, int tense,
int gender, int number, int person, int degree,
int srclang, int tgtlang, Discourse *dc)
{
int number1;
ObjList *p;
ObjToLexEntry *ole_src, *ole_tgt;
Word *infl;
PNode *r, *r1;
r = NULL;
for (p = TranslateGetAllcons(pn, pn->lexitem ? pn->lexitem->le : NULL);
p; p = p->next) {
if (ObjIsList(p->obj)) {
continue;
}
/* todo: Put FT_FILTER checks here. */
if (!TranslateGetOles(pos, p->obj, pn->lexitem->le, NULL, tgtlang, dc,
&ole_src, &ole_tgt)) {
continue;
}
if (StringIn(F_COMMON_INFL, ole_tgt->features)) {
number1 = FeatureGet(ole_tgt->features, FT_NUMBER);
} else {
number1 = number;
}
if (!(infl = LexEntryGetInflection(ole_tgt->le, tense, gender, number1,
person, F_NULL, degree, 1, dc))) {
continue;
}
r1 = PNodeWord(pos, infl->word, infl->features, ole_tgt->le, p->obj);
if (pn->punc[0]) TranslatePunc(r1->punc, pn->punc, srclang, tgtlang);
r1->ole = ole_tgt;
r1->obj = p->obj;
r1->next_altern = r;
r = r1;
}
if (r == NULL) {
return(pn);
}
return(r);
}
Bool ThetaRoleGetSubcat(int c, char *s, /* RESULTS */ int *isoptional,
int *subcat)
{
*isoptional = 0;
while (*s) {
if (c == *((uc *)s)) {
s++;
if (*s == F_OPTIONAL) {
*isoptional = 1;
s++;
}
if (StringIn(*((uc *)s), FT_SUBCAT)) {
*subcat = *((uc *)s);
s++;
} else {
*subcat = F_NULL;
}
return(1);
}
s++;
}
return(0);
}
Float Syn_ParseFilterYZ_Z(PNode *y, PNode *z, int lang)
{
if (Syn_ParseFilter_IsXE(z)) return(0.0);
if (Syn_ParseFilter_IsPrepRel(y)) {
/* y:[avec qui] z:[vous parlez francais] */
return(1.0);
}
if (StringIn(',', PNodeGetEndPunc(y))) {
/* "par précaution," */
return(1.0);
}
if ((PNodeClassIn(y, N("interrogative-pronoun")) ||
PNodeClassIn(y, N("interrogative-determiner"))) &&
Syn_ParseIsCompleteSentence(z, lang)) {
/* "De quel instrument jouait Paganini?" */
return(1.0);
}
return(0.0);
}
Bool WordFormSkipLink(char *linkfeat)
{
return(StringIn(F_INFREQUENT, linkfeat));
}
