X-Git-Url: http://git.indexdata.com/?p=yaz-moved-to-github.git;a=blobdiff_plain;f=src%2Fcclfind.c;h=07e94dc75fc50b20b137c1077612965b1e81a803;hp=4f24e7246615d31695576bedf9ed3727aa693aff;hb=240156e69d3a6390699644756e367fffd42891d1;hpb=c5b92d32f5619eae0b99c4bfb39374e8ef675e34 diff --git a/src/cclfind.c b/src/cclfind.c index 4f24e72..07e94dc 100644 --- a/src/cclfind.c +++ b/src/cclfind.c @@ -38,23 +38,27 @@ static int qual_val_type(ccl_qualifier_t *qa, int type, int value, char **attset) { int i; - if (!qa) return 0; for (i = 0; qa[i]; i++) { + int got_type = 0; struct ccl_rpn_attr *q = ccl_qual_get_attr(qa[i]); - while (q) + for (; q; q = q->next) { - if (q->type == type && q->kind == CCL_RPN_ATTR_NUMERIC && - q->value.numeric == value) + if (q->type == type && q->kind == CCL_RPN_ATTR_NUMERIC) { - if (attset) - *attset = q->set; - return 1; + got_type = 1; + if (q->value.numeric == value) + { + if (attset) + *attset = q->set; + return 1; + } } - q = q->next; } + if (got_type) + return 0; } return 0; } @@ -214,6 +218,24 @@ void ccl_add_attr_numeric(struct ccl_rpn_node *p, const char *set, n->value.numeric = value; } +void ccl_set_attr_numeric(struct ccl_rpn_node *p, const char *set, + int type, int value) +{ + struct ccl_rpn_attr *n; + for (n = p->u.t.attr_list; n; n = n->next) + if (n->type == type) + { + xfree(n->set); + n->set = set ? xstrdup(set) : 0; + if (n->kind == CCL_RPN_ATTR_STRING) + xfree(n->value.str); + n->kind = CCL_RPN_ATTR_NUMERIC; + n->value.numeric = value; + return; + } + ccl_add_attr_numeric(p, set, type, value); +} + void ccl_add_attr_string(struct ccl_rpn_node *p, const char *set, int type, char *value) { @@ -347,11 +369,7 @@ static struct ccl_rpn_node *ccl_term_one_use(CCL_parser cclp, { struct ccl_rpn_node *p; size_t i; - int relation_value = -1; - int position_value = -1; int structure_value = -1; - int truncation_value = -1; - int completeness_value = -1; int left_trunc = 0; int right_trunc = 0; @@ -408,7 +426,9 @@ static struct ccl_rpn_node *ccl_term_one_use(CCL_parser cclp, { struct ccl_rpn_attr *attr; for (attr = ccl_qual_get_attr(qa[i]); attr; attr = attr->next) - if (attr->type != 1 || !attr_use || attr == attr_use) + if (attr->type == 1 && attr_use && attr != attr_use) + continue; + else { switch (attr->kind) { @@ -421,31 +441,11 @@ static struct ccl_rpn_node *ccl_term_one_use(CCL_parser cclp, { /* deal only with REAL attributes (positive) */ switch (attr->type) { - case CCL_BIB1_REL: - if (relation_value != -1) - continue; - relation_value = attr->value.numeric; - break; - case CCL_BIB1_POS: - if (position_value != -1) - continue; - position_value = attr->value.numeric; - break; case CCL_BIB1_STR: if (structure_value != -1) continue; structure_value = attr->value.numeric; break; - case CCL_BIB1_TRU: - if (truncation_value != -1) - continue; - truncation_value = attr->value.numeric; - break; - case CCL_BIB1_COM: - if (completeness_value != -1) - continue; - completeness_value = attr->value.numeric; - break; } ccl_add_attr_numeric(p, attr->set, attr->type, attr->value.numeric); @@ -586,12 +586,13 @@ static struct ccl_rpn_node *ccl_term_multi_use(CCL_parser cclp, } static struct ccl_rpn_node *split_recur(CCL_parser cclp, ccl_qualifier_t *qa, - struct ccl_token **ar, size_t sz) + struct ccl_token **ar, size_t sz, + size_t sub_len) { size_t l; struct ccl_rpn_node *p_top = 0; assert(sz > 0); - for (l = 1; l <= sz; l++) + for (l = 1; l <= sz && l <= sub_len; l++) { struct ccl_rpn_node *p2 = ccl_term_multi_use(cclp, ar[0], qa, l, @@ -604,7 +605,8 @@ static struct ccl_rpn_node *split_recur(CCL_parser cclp, ccl_qualifier_t *qa, } if (sz > l) { - struct ccl_rpn_node *p1 = split_recur(cclp, qa, ar + l, sz - l); + struct ccl_rpn_node *p1 = split_recur(cclp, qa, ar + l, sz - l, + sub_len); if (!p1) { ccl_rpn_delete(p2); @@ -625,7 +627,7 @@ static struct ccl_rpn_node *search_term_split_list(CCL_parser cclp, struct ccl_rpn_node *p; struct ccl_token **ar; struct ccl_token *lookahead = cclp->look_token; - size_t i, sz; + size_t i, sz, sub_len; for (sz = 0; is_term_ok(lookahead->kind, term_list); sz++) lookahead = lookahead->next; if (sz == 0) @@ -640,7 +642,14 @@ static struct ccl_rpn_node *search_term_split_list(CCL_parser cclp, ar[i] = lookahead; lookahead = lookahead->next; } - p = split_recur(cclp, qa, ar, sz); + /* choose sub phrase carefully to avoid huge expansions */ + if (sz >= 7) + sub_len = 1; + else if (sz >= 5) + sub_len = 2; + else + sub_len = 3; + p = split_recur(cclp, qa, ar, sz, sub_len); xfree(ar); for (i = 0; i < sz; i++) ADVANCE; @@ -887,14 +896,14 @@ struct ccl_rpn_node *qualifiers_order(CCL_parser cclp, } p = ccl_rpn_node_create(CCL_RPN_AND); p->u.p[0] = p1; - ccl_add_attr_numeric(p1, attset, CCL_BIB1_REL, 4); + ccl_set_attr_numeric(p1, attset, CCL_BIB1_REL, 4); p->u.p[1] = p2; - ccl_add_attr_numeric(p2, attset, CCL_BIB1_REL, 2); + ccl_set_attr_numeric(p2, attset, CCL_BIB1_REL, 2); return p; } else /* = term - */ { - ccl_add_attr_numeric(p1, attset, CCL_BIB1_REL, 4); + ccl_set_attr_numeric(p1, attset, CCL_BIB1_REL, 4); return p1; } } @@ -905,7 +914,7 @@ struct ccl_rpn_node *qualifiers_order(CCL_parser cclp, ADVANCE; if (!(p = search_term(cclp, ap))) return NULL; - ccl_add_attr_numeric(p, attset, CCL_BIB1_REL, 2); + ccl_set_attr_numeric(p, attset, CCL_BIB1_REL, 2); return p; } else @@ -914,7 +923,7 @@ struct ccl_rpn_node *qualifiers_order(CCL_parser cclp, return NULL; if (rel != 3 || !qual_val_type(ap, CCL_BIB1_REL, CCL_BIB1_REL_OMIT_EQUALS, 0)) - ccl_add_attr_numeric(p, attset, CCL_BIB1_REL, rel); + ccl_set_attr_numeric(p, attset, CCL_BIB1_REL, rel); return p; } return NULL;