X-Git-Url: http://git.indexdata.com/?p=yaz-moved-to-github.git;a=blobdiff_plain;f=src%2Fcclfind.c;h=f4dbaa4b9debbcb1b6da14ee53ebf3b4c60e4059;hp=93ea05da6b5943f26b29c8ad0d962230b40500ba;hb=ba058997e52b990c7fa140301a1970be96ab6a4b;hpb=0c46d2e66bdeea1600e700124a81a5d0a65d349e diff --git a/src/cclfind.c b/src/cclfind.c index 93ea05d..f4dbaa4 100644 --- a/src/cclfind.c +++ b/src/cclfind.c @@ -1,5 +1,5 @@ /* This file is part of the YAZ toolkit. - * Copyright (C) 1995-2013 Index Data + * Copyright (C) Index Data * See the file LICENSE for details. */ /** @@ -229,9 +229,38 @@ static size_t cmp_operator(const char **aliases, const char *input) #define REGEX_CHARS "^[]{}()|.*+?!$" #define CCL_CHARS "#?\\" + +static int has_ccl_masking(const char *src_str, + size_t src_len, + const char **truncation_aliases, + const char **mask_aliases) +{ + size_t j; + int quote_mode = 0; + + for (j = 0; j < src_len; j++) + { + size_t op_size; + if (j > 0 && src_str[j-1] == '\\') + ; + else if (src_str[j] == '"') + quote_mode = !quote_mode; + else if (!quote_mode && + (op_size = cmp_operator(truncation_aliases, + src_str + j))) + return 1; + else if (!quote_mode && + (op_size = cmp_operator(mask_aliases, + src_str + j))) + return 1; + } + return 0; +} + static int append_term(CCL_parser cclp, const char *src_str, size_t src_len, - char *dst_term, int *regex_trunc, int *z3958_trunc, + char *dst_term, int regex_trunc, int z3958_trunc, const char **truncation_aliases, + const char **mask_aliases, int is_first, int is_last, int *left_trunc, int *right_trunc) { @@ -243,16 +272,10 @@ static int append_term(CCL_parser cclp, const char *src_str, size_t src_len, size_t op_size; if (j > 0 && src_str[j-1] == '\\') { - if (*regex_trunc && strchr(REGEX_CHARS "\\", src_str[j])) - { - *regex_trunc = 2; + if (regex_trunc && strchr(REGEX_CHARS "\\", src_str[j])) strcat(dst_term, "\\"); - } - else if (*z3958_trunc && strchr(CCL_CHARS "\\", src_str[j])) - { - *z3958_trunc = 2; + else if (z3958_trunc && strchr(CCL_CHARS "\\", src_str[j])) strcat(dst_term, "\\"); - } strxcat(dst_term, src_str + j, 1); } else if (src_str[j] == '"') @@ -263,16 +286,10 @@ static int append_term(CCL_parser cclp, const char *src_str, size_t src_len, ) { j += (op_size - 1); /* j++ in for loop */ - if (*regex_trunc) - { + if (regex_trunc) strcat(dst_term, ".*"); - *regex_trunc = 2; /* regex trunc is really needed */ - } - else if (*z3958_trunc) - { + else if (z3958_trunc) strcat(dst_term, "?"); - *z3958_trunc = 2; - } else if (is_first && j == 0) *left_trunc = 1; else if (is_last && j == src_len - 1) @@ -283,18 +300,14 @@ static int append_term(CCL_parser cclp, const char *src_str, size_t src_len, return -1; } } - else if (!quote_mode && src_str[j] == '#') + else if (!quote_mode && + (op_size = cmp_operator(mask_aliases, src_str + j))) { - if (*regex_trunc) - { + j += (op_size - 1); /* j++ in for loop */ + if (regex_trunc) strcat(dst_term, "."); - *regex_trunc = 2; /* regex trunc is really needed */ - } - else if (*z3958_trunc) - { + else if (z3958_trunc) strcat(dst_term, "#"); - *z3958_trunc = 2; - } else { cclp->error_code = CCL_ERR_TRUNC_NOT_SINGLE; @@ -303,42 +316,43 @@ static int append_term(CCL_parser cclp, const char *src_str, size_t src_len, } else if (src_str[j] != '\\') { - if (*regex_trunc && strchr(REGEX_CHARS, src_str[j])) - { - *regex_trunc = 2; + if (regex_trunc && strchr(REGEX_CHARS, src_str[j])) strcat(dst_term, "\\"); - } - else if (*z3958_trunc && strchr(CCL_CHARS, src_str[j])) - { - *z3958_trunc = 2; + else if (z3958_trunc && strchr(CCL_CHARS, src_str[j])) strcat(dst_term, "\\"); - } strxcat(dst_term, src_str + j, 1); } } return 0; } -/** - * search_term: Parse CCL search term. - * cclp: CCL Parser - * qa: Qualifier attributes already applied. - * term_list: tokens we accept as terms in context - * multi: whether we accept "multiple" tokens - * return: pointer to node(s); NULL on error. - */ -static struct ccl_rpn_node *search_term_x(CCL_parser cclp, - ccl_qualifier_t *qa, - int *term_list, int multi) + +static struct ccl_rpn_node *ccl_term_one_use(CCL_parser cclp, + struct ccl_rpn_attr *attr_use, + ccl_qualifier_t *qa, + size_t no, int term_len, + int is_phrase, + int auto_group) { - struct ccl_rpn_node *p_top = 0; - struct ccl_token *lookahead = cclp->look_token; - int and_list = 0; - int auto_group = 0; - int or_list = 0; + 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; + int regex_trunc = 0; + int z3958_trunc = 0; + int is_ccl_masked = 0; char *attset; + struct ccl_token *lookahead = cclp->look_token; const char **truncation_aliases; const char *t_default[2]; + const char **mask_aliases; + const char *m_default[2]; truncation_aliases = ccl_qual_search_special(cclp->bibset, "truncation"); @@ -349,78 +363,43 @@ static struct ccl_rpn_node *search_term_x(CCL_parser cclp, t_default[1] = 0; } - if (qual_val_type(qa, CCL_BIB1_STR, CCL_BIB1_STR_AND_LIST, 0)) - and_list = 1; - if (qual_val_type(qa, CCL_BIB1_STR, CCL_BIB1_STR_AUTO_GROUP, 0)) - auto_group = 1; - if (qual_val_type(qa, CCL_BIB1_STR, CCL_BIB1_STR_OR_LIST, 0)) - or_list = 1; - while (1) + mask_aliases = + ccl_qual_search_special(cclp->bibset, "mask"); + if (!mask_aliases) { - struct ccl_rpn_node *p; - size_t no, i; - int is_phrase = 0; - int relation_value = -1; - int position_value = -1; - int structure_value = -1; - int truncation_value = -1; - int completeness_value = -1; - int len = 0; - int left_trunc = 0; - int right_trunc = 0; - int regex_trunc = 0; - int z3958_trunc = 0; - size_t max = 200; - if (and_list || or_list || !multi) - max = 1; - - /* ignore commas when dealing with and-lists .. */ - if (and_list && lookahead && lookahead->kind == CCL_TOK_COMMA) - { - lookahead = lookahead->next; - ADVANCE; - continue; - } - for (no = 0; no < max && is_term_ok(lookahead->kind, term_list); no++) - { - int this_is_phrase = 0; - for (i = 0; ilen; i++) - if (lookahead->name[i] == ' ') - this_is_phrase = 1; - - if (auto_group) - { - if (no > 0 && (is_phrase || is_phrase != this_is_phrase)) - break; - is_phrase = this_is_phrase; - } - else if (this_is_phrase || no > 0) - is_phrase = 1; - len += 1+lookahead->len+lookahead->ws_prefix_len; - lookahead = lookahead->next; - } - - if (len == 0) - break; /* no more terms . stop . */ + mask_aliases = m_default; + m_default[0] = "#"; + m_default[1] = 0; + } - /* create the term node, but wait a moment before adding the term */ - p = ccl_rpn_node_create(CCL_RPN_TERM); - p->u.t.attr_list = NULL; - p->u.t.term = NULL; - if (qa && qa[0]) - { - const char *n = ccl_qual_get_name(qa[0]); - if (n) - p->u.t.qual = xstrdup(n); - } - /* go through all attributes and add them to the attribute list */ - for (i=0; qa && qa[i]; i++) - { - struct ccl_rpn_attr *attr; + for (i = 0; i < no; i++) + { + if (has_ccl_masking(lookahead->name, lookahead->len, + truncation_aliases, + mask_aliases)) + is_ccl_masked = 1; + lookahead = lookahead->next; + } + lookahead = cclp->look_token; - for (attr = ccl_qual_get_attr(qa[i]); attr; attr = attr->next) - switch(attr->kind) + p = ccl_rpn_node_create(CCL_RPN_TERM); + p->u.t.attr_list = NULL; + p->u.t.term = NULL; + if (qa && qa[0]) + { + const char *n = ccl_qual_get_name(qa[0]); + if (n) + p->u.t.qual = xstrdup(n); + } + /* go through all attributes and add them to the attribute list */ + for (i = 0; qa && qa[i]; i++) + { + struct ccl_rpn_attr *attr; + for (attr = ccl_qual_get_attr(qa[i]); attr; attr = attr->next) + if (attr->type != 1 || attr == attr_use) + { + switch (attr->kind) { case CCL_RPN_ATTR_STRING: ccl_add_attr_string(p, attr->set, attr->type, @@ -461,53 +440,204 @@ static struct ccl_rpn_node *search_term_x(CCL_parser cclp, attr->value.numeric); } } + } + } + attset = 0; + if (structure_value == -1 && ( + auto_group || + qual_val_type(qa, CCL_BIB1_STR, CCL_BIB1_STR_WP, &attset)) + ) + { + if (!is_phrase) + ccl_add_attr_numeric(p, attset, CCL_BIB1_STR, 2); + else + ccl_add_attr_numeric(p, attset, CCL_BIB1_STR, 1); + } + if (qual_val_type(qa, CCL_BIB1_TRU, CCL_BIB1_TRU_CAN_REGEX, + &attset)) + { + if (is_ccl_masked) + regex_trunc = 1; /* regex trunc (102) allowed */ + } + else if (qual_val_type(qa, CCL_BIB1_TRU, CCL_BIB1_TRU_CAN_Z3958, + &attset)) + { + if (is_ccl_masked) + z3958_trunc = 1; /* Z39.58 trunc (CCL) trunc allowed */ + } + /* make the RPN token */ + p->u.t.term = (char *)xmalloc(term_len * 2 + 2); + ccl_assert(p->u.t.term); + p->u.t.term[0] = '\0'; + + for (i = 0; i < no; i++) + { + const char *src_str = lookahead->name; + size_t src_len = lookahead->len; + + if (p->u.t.term[0] && lookahead->ws_prefix_len) + { + strxcat(p->u.t.term, lookahead->ws_prefix_buf, + lookahead->ws_prefix_len); } - attset = 0; - if (structure_value == -1 && ( - auto_group || - qual_val_type(qa, CCL_BIB1_STR, CCL_BIB1_STR_WP, &attset)) - ) + if (append_term(cclp, src_str, src_len, p->u.t.term, regex_trunc, + z3958_trunc, truncation_aliases, mask_aliases, + i == 0, i == no - 1, + &left_trunc, &right_trunc)) { - if (!is_phrase) - ccl_add_attr_numeric(p, attset, CCL_BIB1_STR, 2); - else - ccl_add_attr_numeric(p, attset, CCL_BIB1_STR, 1); + ccl_rpn_delete(p); + return NULL; } - - if (qual_val_type(qa, CCL_BIB1_TRU, CCL_BIB1_TRU_CAN_REGEX, - &attset)) + lookahead = lookahead->next; + } + if (left_trunc && right_trunc) + { + if (!qual_val_type(qa, CCL_BIB1_TRU, CCL_BIB1_TRU_CAN_BOTH, + &attset)) { - regex_trunc = 1; /* regex trunc (102) allowed */ + cclp->error_code = CCL_ERR_TRUNC_NOT_BOTH; + ccl_rpn_delete(p); + return NULL; } - else if (qual_val_type(qa, CCL_BIB1_TRU, CCL_BIB1_TRU_CAN_Z3958, - &attset)) + ccl_add_attr_numeric(p, attset, CCL_BIB1_TRU, 3); + } + else if (right_trunc) + { + if (!qual_val_type(qa, CCL_BIB1_TRU, CCL_BIB1_TRU_CAN_RIGHT, + &attset)) { - z3958_trunc = 1; /* Z39.58 trunc (CCL) trunc allowed */ + cclp->error_code = CCL_ERR_TRUNC_NOT_RIGHT; + ccl_rpn_delete(p); + return NULL; } - - /* make the RPN token */ - p->u.t.term = (char *)xmalloc(len * 2 + 2); - ccl_assert(p->u.t.term); - p->u.t.term[0] = '\0'; - for (i = 0; ilook_token->name; - size_t src_len = cclp->look_token->len; + cclp->error_code = CCL_ERR_TRUNC_NOT_LEFT; + ccl_rpn_delete(p); + return NULL; + } + ccl_add_attr_numeric(p, attset, CCL_BIB1_TRU, 2); + } + else if (regex_trunc) + { + ccl_add_attr_numeric(p, attset, CCL_BIB1_TRU, 102); + } + else if (z3958_trunc) + { + ccl_add_attr_numeric(p, attset, CCL_BIB1_TRU, 104); + } + else + { + if (qual_val_type(qa, CCL_BIB1_TRU, CCL_BIB1_TRU_CAN_NONE, + &attset)) + ccl_add_attr_numeric(p, attset, CCL_BIB1_TRU, 100); + } + return p; +} - if (p->u.t.term[0] && cclp->look_token->ws_prefix_len) - { - strxcat(p->u.t.term, cclp->look_token->ws_prefix_buf, - cclp->look_token->ws_prefix_len); - } - if (append_term(cclp, src_str, src_len, p->u.t.term, ®ex_trunc, - &z3958_trunc, truncation_aliases, i == 0, i == no - 1, - &left_trunc, &right_trunc)) +/** + * search_term: Parse CCL search term. + * cclp: CCL Parser + * qa: Qualifier attributes already applied. + * term_list: tokens we accept as terms in context + * multi: whether we accept "multiple" tokens + * return: pointer to node(s); NULL on error. + */ +static struct ccl_rpn_node *search_term_x(CCL_parser cclp, + ccl_qualifier_t *qa, + int *term_list, int multi) +{ + struct ccl_rpn_node *p_top = 0; + struct ccl_token *lookahead = cclp->look_token; + int and_list = 0; + int auto_group = 0; + int or_list = 0; + + if (qual_val_type(qa, CCL_BIB1_STR, CCL_BIB1_STR_AND_LIST, 0)) + and_list = 1; + if (qual_val_type(qa, CCL_BIB1_STR, CCL_BIB1_STR_AUTO_GROUP, 0)) + auto_group = 1; + if (qual_val_type(qa, CCL_BIB1_STR, CCL_BIB1_STR_OR_LIST, 0)) + or_list = 1; + while (1) + { + struct ccl_rpn_node *p = 0; + size_t no, i; + int len = 0; + int is_phrase = 0; + size_t max = 200; + if (and_list || or_list || !multi) + max = 1; + + /* ignore commas when dealing with and-lists .. */ + if (and_list && lookahead && lookahead->kind == CCL_TOK_COMMA) + { + lookahead = lookahead->next; + ADVANCE; + continue; + } + for (no = 0; no < max && is_term_ok(lookahead->kind, term_list); no++) + { + int this_is_phrase = 0; + for (i = 0; ilen; i++) + if (lookahead->name[i] == ' ') + this_is_phrase = 1; + if (auto_group) { - ccl_rpn_delete(p); - return NULL; + if (no > 0 && (is_phrase || is_phrase != this_is_phrase)) + break; + is_phrase = this_is_phrase; } - ADVANCE; + else if (this_is_phrase || no > 0) + is_phrase = 1; + len += 1+lookahead->len+lookahead->ws_prefix_len; + lookahead = lookahead->next; } + + if (len == 0) + break; /* no more terms . stop . */ + + /* go through all attributes and add them to the attribute list */ + for (i = 0; qa && qa[i]; i++) + { + struct ccl_rpn_attr *attr; + + for (attr = ccl_qual_get_attr(qa[i]); attr; attr = attr->next) + if (attr->type == 1) + { + struct ccl_rpn_node *tmp2; + tmp2 = ccl_term_one_use(cclp, attr, qa, no, len, + is_phrase, + auto_group); + if (!tmp2) + { + ccl_rpn_delete(p); + return 0; + } + if (!p) + p = tmp2; + else + { + struct ccl_rpn_node *tmp1; + tmp1 = ccl_rpn_node_create(CCL_RPN_OR); + tmp1->u.p[0] = p; + tmp1->u.p[1] = tmp2; + p = tmp1; + } + } + } + if (!p) + p = ccl_term_one_use(cclp, 0 /* attr: no use */, qa, no, len, + is_phrase, auto_group); + for (i = 0; i < no; i++) + ADVANCE; + if (!p) + return 0; /* make the top node point to us.. */ if (p_top) { @@ -527,54 +657,6 @@ static struct ccl_rpn_node *search_term_x(CCL_parser cclp, else p_top = p; - - if (left_trunc && right_trunc) - { - if (!qual_val_type(qa, CCL_BIB1_TRU, CCL_BIB1_TRU_CAN_BOTH, - &attset)) - { - cclp->error_code = CCL_ERR_TRUNC_NOT_BOTH; - ccl_rpn_delete(p); - return NULL; - } - ccl_add_attr_numeric(p, attset, CCL_BIB1_TRU, 3); - } - else if (right_trunc) - { - if (!qual_val_type(qa, CCL_BIB1_TRU, CCL_BIB1_TRU_CAN_RIGHT, - &attset)) - { - cclp->error_code = CCL_ERR_TRUNC_NOT_RIGHT; - ccl_rpn_delete(p); - return NULL; - } - ccl_add_attr_numeric(p, attset, CCL_BIB1_TRU, 1); - } - else if (left_trunc) - { - if (!qual_val_type(qa, CCL_BIB1_TRU, CCL_BIB1_TRU_CAN_LEFT, - &attset)) - { - cclp->error_code = CCL_ERR_TRUNC_NOT_LEFT; - ccl_rpn_delete(p); - return NULL; - } - ccl_add_attr_numeric(p, attset, CCL_BIB1_TRU, 2); - } - else if (regex_trunc == 2) - { - ccl_add_attr_numeric(p, attset, CCL_BIB1_TRU, 102); - } - else if (z3958_trunc == 2) - { - ccl_add_attr_numeric(p, attset, CCL_BIB1_TRU, 104); - } - else - { - if (qual_val_type(qa, CCL_BIB1_TRU, CCL_BIB1_TRU_CAN_NONE, - &attset)) - ccl_add_attr_numeric(p, attset, CCL_BIB1_TRU, 100); - } if (!multi) break; } @@ -659,9 +741,14 @@ struct ccl_rpn_node *qualifiers_order(CCL_parser cclp, if (KIND == CCL_TOK_TERM) { size_t i; + int quote_mode = 0; for (i = 0; ilook_token->len; i++) { - if (cclp->look_token->name[i] == '-') + if (i > 0 && cclp->look_token->name[i] == '\\') + ; + else if (cclp->look_token->name[i] == '"') + quote_mode = !quote_mode; + else if (cclp->look_token->name[i] == '-' && !quote_mode) break; } @@ -766,10 +853,11 @@ struct ccl_rpn_node *qualifiers_order(CCL_parser cclp, { if (!(p = search_terms(cclp, ap))) return NULL; - ccl_add_attr_numeric(p, attset, CCL_BIB1_REL, rel); + 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); return p; } - cclp->error_code = CCL_ERR_TERM_EXPECTED; return NULL; }