X-Git-Url: http://git.indexdata.com/?p=idzebra-moved-to-github.git;a=blobdiff_plain;f=index%2Fzrpn.c;h=0f1ef439480898de80ee0568014ffb0f3f90029d;hp=bb8dc7a24358cd8ef6f80c558248028b45588689;hb=49d0ee122a9f86ec2967b577dcc297c501785edd;hpb=5e7516d00451ada59a311fad05e8e3959242cdc7 diff --git a/index/zrpn.c b/index/zrpn.c index bb8dc7a..0f1ef43 100644 --- a/index/zrpn.c +++ b/index/zrpn.c @@ -1,10 +1,26 @@ -/* - * Copyright (C) 1995-2002, Index Data - * All rights reserved. - * Sebastian Hammer, Adam Dickmeiss - * - * $Id: zrpn.c,v 1.114 2002-04-12 14:55:22 adam Exp $ - */ +/* $Id: zrpn.c,v 1.175 2005-04-15 10:47:48 adam Exp $ + Copyright (C) 1995-2005 + Index Data ApS + +This file is part of the Zebra server. + +Zebra is free software; you can redistribute it and/or modify it under +the terms of the GNU General Public License as published by the Free +Software Foundation; either version 2, or (at your option) any later +version. + +Zebra is distributed in the hope that it will be useful, but WITHOUT ANY +WARRANTY; without even the implied warranty of MERCHANTABILITY or +FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License +for more details. + +You should have received a copy of the GNU General Public License +along with Zebra; see the file LICENSE.zebra. If not, write to the +Free Software Foundation, 59 Temple Place - Suite 330, Boston, MA +02111-1307, USA. +*/ + + #include #include #ifdef WIN32 @@ -15,58 +31,82 @@ #include #include "index.h" +#include #include -#include -#include -#include -#include +#include + +/* maximum number of terms in an and/or/phrase item */ +#define TERM_LIST_LENGTH_MAX 256 + +static const struct key_control it_ctrl = +{ + sizeof(struct it_key), + 2, /* we have sysnos and seqnos in this key, nothing more */ + key_compare_it, + key_logdump_txt, /* FIXME - clean up these functions */ + key_get_seq, +}; + -struct rpn_char_map_info { +const struct key_control *key_it_ctrl = &it_ctrl; + +struct rpn_char_map_info +{ ZebraMaps zm; int reg_type; }; -static const char **rpn_char_map_handler (void *vp, const char **from, int len) +typedef struct +{ + int type; + int major; + int minor; + Z_AttributesPlusTerm *zapt; +} AttrType; + + +static int log_level_set = 0; +static int log_level_rpn = 0; + +static const char **rpn_char_map_handler(void *vp, const char **from, int len) { struct rpn_char_map_info *p = (struct rpn_char_map_info *) vp; - return zebra_maps_input (p->zm, p->reg_type, from, len); + const char **out = zebra_maps_input(p->zm, p->reg_type, from, len, 0); +#if 0 + if (out && *out) + { + const char *outp = *out; + yaz_log(YLOG_LOG, "---"); + while (*outp) + { + yaz_log(YLOG_LOG, "%02X", *outp); + outp++; + } + } +#endif + return out; } -static void rpn_char_map_prepare (struct zebra_register *reg, int reg_type, - struct rpn_char_map_info *map_info) +static void rpn_char_map_prepare(struct zebra_register *reg, int reg_type, + struct rpn_char_map_info *map_info) { map_info->zm = reg->zebra_maps; map_info->reg_type = reg_type; - dict_grep_cmap (reg->dict, map_info, rpn_char_map_handler); + dict_grep_cmap(reg->dict, map_info, rpn_char_map_handler); } -typedef struct { - int type; - int major; - int minor; - Z_AttributesPlusTerm *zapt; -} AttrType; - -static int attr_find_ex (AttrType *src, oid_value *attributeSetP, - const char **string_value) +static int attr_find_ex(AttrType *src, oid_value *attributeSetP, + const char **string_value) { int num_attributes; -#ifdef ASN_COMPILED num_attributes = src->zapt->attributes->num_attributes; -#else - num_attributes = src->zapt->num_attributes; -#endif while (src->major < num_attributes) { Z_AttributeElement *element; -#ifdef ASN_COMPILED element = src->zapt->attributes->attributes[src->major]; -#else - element = src->zapt->attributeList[src->major]; -#endif if (src->type == *element->attributeType) { switch (element->which) @@ -77,42 +117,42 @@ static int attr_find_ex (AttrType *src, oid_value *attributeSetP, { oident *attrset; - attrset = oid_getentbyoid (element->attributeSet); + attrset = oid_getentbyoid(element->attributeSet); *attributeSetP = attrset->value; } return *element->value.numeric; break; case Z_AttributeValue_complex: if (src->minor >= element->value.complex->num_list) - break; + break; + if (element->attributeSet && attributeSetP) + { + oident *attrset; + + attrset = oid_getentbyoid(element->attributeSet); + *attributeSetP = attrset->value; + } if (element->value.complex->list[src->minor]->which == Z_StringOrNumeric_numeric) - { - ++(src->minor); - if (element->attributeSet && attributeSetP) - { - oident *attrset; - - attrset = oid_getentbyoid (element->attributeSet); - *attributeSetP = attrset->value; - } - return - *element->value.complex->list[src->minor-1]->u.numeric; - } - else if (element->value.complex->list[src->minor]->which == - Z_StringOrNumeric_string) - { - if (!string_value) - break; - ++(src->minor); - *string_value = - element->value.complex->list[src->minor-1]->u.string; - return -2; - } - else - break; + { + ++(src->minor); + return + *element->value.complex->list[src->minor-1]->u.numeric; + } + else if (element->value.complex->list[src->minor]->which == + Z_StringOrNumeric_string) + { + if (!string_value) + break; + ++(src->minor); + *string_value = + element->value.complex->list[src->minor-1]->u.string; + return -2; + } + else + break; default: - assert (0); + assert(0); } } ++(src->major); @@ -120,12 +160,12 @@ static int attr_find_ex (AttrType *src, oid_value *attributeSetP, return -1; } -static int attr_find (AttrType *src, oid_value *attributeSetP) +static int attr_find(AttrType *src, oid_value *attributeSetP) { - return attr_find_ex (src, attributeSetP, 0); + return attr_find_ex(src, attributeSetP, 0); } -static void attr_init (AttrType *src, Z_AttributesPlusTerm *zapt, +static void attr_init(AttrType *src, Z_AttributesPlusTerm *zapt, int type) { src->zapt = zapt; @@ -140,7 +180,7 @@ struct grep_info { #ifdef TERM_COUNT int *term_no; #endif - ISAMS_P *isam_p_buf; + ISAM_P *isam_p_buf; int isam_p_size; int isam_p_indx; ZebraHandle zh; @@ -148,87 +188,92 @@ struct grep_info { ZebraSet termset; }; -static void term_untrans (ZebraHandle zh, int reg_type, - char *dst, const char *src) +static void term_untrans(ZebraHandle zh, int reg_type, + char *dst, const char *src) { + int len = 0; while (*src) { - const char *cp = zebra_maps_output (zh->reg->zebra_maps, - reg_type, &src); - if (!cp) - *dst++ = *src++; - else - while (*cp) - *dst++ = *cp++; + const char *cp = zebra_maps_output(zh->reg->zebra_maps, + reg_type, &src); + if (!cp && len < IT_MAX_WORD-1) + dst[len++] = *src++; + else + while (*cp && len < IT_MAX_WORD-1) + dst[len++] = *cp++; } - *dst = '\0'; + dst[len] = '\0'; } -static void add_isam_p (const char *name, const char *info, - struct grep_info *p) +static void add_isam_p(const char *name, const char *info, + struct grep_info *p) { + if (!log_level_set) + { + log_level_rpn = yaz_log_module_level("rpn"); + log_level_set = 1; + } if (p->isam_p_indx == p->isam_p_size) { - ISAMS_P *new_isam_p_buf; + ISAM_P *new_isam_p_buf; #ifdef TERM_COUNT int *new_term_no; #endif p->isam_p_size = 2*p->isam_p_size + 100; - new_isam_p_buf = (ISAMS_P *) xmalloc (sizeof(*new_isam_p_buf) * - p->isam_p_size); + new_isam_p_buf = (ISAM_P *) xmalloc(sizeof(*new_isam_p_buf) * + p->isam_p_size); if (p->isam_p_buf) { - memcpy (new_isam_p_buf, p->isam_p_buf, + memcpy(new_isam_p_buf, p->isam_p_buf, p->isam_p_indx * sizeof(*p->isam_p_buf)); - xfree (p->isam_p_buf); + xfree(p->isam_p_buf); } p->isam_p_buf = new_isam_p_buf; #ifdef TERM_COUNT - new_term_no = (int *) xmalloc (sizeof(*new_term_no) * - p->isam_p_size); + new_term_no = (int *) xmalloc(sizeof(*new_term_no) * p->isam_p_size); if (p->term_no) { - memcpy (new_term_no, p->isam_p_buf, + memcpy(new_term_no, p->isam_p_buf, p->isam_p_indx * sizeof(*p->term_no)); - xfree (p->term_no); + xfree(p->term_no); } p->term_no = new_term_no; #endif } - assert (*info == sizeof(*p->isam_p_buf)); - memcpy (p->isam_p_buf + p->isam_p_indx, info+1, sizeof(*p->isam_p_buf)); + assert(*info == sizeof(*p->isam_p_buf)); + memcpy(p->isam_p_buf + p->isam_p_indx, info+1, sizeof(*p->isam_p_buf)); #if 1 if (p->termset) { - const char *db; - int set, use; - char term_tmp[512]; - int su_code = 0; - int len = key_SU_decode (&su_code, name); - - term_untrans (p->zh, p->reg_type, term_tmp, name+len+1); - logf (LOG_LOG, "grep: %d %c %s", su_code, name[len], term_tmp); - zebraExplain_lookup_ord (p->zh->reg->zei, - su_code, &db, &set, &use); - logf (LOG_LOG, "grep: set=%d use=%d db=%s", set, use, db); - - resultSetAddTerm (p->zh, p->termset, name[len], db, - set, use, term_tmp); + const char *db; + int set, use; + char term_tmp[IT_MAX_WORD]; + int su_code = 0; + int len = key_SU_decode (&su_code, name); + + term_untrans (p->zh, p->reg_type, term_tmp, name+len+1); + yaz_log(log_level_rpn, "grep: %d %c %s", su_code, name[len], term_tmp); + zebraExplain_lookup_ord (p->zh->reg->zei, + su_code, &db, &set, &use); + yaz_log(log_level_rpn, "grep: set=%d use=%d db=%s", set, use, db); + + resultSetAddTerm(p->zh, p->termset, name[len], db, + set, use, term_tmp); } #endif (p->isam_p_indx)++; } -static int grep_handle (char *name, const char *info, void *p) +static int grep_handle(char *name, const char *info, void *p) { - add_isam_p (name, info, (struct grep_info *) p); + add_isam_p(name, info, (struct grep_info *) p); return 0; } -static int term_pre (ZebraMaps zebra_maps, int reg_type, const char **src, - const char *ct1, const char *ct2) +static int term_pre(ZebraMaps zebra_maps, int reg_type, const char **src, + const char *ct1, const char *ct2, int first) { const char *s1, *s0 = *src; const char **map; @@ -236,12 +281,12 @@ static int term_pre (ZebraMaps zebra_maps, int reg_type, const char **src, /* skip white space */ while (*s0) { - if (ct1 && strchr (ct1, *s0)) + if (ct1 && strchr(ct1, *s0)) break; - if (ct2 && strchr (ct2, *s0)) + if (ct2 && strchr(ct2, *s0)) break; s1 = s0; - map = zebra_maps_input (zebra_maps, reg_type, &s1, strlen(s1)); + map = zebra_maps_input(zebra_maps, reg_type, &s1, strlen(s1), first); if (**map != *CHR_SPACE) break; s0 = s1; @@ -250,12 +295,41 @@ static int term_pre (ZebraMaps zebra_maps, int reg_type, const char **src, return *s0; } -/* term_100: handle term, where trunc=none (no operators at all) */ -static int term_100 (ZebraMaps zebra_maps, int reg_type, - const char **src, char *dst, int space_split, - char *dst_term) + +static void esc_str(char *out_buf, int out_size, + const char *in_buf, int in_size) { - const char *s0, *s1; + int k; + + assert(out_buf); + assert(in_buf); + assert(out_size > 20); + *out_buf = '\0'; + for (k = 0; k 126) + pc = '?'; + else + pc = c; + sprintf(out_buf +strlen(out_buf), "%02X:%c ", c, pc); + if (strlen(out_buf) > out_size-20) + { + strcat(out_buf, ".."); + break; + } + } +} + +#define REGEX_CHARS " []()|.*+?!" + +/* term_100: handle term, where trunc = none(no operators at all) */ +static int term_100(ZebraMaps zebra_maps, int reg_type, + const char **src, char *dst, int space_split, + char *dst_term) +{ + const char *s0; const char **map; int i = 0; int j = 0; @@ -263,47 +337,61 @@ static int term_100 (ZebraMaps zebra_maps, int reg_type, const char *space_start = 0; const char *space_end = 0; - if (!term_pre (zebra_maps, reg_type, src, NULL, NULL)) + if (!term_pre(zebra_maps, reg_type, src, NULL, NULL, !space_split)) return 0; s0 = *src; while (*s0) { - s1 = s0; - map = zebra_maps_input (zebra_maps, reg_type, &s0, strlen(s0)); - if (space_split) + const char *s1 = s0; + int q_map_match = 0; + map = zebra_maps_search(zebra_maps, reg_type, &s0, strlen(s0), + &q_map_match); + if (space_split) + { + if (**map == *CHR_SPACE) + break; + } + else /* complete subfield only. */ + { + if (**map == *CHR_SPACE) + { /* save space mapping for later .. */ + space_start = s1; + space_end = s0; + continue; + } + else if (space_start) + { /* reload last space */ + while (space_start < space_end) + { + if (strchr(REGEX_CHARS, *space_start)) + dst[i++] = '\\'; + dst_term[j++] = *space_start; + dst[i++] = *space_start++; + } + /* and reset */ + space_start = space_end = 0; + } + } + /* add non-space char */ + memcpy(dst_term+j, s1, s0 - s1); + j += (s0 - s1); + if (!q_map_match) { - if (**map == *CHR_SPACE) - break; + while (s1 < s0) + { + if (strchr(REGEX_CHARS, *s1)) + dst[i++] = '\\'; + dst[i++] = *s1++; + } } - else /* complete subfield only. */ + else { - if (**map == *CHR_SPACE) - { /* save space mapping for later .. */ - space_start = s1; - space_end = s0; - continue; - } - else if (space_start) - { /* reload last space */ - while (space_start < space_end) - { - if (!isalnum (*space_start) && *space_start != '-') - dst[i++] = '\\'; - dst_term[j++] = *space_start; - dst[i++] = *space_start++; - } - /* and reset */ - space_start = space_end = 0; - } + char tmpbuf[80]; + esc_str(tmpbuf, sizeof(tmpbuf), map[0], strlen(map[0])); + + strcpy(dst + i, map[0]); + i += strlen(map[0]); } - /* add non-space char */ - while (s1 < s0) - { - if (!isalnum (*s1) && *s1 != '-') - dst[i++] = '\\'; - dst_term[j++] = *s1; - dst[i++] = *s1++; - } } dst[i] = '\0'; dst_term[j] = '\0'; @@ -311,17 +399,17 @@ static int term_100 (ZebraMaps zebra_maps, int reg_type, return i; } -/* term_101: handle term, where trunc=Process # */ -static int term_101 (ZebraMaps zebra_maps, int reg_type, - const char **src, char *dst, int space_split, - char *dst_term) +/* term_101: handle term, where trunc = Process # */ +static int term_101(ZebraMaps zebra_maps, int reg_type, + const char **src, char *dst, int space_split, + char *dst_term) { - const char *s0, *s1; + const char *s0; const char **map; int i = 0; int j = 0; - if (!term_pre (zebra_maps, reg_type, src, "#", "#")) + if (!term_pre(zebra_maps, reg_type, src, "#", "#", !space_split)) return 0; s0 = *src; while (*s0) @@ -330,21 +418,37 @@ static int term_101 (ZebraMaps zebra_maps, int reg_type, { dst[i++] = '.'; dst[i++] = '*'; - dst_term[j++] = *s0++; + dst_term[j++] = *s0++; } else { - s1 = s0; - map = zebra_maps_input (zebra_maps, reg_type, &s0, strlen(s0)); + const char *s1 = s0; + int q_map_match = 0; + map = zebra_maps_search(zebra_maps, reg_type, &s0, strlen(s0), + &q_map_match); if (space_split && **map == *CHR_SPACE) break; - while (s1 < s0) - { - if (!isalnum (*s1)) - dst[i++] = '\\'; - dst_term[j++] = *s1; - dst[i++] = *s1++; - } + + /* add non-space char */ + memcpy(dst_term+j, s1, s0 - s1); + j += (s0 - s1); + if (!q_map_match) + { + while (s1 < s0) + { + if (strchr(REGEX_CHARS, *s1)) + dst[i++] = '\\'; + dst[i++] = *s1++; + } + } + else + { + char tmpbuf[80]; + esc_str(tmpbuf, sizeof(tmpbuf), map[0], strlen(map[0])); + + strcpy(dst + i, map[0]); + i += strlen(map[0]); + } } } dst[i] = '\0'; @@ -353,21 +457,21 @@ static int term_101 (ZebraMaps zebra_maps, int reg_type, return i; } -/* term_103: handle term, where trunc=re-2 (regular expressions) */ -static int term_103 (ZebraMaps zebra_maps, int reg_type, const char **src, - char *dst, int *errors, int space_split, - char *dst_term) +/* term_103: handle term, where trunc = re-2 (regular expressions) */ +static int term_103(ZebraMaps zebra_maps, int reg_type, const char **src, + char *dst, int *errors, int space_split, + char *dst_term) { int i = 0; int j = 0; - const char *s0, *s1; + const char *s0; const char **map; - if (!term_pre (zebra_maps, reg_type, src, "^\\()[].*+?|", "(")) + if (!term_pre(zebra_maps, reg_type, src, "^\\()[].*+?|", "(", !space_split)) return 0; s0 = *src; if (errors && *s0 == '+' && s0[1] && s0[2] == '+' && s0[3] && - isdigit (s0[1])) + isdigit(((const unsigned char *)s0)[1])) { *errors = s0[1] - '0'; s0 += 3; @@ -376,77 +480,120 @@ static int term_103 (ZebraMaps zebra_maps, int reg_type, const char **src, } while (*s0) { - if (strchr ("^\\()[].*+?|-", *s0)) - { - dst_term[j++] = *s0; + if (strchr("^\\()[].*+?|-", *s0)) + { + dst_term[j++] = *s0; dst[i++] = *s0++; - } + } else { - s1 = s0; - map = zebra_maps_input (zebra_maps, reg_type, &s0, strlen(s0)); - if (**map == *CHR_SPACE) + const char *s1 = s0; + int q_map_match = 0; + map = zebra_maps_search(zebra_maps, reg_type, &s0, strlen(s0), + &q_map_match); + if (space_split && **map == *CHR_SPACE) break; - while (s1 < s0) - { - if (!isalnum (*s1)) - dst[i++] = '\\'; - dst_term[j++] = *s1; - dst[i++] = *s1++; - } + + /* add non-space char */ + memcpy(dst_term+j, s1, s0 - s1); + j += (s0 - s1); + if (!q_map_match) + { + while (s1 < s0) + { + if (strchr(REGEX_CHARS, *s1)) + dst[i++] = '\\'; + dst[i++] = *s1++; + } + } + else + { + char tmpbuf[80]; + esc_str(tmpbuf, sizeof(tmpbuf), map[0], strlen(map[0])); + + strcpy(dst + i, map[0]); + i += strlen(map[0]); + } } } dst[i] = '\0'; dst_term[j] = '\0'; *src = s0; + return i; } -/* term_103: handle term, where trunc=re-1 (regular expressions) */ +/* term_103: handle term, where trunc = re-1 (regular expressions) */ static int term_102 (ZebraMaps zebra_maps, int reg_type, const char **src, - char *dst, int space_split, char *dst_term) + char *dst, int space_split, char *dst_term) { - return term_103 (zebra_maps, reg_type, src, dst, NULL, space_split, - dst_term); + return term_103(zebra_maps, reg_type, src, dst, NULL, space_split, + dst_term); } -/* term_104: handle term, where trunc=Process # and ! */ -static int term_104 (ZebraMaps zebra_maps, int reg_type, - const char **src, char *dst, int space_split, - char *dst_term) +/* term_104: handle term, where trunc = Process # and ! */ +static int term_104(ZebraMaps zebra_maps, int reg_type, + const char **src, char *dst, int space_split, + char *dst_term) { const char *s0, *s1; const char **map; int i = 0; int j = 0; - if (!term_pre (zebra_maps, reg_type, src, "#!", "#!")) + if (!term_pre(zebra_maps, reg_type, src, "?*#", "?*#", !space_split)) return 0; s0 = *src; while (*s0) { - if (*s0 == '#') + if (*s0 == '?') + { + dst_term[j++] = *s0++; + if (*s0 >= '0' && *s0 <= '9') + { + int limit = 0; + while (*s0 >= '0' && *s0 <= '9') + { + limit = limit * 10 + (*s0 - '0'); + dst_term[j++] = *s0++; + } + if (limit > 20) + limit = 20; + while (--limit >= 0) + { + dst[i++] = '.'; + dst[i++] = '?'; + } + } + else + { + dst[i++] = '.'; + dst[i++] = '*'; + } + } + else if (*s0 == '*') { dst[i++] = '.'; dst[i++] = '*'; - dst_term[j++] = *s0++; + dst_term[j++] = *s0++; } - else if (*s0 == '!') - { + else if (*s0 == '#') + { dst[i++] = '.'; - dst_term[j++] = *s0++; - } + dst_term[j++] = *s0++; + } + else { s1 = s0; - map = zebra_maps_input (zebra_maps, reg_type, &s0, strlen(s0)); + map = zebra_maps_input(zebra_maps, reg_type, &s0, strlen(s0), 0); if (space_split && **map == *CHR_SPACE) break; while (s1 < s0) { - if (!isalnum (*s1)) + if (strchr(REGEX_CHARS, *s1)) dst[i++] = '\\'; - dst_term[j++] = *s1; + dst_term[j++] = *s1; dst[i++] = *s1++; } } @@ -457,17 +604,17 @@ static int term_104 (ZebraMaps zebra_maps, int reg_type, return i; } -/* term_105/106: handle term, where trunc=Process * and ! and right trunc */ +/* term_105/106: handle term, where trunc = Process * and ! and right trunc */ static int term_105 (ZebraMaps zebra_maps, int reg_type, - const char **src, char *dst, int space_split, - char *dst_term, int right_truncate) + const char **src, char *dst, int space_split, + char *dst_term, int right_truncate) { const char *s0, *s1; const char **map; int i = 0; int j = 0; - if (!term_pre (zebra_maps, reg_type, src, "*!", "*!")) + if (!term_pre(zebra_maps, reg_type, src, "*!", "*!", !space_split)) return 0; s0 = *src; while (*s0) @@ -476,23 +623,24 @@ static int term_105 (ZebraMaps zebra_maps, int reg_type, { dst[i++] = '.'; dst[i++] = '*'; - dst_term[j++] = *s0++; + dst_term[j++] = *s0++; } else if (*s0 == '!') - { + { dst[i++] = '.'; - dst_term[j++] = *s0++; - } + dst_term[j++] = *s0++; + } + else { s1 = s0; - map = zebra_maps_input (zebra_maps, reg_type, &s0, strlen(s0)); + map = zebra_maps_input(zebra_maps, reg_type, &s0, strlen(s0), 0); if (space_split && **map == *CHR_SPACE) break; while (s1 < s0) { - if (!isalnum (*s1)) + if (strchr(REGEX_CHARS, *s1)) dst[i++] = '\\'; - dst_term[j++] = *s1; + dst_term[j++] = *s1; dst[i++] = *s1++; } } @@ -514,38 +662,38 @@ static int term_105 (ZebraMaps zebra_maps, int reg_type, * val: border value (inclusive) * islt: 1 if <=; 0 if >=. */ -static void gen_regular_rel (char *dst, int val, int islt) +static void gen_regular_rel(char *dst, int val, int islt) { int dst_p; int w, d, i; int pos = 0; char numstr[20]; - logf (LOG_DEBUG, "gen_regular_rel. val=%d, islt=%d", val, islt); + yaz_log(YLOG_DEBUG, "gen_regular_rel. val=%d, islt=%d", val, islt); if (val >= 0) { if (islt) - strcpy (dst, "(-[0-9]+|("); + strcpy(dst, "(-[0-9]+|("); else - strcpy (dst, "(("); + strcpy(dst, "(("); } else { if (!islt) { - strcpy (dst, "([0-9]+|-("); - dst_p = strlen (dst); + strcpy(dst, "([0-9]+|-("); + dst_p = strlen(dst); islt = 1; } else { - strcpy (dst, "(-("); + strcpy(dst, "(-("); islt = 0; } val = -val; } - dst_p = strlen (dst); - sprintf (numstr, "%d", val); + dst_p = strlen(dst); + sprintf(numstr, "%d", val); for (w = strlen(numstr); --w >= 0; pos++) { d = numstr[w]; @@ -565,7 +713,7 @@ static void gen_regular_rel (char *dst, int val, int islt) } } - strcpy (dst + dst_p, numstr); + strcpy(dst + dst_p, numstr); dst_p = strlen(dst) - pos - 1; if (islt) @@ -607,25 +755,25 @@ static void gen_regular_rel (char *dst, int val, int islt) dst[dst_p] = '\0'; if (islt) { - /* match everything less than 10^(pos-1) */ - strcat (dst, "0*"); - for (i=1; ireg->zebra_maps, reg_type, - term_sub, term_component, - space_split, term_dst)) + term_sub, term_component, + space_split, term_dst)) return 0; - logf (LOG_DEBUG, "Relation <"); - - *term_tmp++ = '('; - for (i = 0; term_component[i]; ) - { - int j = 0; + yaz_log(log_level_rpn, "Relation <"); + + *term_tmp++ = '('; + for (i = 0; term_component[i]; ) + { + int j = 0; - if (i) - *term_tmp++ = '|'; - while (j < i) - string_rel_add_char (&term_tmp, term_component, &j); + if (i) + *term_tmp++ = '|'; + while (j < i) + string_rel_add_char (&term_tmp, term_component, &j); - *term_tmp++ = '['; + *term_tmp++ = '['; - *term_tmp++ = '^'; - string_rel_add_char (&term_tmp, term_component, &i); - *term_tmp++ = '-'; + *term_tmp++ = '^'; + string_rel_add_char (&term_tmp, term_component, &i); + *term_tmp++ = '-'; - *term_tmp++ = ']'; - *term_tmp++ = '.'; - *term_tmp++ = '*'; - } - *term_tmp++ = ')'; - *term_tmp = '\0'; + *term_tmp++ = ']'; + *term_tmp++ = '.'; + *term_tmp++ = '*'; + + if ((term_tmp - term_dict) > IT_MAX_WORD) + break; + } + *term_tmp++ = ')'; + *term_tmp = '\0'; break; case 2: if (!term_100 (zh->reg->zebra_maps, reg_type, - term_sub, term_component, - space_split, term_dst)) + term_sub, term_component, + space_split, term_dst)) return 0; - logf (LOG_DEBUG, "Relation <="); + yaz_log(log_level_rpn, "Relation <="); - *term_tmp++ = '('; - for (i = 0; term_component[i]; ) - { - int j = 0; + *term_tmp++ = '('; + for (i = 0; term_component[i]; ) + { + int j = 0; - while (j < i) - string_rel_add_char (&term_tmp, term_component, &j); - *term_tmp++ = '['; + while (j < i) + string_rel_add_char (&term_tmp, term_component, &j); + *term_tmp++ = '['; - *term_tmp++ = '^'; - string_rel_add_char (&term_tmp, term_component, &i); - *term_tmp++ = '-'; + *term_tmp++ = '^'; + string_rel_add_char (&term_tmp, term_component, &i); + *term_tmp++ = '-'; - *term_tmp++ = ']'; - *term_tmp++ = '.'; - *term_tmp++ = '*'; + *term_tmp++ = ']'; + *term_tmp++ = '.'; + *term_tmp++ = '*'; - *term_tmp++ = '|'; - } - for (i = 0; term_component[i]; ) - string_rel_add_char (&term_tmp, term_component, &i); - *term_tmp++ = ')'; - *term_tmp = '\0'; + *term_tmp++ = '|'; + + if ((term_tmp - term_dict) > IT_MAX_WORD) + break; + } + for (i = 0; term_component[i]; ) + string_rel_add_char (&term_tmp, term_component, &i); + *term_tmp++ = ')'; + *term_tmp = '\0'; break; case 5: if (!term_100 (zh->reg->zebra_maps, reg_type, - term_sub, term_component, space_split, term_dst)) + term_sub, term_component, space_split, term_dst)) return 0; - logf (LOG_DEBUG, "Relation >"); + yaz_log(log_level_rpn, "Relation >"); - *term_tmp++ = '('; - for (i = 0; term_component[i];) - { - int j = 0; + *term_tmp++ = '('; + for (i = 0; term_component[i];) + { + int j = 0; - while (j < i) - string_rel_add_char (&term_tmp, term_component, &j); - *term_tmp++ = '['; - - *term_tmp++ = '^'; - *term_tmp++ = '-'; - string_rel_add_char (&term_tmp, term_component, &i); + while (j < i) + string_rel_add_char (&term_tmp, term_component, &j); + *term_tmp++ = '['; + + *term_tmp++ = '^'; + *term_tmp++ = '-'; + string_rel_add_char (&term_tmp, term_component, &i); - *term_tmp++ = ']'; - *term_tmp++ = '.'; - *term_tmp++ = '*'; + *term_tmp++ = ']'; + *term_tmp++ = '.'; + *term_tmp++ = '*'; - *term_tmp++ = '|'; - } - for (i = 0; term_component[i];) - string_rel_add_char (&term_tmp, term_component, &i); - *term_tmp++ = '.'; - *term_tmp++ = '+'; - *term_tmp++ = ')'; - *term_tmp = '\0'; + *term_tmp++ = '|'; + + if ((term_tmp - term_dict) > IT_MAX_WORD) + break; + } + for (i = 0; term_component[i];) + string_rel_add_char (&term_tmp, term_component, &i); + *term_tmp++ = '.'; + *term_tmp++ = '+'; + *term_tmp++ = ')'; + *term_tmp = '\0'; break; case 4: if (!term_100 (zh->reg->zebra_maps, reg_type, term_sub, - term_component, space_split, term_dst)) + term_component, space_split, term_dst)) return 0; - logf (LOG_DEBUG, "Relation >="); + yaz_log(log_level_rpn, "Relation >="); - *term_tmp++ = '('; - for (i = 0; term_component[i];) - { - int j = 0; + *term_tmp++ = '('; + for (i = 0; term_component[i];) + { + int j = 0; - if (i) - *term_tmp++ = '|'; - while (j < i) - string_rel_add_char (&term_tmp, term_component, &j); - *term_tmp++ = '['; + if (i) + *term_tmp++ = '|'; + while (j < i) + string_rel_add_char (&term_tmp, term_component, &j); + *term_tmp++ = '['; - if (term_component[i+1]) - { - *term_tmp++ = '^'; - *term_tmp++ = '-'; - string_rel_add_char (&term_tmp, term_component, &i); - } - else - { - string_rel_add_char (&term_tmp, term_component, &i); - *term_tmp++ = '-'; - } - *term_tmp++ = ']'; - *term_tmp++ = '.'; - *term_tmp++ = '*'; - } - *term_tmp++ = ')'; - *term_tmp = '\0'; + if (term_component[i+1]) + { + *term_tmp++ = '^'; + *term_tmp++ = '-'; + string_rel_add_char (&term_tmp, term_component, &i); + } + else + { + string_rel_add_char (&term_tmp, term_component, &i); + *term_tmp++ = '-'; + } + *term_tmp++ = ']'; + *term_tmp++ = '.'; + *term_tmp++ = '*'; + + if ((term_tmp - term_dict) > IT_MAX_WORD) + break; + } + *term_tmp++ = ')'; + *term_tmp = '\0'; break; case 3: - default: - logf (LOG_DEBUG, "Relation ="); + case 102: + case -1: + yaz_log(log_level_rpn, "Relation ="); if (!term_100 (zh->reg->zebra_maps, reg_type, term_sub, - term_component, space_split, term_dst)) + term_component, space_split, term_dst)) return 0; - strcat (term_tmp, "("); - strcat (term_tmp, term_component); - strcat (term_tmp, ")"); + strcat(term_tmp, "("); + strcat(term_tmp, term_component); + strcat(term_tmp, ")"); + break; + default: + *error_code = 117; + return 0; } return 1; } -static int string_term (ZebraHandle zh, Z_AttributesPlusTerm *zapt, - const char **term_sub, - oid_value attributeSet, NMEM stream, - struct grep_info *grep_info, - int reg_type, int complete_flag, - int num_bases, char **basenames, - char *term_dst); - -static RSET term_trunc (ZebraHandle zh, Z_AttributesPlusTerm *zapt, - const char **term_sub, - oid_value attributeSet, NMEM stream, - struct grep_info *grep_info, - int reg_type, int complete_flag, - int num_bases, char **basenames, - char *term_dst, - const char *rank_type) +static ZEBRA_RES string_term(ZebraHandle zh, Z_AttributesPlusTerm *zapt, + const char **term_sub, + oid_value attributeSet, NMEM stream, + struct grep_info *grep_info, + int reg_type, int complete_flag, + int num_bases, char **basenames, + char *term_dst, int xpath_use); + +static RSET term_trunc(ZebraHandle zh, Z_AttributesPlusTerm *zapt, + const char **term_sub, + oid_value attributeSet, NMEM stream, + struct grep_info *grep_info, + int reg_type, int complete_flag, + int num_bases, char **basenames, + char *term_dst, + const char *rank_type, int xpath_use, + NMEM rset_nmem) { - int r; + ZEBRA_RES res; grep_info->isam_p_indx = 0; - r = string_term (zh, zapt, term_sub, attributeSet, stream, grep_info, - reg_type, complete_flag, num_bases, basenames, - term_dst); - if (r < 1) + res = string_term(zh, zapt, term_sub, attributeSet, stream, grep_info, + reg_type, complete_flag, num_bases, basenames, + term_dst, xpath_use); + if (res != ZEBRA_OK) return 0; - logf (LOG_DEBUG, "term: %s", term_dst); - return rset_trunc (zh, grep_info->isam_p_buf, - grep_info->isam_p_indx, term_dst, - strlen(term_dst), rank_type, 1 /* preserve pos */); + if (!*term_sub) /* no more terms ? */ + return 0; + yaz_log(log_level_rpn, "term: %s", term_dst); + return rset_trunc(zh, grep_info->isam_p_buf, + grep_info->isam_p_indx, term_dst, + strlen(term_dst), rank_type, 1 /* preserve pos */, + zapt->term->which, rset_nmem, + key_it_ctrl, key_it_ctrl->scope); +} +static char *nmem_strdup_i(NMEM nmem, int v) +{ + char val_str[64]; + sprintf (val_str, "%d", v); + return nmem_strdup(nmem, val_str); } - -static int string_term (ZebraHandle zh, Z_AttributesPlusTerm *zapt, - const char **term_sub, - oid_value attributeSet, NMEM stream, - struct grep_info *grep_info, - int reg_type, int complete_flag, - int num_bases, char **basenames, - char *term_dst) +static ZEBRA_RES string_term(ZebraHandle zh, Z_AttributesPlusTerm *zapt, + const char **term_sub, + oid_value attributeSet, NMEM stream, + struct grep_info *grep_info, + int reg_type, int complete_flag, + int num_bases, char **basenames, + char *term_dst, int xpath_use) { char term_dict[2*IT_MAX_WORD+4000]; int j, r, base_no; @@ -848,25 +1025,31 @@ static int string_term (ZebraHandle zh, Z_AttributesPlusTerm *zapt, struct rpn_char_map_info rcmi; int space_split = complete_flag ? 0 : 1; + int bases_ok = 0; /* no of databases with OK attribute */ + int errCode = 0; /* err code (if any is not OK) */ + char *errString = 0; /* addinfo */ + rpn_char_map_prepare (zh->reg, reg_type, &rcmi); attr_init (&use, zapt, 1); use_value = attr_find_ex (&use, &curAttributeSet, &use_string); - logf (LOG_DEBUG, "string_term, use value %d", use_value); + yaz_log(log_level_rpn, "string_term, use value %d", use_value); attr_init (&truncation, zapt, 5); truncation_value = attr_find (&truncation, NULL); - logf (LOG_DEBUG, "truncation value %d", truncation_value); + yaz_log(log_level_rpn, "truncation value %d", truncation_value); - if (use_value == -1) - use_value = 1016; - if (use_value == -2) + if (use_value == -1) /* no attribute - assumy "any" */ use_value = 1016; - for (base_no = 0; base_no < num_bases; base_no++) { + int ord = -1; + int attr_ok = 0; + int regex_range = 0; + int init_pos = 0; attent attp; data1_local_attribute id_xpath_attr; data1_local_attribute *local_attr; int max_pos, prefix_len = 0; + int relation_error; termp = *term_sub; @@ -874,27 +1057,63 @@ static int string_term (ZebraHandle zh, Z_AttributesPlusTerm *zapt, { zh->errCode = 109; /* Database unavailable */ zh->errString = basenames[base_no]; - return -1; + return ZEBRA_FAIL; } - if (curAttributeSet == VAL_IDXPATH) + if (xpath_use > 0 && use_value == -2) { + /* xpath mode and we have a string attribute */ attp.local_attributes = &id_xpath_attr; - attp.attset_ordinal = curAttributeSet; + attp.attset_ordinal = VAL_IDXPATH; id_xpath_attr.next = 0; + + use_value = xpath_use; /* xpath_use as use-attribute now */ id_xpath_attr.local = use_value; } - else + else if (curAttributeSet == VAL_IDXPATH && use_value >= 0) + { + /* X-Path attribute, use numeric value directly */ + attp.local_attributes = &id_xpath_attr; + attp.attset_ordinal = VAL_IDXPATH; + id_xpath_attr.next = 0; + id_xpath_attr.local = use_value; + } + else if (use_string && + (ord = zebraExplain_lookup_attr_str(zh->reg->zei, + use_string)) >= 0) + { + /* we have a match for a raw string attribute */ + char ord_buf[32]; + int i, ord_len; + + if (prefix_len) + term_dict[prefix_len++] = '|'; + else + term_dict[prefix_len++] = '('; + + ord_len = key_SU_encode (ord, ord_buf); + for (i = 0; ierrCode = 114; - zh->errString = nmem_strdup (stream, val_str); + /* set was found, but value wasn't defined */ + errCode = 114; + if (use_string) + errString = nmem_strdup(stream, use_string); + else + errString = nmem_strdup_i (stream, use_value); } else { @@ -906,439 +1125,279 @@ static int string_term (ZebraHandle zh, Z_AttributesPlusTerm *zapt, oident.value = curAttributeSet; oid_ent_to_oid (&oident, oid); - zh->errCode = 121; - zh->errString = nmem_strdup (stream, oident.desc); + errCode = 121; + errString = nmem_strdup (stream, oident.desc); } - return -1; + continue; } } - for (local_attr = attp.local_attributes; local_attr; - local_attr = local_attr->next) - { - int ord; - char ord_buf[32]; - int i, ord_len; - - ord = zebraExplain_lookupSU (zh->reg->zei, attp.attset_ordinal, - local_attr->local); - if (ord < 0) - continue; - if (prefix_len) - term_dict[prefix_len++] = '|'; - else - term_dict[prefix_len++] = '('; - - ord_len = key_SU_encode (ord, ord_buf); - for (i = 0; ierrCode = 114; - zh->errString = nmem_strdup (stream, val_str); - return -1; - } + for (local_attr = attp.local_attributes; local_attr; + local_attr = local_attr->next) + { + char ord_buf[32]; + int i, ord_len; + + ord = zebraExplain_lookup_attr_su(zh->reg->zei, + attp.attset_ordinal, + local_attr->local); + if (ord < 0) + continue; + if (prefix_len) + term_dict[prefix_len++] = '|'; + else + term_dict[prefix_len++] = '('; + + ord_len = key_SU_encode (ord, ord_buf); + for (i = 0; ireg->dict, term_dict, 0, - grep_info, &max_pos, 0, grep_handle); - if (r) - logf (LOG_WARN, "dict_lookup_grep fail, rel=gt: %d", r); - break; - case 1: /* right truncation */ - term_dict[j++] = '('; - if (!term_100 (zh->reg->zebra_maps, reg_type, - &termp, term_dict + j, space_split, term_dst)) - return 0; - strcat (term_dict, ".*)"); - dict_lookup_grep (zh->reg->dict, term_dict, 0, grep_info, - &max_pos, 0, grep_handle); - break; - case 2: /* keft truncation */ - term_dict[j++] = '('; term_dict[j++] = '.'; term_dict[j++] = '*'; - if (!term_100 (zh->reg->zebra_maps, reg_type, - &termp, term_dict + j, space_split, term_dst)) - return 0; - strcat (term_dict, ")"); - dict_lookup_grep (zh->reg->dict, term_dict, 0, grep_info, - &max_pos, 0, grep_handle); - break; - case 3: /* left&right truncation */ - term_dict[j++] = '('; term_dict[j++] = '.'; term_dict[j++] = '*'; - if (!term_100 (zh->reg->zebra_maps, reg_type, - &termp, term_dict + j, space_split, term_dst)) - return 0; - strcat (term_dict, ".*)"); - dict_lookup_grep (zh->reg->dict, term_dict, 0, grep_info, - &max_pos, 0, grep_handle); + j = prefix_len; + switch (truncation_value) + { + case -1: /* not specified */ + case 100: /* do not truncate */ + if (!string_relation (zh, zapt, &termp, term_dict, + attributeSet, + reg_type, space_split, term_dst, + &relation_error)) + { + if (relation_error) + { + zh->errCode = relation_error; + return ZEBRA_FAIL; + } + *term_sub = 0; + return ZEBRA_OK; + } break; + case 1: /* right truncation */ + term_dict[j++] = '('; + if (!term_100(zh->reg->zebra_maps, reg_type, + &termp, term_dict + j, space_split, term_dst)) + { + *term_sub = 0; + return ZEBRA_OK; + } + strcat(term_dict, ".*)"); + break; + case 2: /* keft truncation */ + term_dict[j++] = '('; term_dict[j++] = '.'; term_dict[j++] = '*'; + if (!term_100(zh->reg->zebra_maps, reg_type, + &termp, term_dict + j, space_split, term_dst)) + { + *term_sub = 0; + return ZEBRA_OK; + } + strcat(term_dict, ")"); + break; + case 3: /* left&right truncation */ + term_dict[j++] = '('; term_dict[j++] = '.'; term_dict[j++] = '*'; + if (!term_100(zh->reg->zebra_maps, reg_type, + &termp, term_dict + j, space_split, term_dst)) + { + *term_sub = 0; + return ZEBRA_OK; + } + strcat(term_dict, ".*)"); + break; + case 101: /* process # in term */ + term_dict[j++] = '('; + if (!term_101(zh->reg->zebra_maps, reg_type, + &termp, term_dict + j, space_split, term_dst)) + { + *term_sub = 0; + return ZEBRA_OK; + } + strcat(term_dict, ")"); + break; + case 102: /* Regexp-1 */ + term_dict[j++] = '('; + if (!term_102(zh->reg->zebra_maps, reg_type, + &termp, term_dict + j, space_split, term_dst)) + { + *term_sub = 0; + return ZEBRA_OK; + } + strcat(term_dict, ")"); + break; + case 103: /* Regexp-2 */ + r = 1; + term_dict[j++] = '('; + init_pos = 2; + if (!term_103 (zh->reg->zebra_maps, reg_type, + &termp, term_dict + j, ®ex_range, + space_split, term_dst)) + { + *term_sub = 0; + return ZEBRA_OK; + } + strcat(term_dict, ")"); + case 104: /* process # and ! in term */ + term_dict[j++] = '('; + if (!term_104 (zh->reg->zebra_maps, reg_type, + &termp, term_dict + j, space_split, term_dst)) + { + *term_sub = 0; + return ZEBRA_OK; + } + strcat(term_dict, ")"); + break; + case 105: /* process * and ! in term */ + term_dict[j++] = '('; + if (!term_105 (zh->reg->zebra_maps, reg_type, + &termp, term_dict + j, space_split, term_dst, 1)) + { + *term_sub = 0; + return ZEBRA_OK; + } + strcat(term_dict, ")"); + break; + case 106: /* process * and ! in term */ + term_dict[j++] = '('; + if (!term_105 (zh->reg->zebra_maps, reg_type, + &termp, term_dict + j, space_split, term_dst, 0)) + { + *term_sub = 0; + return ZEBRA_OK; + } + strcat(term_dict, ")"); + break; + default: zh->errCode = 120; - return -1; - case 101: /* process # in term */ - term_dict[j++] = '('; - if (!term_101 (zh->reg->zebra_maps, reg_type, - &termp, term_dict + j, space_split, term_dst)) - return 0; - strcat (term_dict, ")"); - r = dict_lookup_grep (zh->reg->dict, term_dict, 0, grep_info, - &max_pos, 0, grep_handle); - if (r) - logf (LOG_WARN, "dict_lookup_grep err, trunc=#: %d", r); - break; - case 102: /* Regexp-1 */ - term_dict[j++] = '('; - if (!term_102 (zh->reg->zebra_maps, reg_type, - &termp, term_dict + j, space_split, term_dst)) - return 0; - strcat (term_dict, ")"); - logf (LOG_DEBUG, "Regexp-1 tolerance=%d", r); - r = dict_lookup_grep (zh->reg->dict, term_dict, 0, grep_info, - &max_pos, 0, grep_handle); - if (r) - logf (LOG_WARN, "dict_lookup_grep err, trunc=regular: %d", - r); - break; - case 103: /* Regexp-2 */ - r = 1; - term_dict[j++] = '('; - if (!term_103 (zh->reg->zebra_maps, reg_type, - &termp, term_dict + j, &r, space_split, term_dst)) - return 0; - strcat (term_dict, ")"); - logf (LOG_DEBUG, "Regexp-2 tolerance=%d", r); - r = dict_lookup_grep (zh->reg->dict, term_dict, r, grep_info, - &max_pos, 2, grep_handle); - if (r) - logf (LOG_WARN, "dict_lookup_grep err, trunc=eregular: %d", - r); - break; - case 104: /* process # and ! in term */ - term_dict[j++] = '('; - if (!term_104 (zh->reg->zebra_maps, reg_type, - &termp, term_dict + j, space_split, term_dst)) - return 0; - strcat (term_dict, ")"); - r = dict_lookup_grep (zh->reg->dict, term_dict, 0, grep_info, - &max_pos, 0, grep_handle); - if (r) - logf (LOG_WARN, "dict_lookup_grep err, trunc=#/!: %d", r); - break; - case 105: /* process * and ! in term */ - term_dict[j++] = '('; - if (!term_105 (zh->reg->zebra_maps, reg_type, - &termp, term_dict + j, space_split, term_dst, 1)) - return 0; - strcat (term_dict, ")"); - r = dict_lookup_grep (zh->reg->dict, term_dict, 0, grep_info, - &max_pos, 0, grep_handle); - if (r) - logf (LOG_WARN, "dict_lookup_grep err, trunc=*/!: %d", r); - break; - case 106: /* process * and ! in term */ - term_dict[j++] = '('; - if (!term_105 (zh->reg->zebra_maps, reg_type, - &termp, term_dict + j, space_split, term_dst, 0)) - return 0; - strcat (term_dict, ")"); - r = dict_lookup_grep (zh->reg->dict, term_dict, 0, grep_info, - &max_pos, 0, grep_handle); - if (r) - logf (LOG_WARN, "dict_lookup_grep err, trunc=*/!: %d", r); - break; + zh->errString = nmem_strdup_i(stream, truncation_value); + return ZEBRA_FAIL; } + if (attr_ok) + { + char buf[80]; + const char *input = term_dict + prefix_len; + esc_str(buf, sizeof(buf), input, strlen(input)); + } + if (attr_ok) + { + yaz_log(log_level_rpn, "dict_lookup_grep: %s", term_dict+prefix_len); + r = dict_lookup_grep(zh->reg->dict, term_dict, regex_range, + grep_info, &max_pos, init_pos, + grep_handle); + if (r) + yaz_log(YLOG_WARN, "dict_lookup_grep fail %d", r); + } + } + if (!bases_ok) + { + zh->errCode = errCode; + zh->errString = errString; + return ZEBRA_FAIL; } *term_sub = termp; - logf (LOG_DEBUG, "%d positions", grep_info->isam_p_indx); - return 1; + yaz_log(YLOG_DEBUG, "%d positions", grep_info->isam_p_indx); + return ZEBRA_OK; } -static void trans_term (ZebraHandle zh, Z_AttributesPlusTerm *zapt, - char *termz) + +/* convert APT search term to UTF8 */ +static int zapt_term_to_utf8 (ZebraHandle zh, Z_AttributesPlusTerm *zapt, + char *termz) { size_t sizez; Z_Term *term = zapt->term; - sizez = term->u.general->len; - if (sizez > IT_MAX_WORD-1) - sizez = IT_MAX_WORD-1; - memcpy (termz, term->u.general->buf, sizez); - termz[sizez] = '\0'; -} - -static void trans_scan_term (ZebraHandle zh, Z_AttributesPlusTerm *zapt, - char *termz, int reg_type) -{ - Z_Term *term = zapt->term; - const char **map; - const char *cp = (const char *) term->u.general->buf; - const char *cp_end = cp + term->u.general->len; - const char *src; - int i = 0; - const char *space_map = NULL; - int len; - - while ((len = (cp_end - cp)) > 0) + switch (term->which) { - map = zebra_maps_input (zh->reg->zebra_maps, reg_type, &cp, len); - if (**map == *CHR_SPACE) - space_map = *map; + case Z_Term_general: + if (zh->iconv_to_utf8 != 0) + { + char *inbuf = term->u.general->buf; + size_t inleft = term->u.general->len; + char *outbuf = termz; + size_t outleft = IT_MAX_WORD-1; + size_t ret; + + ret = yaz_iconv(zh->iconv_to_utf8, &inbuf, &inleft, + &outbuf, &outleft); + if (ret == (size_t)(-1)) + { + ret = yaz_iconv(zh->iconv_to_utf8, 0, 0, 0, 0); + zh->errCode = 125; + return -1; + } + *outbuf = 0; + } else { - if (i && space_map) - for (src = space_map; *src; src++) - termz[i++] = *src; - space_map = NULL; - for (src = *map; *src; src++) - termz[i++] = *src; + sizez = term->u.general->len; + if (sizez > IT_MAX_WORD-1) + sizez = IT_MAX_WORD-1; + memcpy (termz, term->u.general->buf, sizez); + termz[sizez] = '\0'; } + break; + case Z_Term_characterString: + sizez = strlen(term->u.characterString); + if (sizez > IT_MAX_WORD-1) + sizez = IT_MAX_WORD-1; + memcpy (termz, term->u.characterString, sizez); + termz[sizez] = '\0'; + break; + default: + zh->errCode = 124; + return -1; } - termz[i] = '\0'; + return 0; } -static RSET rpn_prox (ZebraHandle zh, RSET *rset, int rset_no, - int ordered, int exclusion, int relation, int distance) +/* convert APT SCAN term to internal cmap */ +static int trans_scan_term(ZebraHandle zh, Z_AttributesPlusTerm *zapt, + char *termz, int reg_type) { - int i; - RSFD *rsfd; - int *more; - struct it_key **buf; - RSET result; - char prox_term[1024]; - int length_prox_term = 0; - int min_nn = 10000000; - int term_index; - const char *flags = NULL; - - rsfd = (RSFD *) xmalloc (sizeof(*rsfd)*rset_no); - more = (int *) xmalloc (sizeof(*more)*rset_no); - buf = (struct it_key **) xmalloc (sizeof(*buf)*rset_no); + char termz0[IT_MAX_WORD]; - *prox_term = '\0'; - for (i = 0; ino_rset_terms; j++) - { - const char *nflags = rset[i]->rset_terms[j]->flags; - char *term = rset[i]->rset_terms[j]->name; - int lterm = strlen(term); - if (lterm + length_prox_term < sizeof(prox_term)-1) - { - if (length_prox_term) - prox_term[length_prox_term++] = ' '; - strcpy (prox_term + length_prox_term, term); - length_prox_term += lterm; - } - if (min_nn > rset[i]->rset_terms[j]->nn) - min_nn = rset[i]->rset_terms[j]->nn; - flags = nflags; - } - } - for (i = 0; inn = 0; - result = rset_create (rset_kind_null, &parms); - } - else if (ordered && relation == 3 && exclusion == 0 && distance == 1) - { - /* special proximity case = phrase search ... */ - rset_temp_parms parms; - RSFD rsfd_result; - - parms.rset_term = rset_term_create (prox_term, length_prox_term, - flags); - parms.rset_term->nn = min_nn; - parms.cmp = key_compare_it; - parms.key_size = sizeof (struct it_key); - parms.temp_path = res_get (zh->res, "setTmpDir"); - result = rset_create (rset_kind_temp, &parms); - rsfd_result = rset_open (result, RSETF_WRITE); - - while (*more) - { - for (i = 1; i 1) - { - more[i-1] = rset_read (rset[i-1], rsfd[i-1], - buf[i-1], &term_index); - break; - } - else if (cmp == 1) - { - if (buf[i-1]->seqno+1 != buf[i]->seqno) - { - more[i-1] = rset_read (rset[i-1], rsfd[i-1], - buf[i-1], &term_index); - break; - } - } - else - { - more[i] = rset_read (rset[i], rsfd[i], buf[i], - &term_index); - break; - } - } - if (i == rset_no) - { - rset_write (result, rsfd_result, buf[0]); - more[0] = rset_read (*rset, *rsfd, *buf, &term_index); - } - } - rset_close (result, rsfd_result); - } - else if (rset_no == 2) - { - /* generic proximity case (two input sets only) ... */ - rset_temp_parms parms; - RSFD rsfd_result; - - logf (LOG_LOG, "generic prox, dist = %d, relation = %d, ordered =%d, exclusion=%d", - distance, relation, ordered, exclusion); - parms.rset_term = rset_term_create (prox_term, length_prox_term, - flags); - parms.rset_term->nn = min_nn; - parms.cmp = key_compare_it; - parms.key_size = sizeof (struct it_key); - parms.temp_path = res_get (zh->res, "setTmpDir"); - result = rset_create (rset_kind_temp, &parms); - rsfd_result = rset_open (result, RSETF_WRITE); - - while (more[0] && more[1]) - { - int cmp = key_compare_it (buf[0], buf[1]); - if (cmp < -1) - more[0] = rset_read (rset[0], rsfd[0], buf[0], &term_index); - else if (cmp > 1) - more[1] = rset_read (rset[1], rsfd[1], buf[1], &term_index); - else - { - int sysno = buf[0]->sysno; - int seqno[500]; - int n = 0; - - seqno[n++] = buf[0]->seqno; - while ((more[0] = rset_read (rset[0], rsfd[0], buf[0], - &term_index)) && - sysno == buf[0]->sysno) - if (n < 500) - seqno[n++] = buf[0]->seqno; - do - { - for (i = 0; iseqno - seqno[i]; - int excl = exclusion; - if (!ordered && diff < 0) - diff = -diff; - switch (relation) - { - case 1: /* < */ - if (diff < distance && diff >= 0) - excl = !excl; - break; - case 2: /* <= */ - if (diff <= distance && diff >= 0) - excl = !excl; - break; - case 3: /* == */ - if (diff == distance && diff >= 0) - excl = !excl; - break; - case 4: /* >= */ - if (diff >= distance && diff >= 0) - excl = !excl; - break; - case 5: /* > */ - if (diff > distance && diff >= 0) - excl = !excl; - break; - case 6: /* != */ - if (diff != distance && diff >= 0) - excl = !excl; - break; - } - if (excl) - { - rset_write (result, rsfd_result, buf[1]); - break; - } - } - } while ((more[1] = rset_read (rset[1], rsfd[1], buf[1], - &term_index)) && - sysno == buf[1]->sysno); - } - } - rset_close (result, rsfd_result); - } + if (zapt_term_to_utf8(zh, zapt, termz0)) + return -1; /* error */ else { - rset_null_parms parms; - - parms.rset_term = rset_term_create (prox_term, length_prox_term, - flags); - parms.rset_term->nn = 0; - result = rset_create (rset_kind_null, &parms); - } - for (i = 0; i 0) + { + map = zebra_maps_input (zh->reg->zebra_maps, reg_type, &cp, len, 0); + if (**map == *CHR_SPACE) + space_map = *map; + else + { + if (i && space_map) + for (src = space_map; *src; src++) + termz[i++] = *src; + space_map = NULL; + for (src = *map; *src; src++) + termz[i++] = *src; + } + } + termz[i] = '\0'; } - xfree (buf); - xfree (more); - xfree (rsfd); - return result; + return 0; } - char *normalize_term(ZebraHandle zh, Z_AttributesPlusTerm *zapt, - const char *termz, NMEM stream, unsigned reg_id) + const char *termz, NMEM stream, unsigned reg_id) { WRBUF wrbuf = 0; AttrType truncation; @@ -1351,33 +1410,33 @@ char *normalize_term(ZebraHandle zh, Z_AttributesPlusTerm *zapt, switch (truncation_value) { default: - ex_list = ""; - break; + ex_list = ""; + break; case 101: - ex_list = "#"; - break; + ex_list = "#"; + break; case 102: case 103: - ex_list = 0; - break; + ex_list = 0; + break; case 104: - ex_list = "!#"; - break; + ex_list = "!#"; + break; case 105: - ex_list = "!*"; - break; + ex_list = "!*"; + break; } if (ex_list) - wrbuf = zebra_replace(zh->reg->zebra_maps, reg_id, ex_list, - termz, strlen(termz)); + wrbuf = zebra_replace(zh->reg->zebra_maps, reg_id, ex_list, + termz, strlen(termz)); if (!wrbuf) - return nmem_strdup(stream, termz); + return nmem_strdup(stream, termz); else { - char *buf = (char*) nmem_malloc (stream, wrbuf_len(wrbuf)+1); - memcpy (buf, wrbuf_buf(wrbuf), wrbuf_len(wrbuf)); - buf[wrbuf_len(wrbuf)] = '\0'; - return buf; + char *buf = (char*) nmem_malloc(stream, wrbuf_len(wrbuf)+1); + memcpy (buf, wrbuf_buf(wrbuf), wrbuf_len(wrbuf)); + buf[wrbuf_len(wrbuf)] = '\0'; + return buf; } } @@ -1390,10 +1449,10 @@ static void grep_info_delete (struct grep_info *grep_info) } static int grep_info_prepare (ZebraHandle zh, - Z_AttributesPlusTerm *zapt, - struct grep_info *grep_info, - int reg_type, - NMEM stream) + Z_AttributesPlusTerm *zapt, + struct grep_info *grep_info, + int reg_type, + NMEM stream) { AttrType termset; int termset_value_numeric; @@ -1408,197 +1467,190 @@ static int grep_info_prepare (ZebraHandle zh, grep_info->reg_type = reg_type; grep_info->termset = 0; + if (!zapt) + return 0; attr_init (&termset, zapt, 8); termset_value_numeric = - attr_find_ex (&termset, NULL, &termset_value_string); + attr_find_ex (&termset, NULL, &termset_value_string); if (termset_value_numeric != -1) { - char resname[32]; - const char *termset_name = 0; - if (termset_value_numeric != -2) - { + char resname[32]; + const char *termset_name = 0; + if (termset_value_numeric != -2) + { - sprintf (resname, "%d", termset_value_numeric); - termset_name = resname; - } - else - termset_name = termset_value_string; - logf (LOG_LOG, "creating termset set %s", termset_name); - grep_info->termset = resultSetAdd (zh, termset_name, 1); - if (!grep_info->termset) - { - zh->errCode = 128; - zh->errString = nmem_strdup (stream, termset_name); - return -1; - } + sprintf (resname, "%d", termset_value_numeric); + termset_name = resname; + } + else + termset_name = termset_value_string; + yaz_log(log_level_rpn, "creating termset set %s", termset_name); + grep_info->termset = resultSetAdd (zh, termset_name, 1); + if (!grep_info->termset) + { + zh->errCode = 128; + zh->errString = nmem_strdup (stream, termset_name); + return -1; + } } return 0; } - + -static RSET rpn_search_APT_phrase (ZebraHandle zh, - Z_AttributesPlusTerm *zapt, - const char *termz_org, - oid_value attributeSet, - NMEM stream, - int reg_type, int complete_flag, - const char *rank_type, - int num_bases, char **basenames) +static RSET rpn_search_APT_phrase(ZebraHandle zh, + Z_AttributesPlusTerm *zapt, + const char *termz_org, + oid_value attributeSet, + NMEM stream, + int reg_type, int complete_flag, + const char *rank_type, int xpath_use, + int num_bases, char **basenames, + NMEM rset_nmem) { char term_dst[IT_MAX_WORD+1]; - RSET rset[60], result; - int i, rset_no = 0; + RSET rset[TERM_LIST_LENGTH_MAX], result; + size_t rset_no = 0; struct grep_info grep_info; char *termz = normalize_term(zh, zapt, termz_org, stream, reg_type); const char *termp = termz; *term_dst = 0; if (grep_info_prepare (zh, zapt, &grep_info, reg_type, stream)) - return 0; - while (1) + return 0; + for (; rset_no < sizeof(rset)/sizeof(*rset); rset_no++) { - logf (LOG_DEBUG, "APT_phrase termp=%s", termp); - rset[rset_no] = term_trunc (zh, zapt, &termp, attributeSet, - stream, &grep_info, - reg_type, complete_flag, - num_bases, basenames, - term_dst, rank_type); - if (!rset[rset_no]) - break; - if (++rset_no >= (int) (sizeof(rset)/sizeof(*rset))) - break; + yaz_log(log_level_rpn, "APT_phrase termp=%s", termp); + rset[rset_no] = term_trunc(zh, zapt, &termp, attributeSet, + stream, &grep_info, + reg_type, complete_flag, + num_bases, basenames, + term_dst, rank_type, + xpath_use, rset_nmem); + if (!rset[rset_no]) /* error or no more terms */ + break; } grep_info_delete (&grep_info); - if (rset_no == 0) + if (termp) { - rset_null_parms parms; - - parms.rset_term = rset_term_create (termz, -1, rank_type); - return rset_create (rset_kind_null, &parms); + int i; + for (i = 0; iscope, + rset_no, rset, + 1 /* ordered */, 0 /* exclusion */, + 3 /* relation */, 1 /* distance */); return result; } -static RSET rpn_search_APT_or_list (ZebraHandle zh, - Z_AttributesPlusTerm *zapt, - const char *termz_org, - oid_value attributeSet, - NMEM stream, - int reg_type, int complete_flag, - const char *rank_type, - int num_bases, char **basenames) +static RSET rpn_search_APT_or_list(ZebraHandle zh, + Z_AttributesPlusTerm *zapt, + const char *termz_org, + oid_value attributeSet, + NMEM stream, + int reg_type, int complete_flag, + const char *rank_type, + int xpath_use, + int num_bases, char **basenames, + NMEM rset_nmem) { char term_dst[IT_MAX_WORD+1]; - RSET rset[60], result; - int i, rset_no = 0; + RSET rset[TERM_LIST_LENGTH_MAX]; + size_t rset_no = 0; struct grep_info grep_info; char *termz = normalize_term(zh, zapt, termz_org, stream, reg_type); const char *termp = termz; - if (grep_info_prepare (zh, zapt, &grep_info, reg_type, stream)) - return 0; - while (1) + if (grep_info_prepare(zh, zapt, &grep_info, reg_type, stream)) + return 0; + for (; rset_no < sizeof(rset)/sizeof(*rset); rset_no++) { - logf (LOG_DEBUG, "APT_or_list termp=%s", termp); - rset[rset_no] = term_trunc (zh, zapt, &termp, attributeSet, - stream, &grep_info, - reg_type, complete_flag, - num_bases, basenames, - term_dst, rank_type); - if (!rset[rset_no]) - break; - if (++rset_no >= (int) (sizeof(rset)/sizeof(*rset))) - break; + yaz_log(log_level_rpn, "APT_or_list termp=%s", termp); + rset[rset_no] = term_trunc(zh, zapt, &termp, attributeSet, + stream, &grep_info, + reg_type, complete_flag, + num_bases, basenames, + term_dst, rank_type, + xpath_use, rset_nmem); + if (!rset[rset_no]) /* error or no more terms */ + break; } grep_info_delete (&grep_info); - if (rset_no == 0) + if (termp) { - rset_null_parms parms; - - parms.rset_term = rset_term_create (termz, -1, rank_type); - return rset_create (rset_kind_null, &parms); - } - result = rset[0]; - for (i = 1; iscope, + rset_no, rset); } -static RSET rpn_search_APT_and_list (ZebraHandle zh, - Z_AttributesPlusTerm *zapt, - const char *termz_org, - oid_value attributeSet, - NMEM stream, - int reg_type, int complete_flag, - const char *rank_type, - int num_bases, char **basenames) +static RSET rpn_search_APT_and_list(ZebraHandle zh, + Z_AttributesPlusTerm *zapt, + const char *termz_org, + oid_value attributeSet, + NMEM stream, + int reg_type, int complete_flag, + const char *rank_type, + int xpath_use, + int num_bases, char **basenames, + NMEM rset_nmem) { char term_dst[IT_MAX_WORD+1]; - RSET rset[60], result; - int i, rset_no = 0; + RSET rset[TERM_LIST_LENGTH_MAX]; + size_t rset_no = 0; struct grep_info grep_info; char *termz = normalize_term(zh, zapt, termz_org, stream, reg_type); const char *termp = termz; - if (grep_info_prepare (zh, zapt, &grep_info, reg_type, stream)) - return 0; - while (1) + if (grep_info_prepare(zh, zapt, &grep_info, reg_type, stream)) + return 0; + for (; rset_no < sizeof(rset)/sizeof(*rset); rset_no++) { - logf (LOG_DEBUG, "APT_and_list termp=%s", termp); - rset[rset_no] = term_trunc (zh, zapt, &termp, attributeSet, - stream, &grep_info, - reg_type, complete_flag, - num_bases, basenames, - term_dst, rank_type); - if (!rset[rset_no]) - break; - assert (rset[rset_no]); - if (++rset_no >= (int) (sizeof(rset)/sizeof(*rset))) - break; + yaz_log(log_level_rpn, "APT_and_list termp=%s", termp); + rset[rset_no] = term_trunc(zh, zapt, &termp, attributeSet, + stream, &grep_info, + reg_type, complete_flag, + num_bases, basenames, + term_dst, rank_type, + xpath_use, rset_nmem); + if (!rset[rset_no]) /* error or no more terms */ + break; } grep_info_delete (&grep_info); - if (rset_no == 0) + if (termp) { - rset_null_parms parms; - - parms.rset_term = rset_term_create (termz, -1, rank_type); - return rset_create (rset_kind_null, &parms); - } - result = rset[0]; - for (i = 1; iscope, + rset_no, rset); } -static int numeric_relation (ZebraHandle zh, Z_AttributesPlusTerm *zapt, - const char **term_sub, - char *term_dict, - oid_value attributeSet, - struct grep_info *grep_info, - int *max_pos, - int reg_type, - char *term_dst) +static int numeric_relation(ZebraHandle zh, Z_AttributesPlusTerm *zapt, + const char **term_sub, + char *term_dict, + oid_value attributeSet, + struct grep_info *grep_info, + int *max_pos, + int reg_type, + char *term_dst, + int *error_code) { AttrType relation; int relation_value; @@ -1606,66 +1658,76 @@ static int numeric_relation (ZebraHandle zh, Z_AttributesPlusTerm *zapt, int r; char *term_tmp = term_dict + strlen(term_dict); + *error_code = 0; attr_init (&relation, zapt, 2); relation_value = attr_find (&relation, NULL); - logf (LOG_DEBUG, "numeric relation value=%d", relation_value); + yaz_log(log_level_rpn, "numeric relation value=%d", relation_value); - if (!term_100 (zh->reg->zebra_maps, reg_type, term_sub, term_tmp, 1, - term_dst)) - return 0; + if (!term_100(zh->reg->zebra_maps, reg_type, term_sub, term_tmp, 1, + term_dst)) + return 0; term_value = atoi (term_tmp); switch (relation_value) { case 1: - logf (LOG_DEBUG, "Relation <"); - gen_regular_rel (term_tmp, term_value-1, 1); + yaz_log(log_level_rpn, "Relation <"); + gen_regular_rel(term_tmp, term_value-1, 1); break; case 2: - logf (LOG_DEBUG, "Relation <="); - gen_regular_rel (term_tmp, term_value, 1); + yaz_log(log_level_rpn, "Relation <="); + gen_regular_rel(term_tmp, term_value, 1); break; case 4: - logf (LOG_DEBUG, "Relation >="); - gen_regular_rel (term_tmp, term_value, 0); + yaz_log(log_level_rpn, "Relation >="); + gen_regular_rel(term_tmp, term_value, 0); break; case 5: - logf (LOG_DEBUG, "Relation >"); - gen_regular_rel (term_tmp, term_value+1, 0); + yaz_log(log_level_rpn, "Relation >"); + gen_regular_rel(term_tmp, term_value+1, 0); break; + case -1: case 3: + yaz_log(log_level_rpn, "Relation ="); + sprintf(term_tmp, "(0*%d)", term_value); + break; default: - logf (LOG_DEBUG, "Relation ="); - sprintf (term_tmp, "(0*%d)", term_value); + *error_code = 117; + return 0; } - logf (LOG_DEBUG, "dict_lookup_grep: %s", term_tmp); - r = dict_lookup_grep (zh->reg->dict, term_dict, 0, grep_info, max_pos, + yaz_log(log_level_rpn, "dict_lookup_grep: %s", term_tmp); + r = dict_lookup_grep(zh->reg->dict, term_dict, 0, grep_info, max_pos, 0, grep_handle); if (r) - logf (LOG_WARN, "dict_lookup_grep fail, rel=gt: %d", r); - logf (LOG_DEBUG, "%d positions", grep_info->isam_p_indx); + yaz_log(YLOG_WARN, "dict_lookup_grep fail, rel = gt: %d", r); + yaz_log(log_level_rpn, "%d positions", grep_info->isam_p_indx); return 1; } -static int numeric_term (ZebraHandle zh, Z_AttributesPlusTerm *zapt, - const char **term_sub, - oid_value attributeSet, struct grep_info *grep_info, - int reg_type, int complete_flag, - int num_bases, char **basenames, - char *term_dst) +static ZEBRA_RES numeric_term(ZebraHandle zh, Z_AttributesPlusTerm *zapt, + const char **term_sub, + oid_value attributeSet, + struct grep_info *grep_info, + int reg_type, int complete_flag, + int num_bases, char **basenames, + char *term_dst, int xpath_use, NMEM stream) { char term_dict[2*IT_MAX_WORD+2]; int r, base_no; AttrType use; int use_value; + const char *use_string = 0; oid_value curAttributeSet = attributeSet; const char *termp; struct rpn_char_map_info rcmi; + int bases_ok = 0; /* no of databases with OK attribute */ + int errCode = 0; /* err code (if any is not OK) */ + char *errString = 0; /* addinfo */ + rpn_char_map_prepare (zh->reg, reg_type, &rcmi); attr_init (&use, zapt, 1); - use_value = attr_find (&use, &curAttributeSet); - logf (LOG_DEBUG, "numeric_term, use value %d", use_value); + use_value = attr_find_ex (&use, &curAttributeSet, &use_string); if (use_value == -1) use_value = 1016; @@ -1673,19 +1735,46 @@ static int numeric_term (ZebraHandle zh, Z_AttributesPlusTerm *zapt, for (base_no = 0; base_no < num_bases; base_no++) { attent attp; + data1_local_attribute id_xpath_attr; data1_local_attribute *local_attr; int max_pos, prefix_len = 0; + int relation_error = 0; termp = *term_sub; - if ((r=att_getentbyatt (zh, &attp, curAttributeSet, use_value))) + if (use_value == -2) /* string attribute (assume IDXPATH/any) */ { - logf (LOG_DEBUG, "att_getentbyatt fail. set=%d use=%d r=%d", - curAttributeSet, use_value, r); - if (r == -1) - zh->errCode = 114; - else - zh->errCode = 121; - return -1; + use_value = xpath_use; + attp.local_attributes = &id_xpath_attr; + attp.attset_ordinal = VAL_IDXPATH; + id_xpath_attr.next = 0; + id_xpath_attr.local = use_value; + } + else if (curAttributeSet == VAL_IDXPATH) + { + attp.local_attributes = &id_xpath_attr; + attp.attset_ordinal = VAL_IDXPATH; + id_xpath_attr.next = 0; + id_xpath_attr.local = use_value; + } + else + { + if ((r = att_getentbyatt (zh, &attp, curAttributeSet, use_value, + use_string))) + { + yaz_log(YLOG_DEBUG, "att_getentbyatt fail. set=%d use=%d r=%d", + curAttributeSet, use_value, r); + if (r == -1) + { + errCode = 114; + if (use_string) + errString = nmem_strdup(stream, use_string); + else + errString = nmem_strdup_i (stream, use_value); + } + else + errCode = 121; + continue; + } } if (zebraExplain_curDatabase (zh->reg->zei, basenames[base_no])) { @@ -1697,11 +1786,12 @@ static int numeric_term (ZebraHandle zh, Z_AttributesPlusTerm *zapt, local_attr = local_attr->next) { int ord; - char ord_buf[32]; - int i, ord_len; + char ord_buf[32]; + int i, ord_len; - ord = zebraExplain_lookupSU (zh->reg->zei, attp.attset_ordinal, - local_attr->local); + ord = zebraExplain_lookup_attr_su(zh->reg->zei, + attp.attset_ordinal, + local_attr->local); if (ord < 0) continue; if (prefix_len) @@ -1709,123 +1799,136 @@ static int numeric_term (ZebraHandle zh, Z_AttributesPlusTerm *zapt, else term_dict[prefix_len++] = '('; - ord_len = key_SU_encode (ord, ord_buf); - for (i = 0; ierrCode = 114; - return -1; + errCode = 114; + errString = nmem_strdup_i(stream, use_value); + continue; } + bases_ok++; term_dict[prefix_len++] = ')'; term_dict[prefix_len++] = 1; term_dict[prefix_len++] = reg_type; - logf (LOG_DEBUG, "reg_type = %d", term_dict[prefix_len-1]); + yaz_log(YLOG_DEBUG, "reg_type = %d", term_dict[prefix_len-1]); term_dict[prefix_len] = '\0'; - if (!numeric_relation (zh, zapt, &termp, term_dict, - attributeSet, grep_info, &max_pos, reg_type, - term_dst)) - return 0; + if (!numeric_relation(zh, zapt, &termp, term_dict, + attributeSet, grep_info, &max_pos, reg_type, + term_dst, &relation_error)) + { + if (relation_error) + { + zh->errCode = relation_error; + zh->errString = 0; + return ZEBRA_FAIL; + } + *term_sub = 0; + return ZEBRA_OK; + } + } + if (!bases_ok) + { + zh->errCode = errCode; + zh->errString = errString; + return ZEBRA_FAIL; } *term_sub = termp; - logf (LOG_DEBUG, "%d positions", grep_info->isam_p_indx); - return 1; + yaz_log(YLOG_DEBUG, "%d positions", grep_info->isam_p_indx); + return ZEBRA_OK; } -static RSET rpn_search_APT_numeric (ZebraHandle zh, - Z_AttributesPlusTerm *zapt, - const char *termz, - oid_value attributeSet, - NMEM stream, - int reg_type, int complete_flag, - const char *rank_type, - int num_bases, char **basenames) +static RSET rpn_search_APT_numeric(ZebraHandle zh, + Z_AttributesPlusTerm *zapt, + const char *termz, + oid_value attributeSet, + NMEM stream, + int reg_type, int complete_flag, + const char *rank_type, int xpath_use, + int num_bases, char **basenames, + NMEM rset_nmem) { char term_dst[IT_MAX_WORD+1]; const char *termp = termz; - RSET rset[60], result; - int i, r, rset_no = 0; + RSET rset[TERM_LIST_LENGTH_MAX]; + ZEBRA_RES r; + size_t rset_no = 0; struct grep_info grep_info; + yaz_log(log_level_rpn, "APT_numeric t='%s'",termz); if (grep_info_prepare (zh, zapt, &grep_info, reg_type, stream)) - return 0; - while (1) + return 0; + for (; rset_no < sizeof(rset)/sizeof(*rset); rset_no++) { - logf (LOG_DEBUG, "APT_numeric termp=%s", termp); - grep_info.isam_p_indx = 0; - r = numeric_term (zh, zapt, &termp, attributeSet, &grep_info, - reg_type, complete_flag, num_bases, basenames, - term_dst); - if (r < 1) - break; - logf (LOG_DEBUG, "term: %s", term_dst); - rset[rset_no] = rset_trunc (zh, grep_info.isam_p_buf, + yaz_log(YLOG_DEBUG, "APT_numeric termp=%s", termp); + grep_info.isam_p_indx = 0; + r = numeric_term(zh, zapt, &termp, attributeSet, &grep_info, + reg_type, complete_flag, num_bases, basenames, + term_dst, xpath_use, + stream); + if (r == ZEBRA_FAIL || termp == 0) + break; + yaz_log(YLOG_DEBUG, "term: %s", term_dst); + rset[rset_no] = rset_trunc(zh, grep_info.isam_p_buf, grep_info.isam_p_indx, term_dst, - strlen(term_dst), rank_type, - 0 /* preserve position */); - assert (rset[rset_no]); - if (++rset_no >= (int) (sizeof(rset)/sizeof(*rset))) - break; + strlen(term_dst), rank_type, + 0 /* preserve position */, + zapt->term->which, rset_nmem, + key_it_ctrl,key_it_ctrl->scope); + if (!rset[rset_no]) + break; } grep_info_delete (&grep_info); - if (rset_no == 0) - { - rset_null_parms parms; - - parms.rset_term = rset_term_create (term_dst, -1, rank_type); - return rset_create (rset_kind_null, &parms); - } - result = rset[0]; - for (i = 1; iscope, + rset_no, rset); } static RSET rpn_search_APT_local (ZebraHandle zh, Z_AttributesPlusTerm *zapt, - const char *termz, + const char *termz, oid_value attributeSet, - NMEM stream, - const char *rank_type) + NMEM stream, + const char *rank_type, NMEM rset_nmem) { RSET result; RSFD rsfd; struct it_key key; - rset_temp_parms parms; - - parms.rset_term = rset_term_create (termz, -1, rank_type); - parms.cmp = key_compare_it; - parms.key_size = sizeof (struct it_key); - parms.temp_path = res_get (zh->res, "setTmpDir"); - result = rset_create (rset_kind_temp, &parms); + int sys; + result = rstemp_create( rset_nmem,key_it_ctrl,key_it_ctrl->scope, + res_get (zh->res, "setTmpDir"),0 ); rsfd = rset_open (result, RSETF_WRITE); - key.sysno = atoi (termz); - key.seqno = 1; - if (key.sysno <= 0) - key.sysno = 1; - rset_write (result, rsfd, &key); - rset_close (result, rsfd); + sys = atoi(termz); + if (sys <= 0) + sys = 1; + key.mem[0] = sys; + key.mem[1] = 1; + key.len = 2; + rset_write (rsfd, &key); + rset_close (rsfd); return result; } static RSET rpn_sort_spec (ZebraHandle zh, Z_AttributesPlusTerm *zapt, - oid_value attributeSet, NMEM stream, - Z_SortKeySpecList *sort_sequence, - const char *rank_type) + oid_value attributeSet, NMEM stream, + Z_SortKeySpecList *sort_sequence, + const char *rank_type) { - rset_null_parms parms; int i; int sort_relation_value; AttrType sort_relation_type; @@ -1846,209 +1949,295 @@ static RSET rpn_sort_spec (ZebraHandle zh, Z_AttributesPlusTerm *zapt, if (!sort_sequence->specs) { - sort_sequence->num_specs = 10; - sort_sequence->specs = (Z_SortKeySpec **) - nmem_malloc (stream, sort_sequence->num_specs * - sizeof(*sort_sequence->specs)); - for (i = 0; inum_specs; i++) - sort_sequence->specs[i] = 0; + sort_sequence->num_specs = 10; + sort_sequence->specs = (Z_SortKeySpec **) + nmem_malloc(stream, sort_sequence->num_specs * + sizeof(*sort_sequence->specs)); + for (i = 0; inum_specs; i++) + sort_sequence->specs[i] = 0; } if (zapt->term->which != Z_Term_general) - i = 0; + i = 0; else - i = atoi_n ((char *) zapt->term->u.general->buf, - zapt->term->u.general->len); + i = atoi_n ((char *) zapt->term->u.general->buf, + zapt->term->u.general->len); if (i >= sort_sequence->num_specs) - i = 0; + i = 0; sprintf (termz, "%d", i); oe.proto = PROTO_Z3950; oe.oclass = CLASS_ATTSET; oe.value = attributeSet; if (!oid_ent_to_oid (&oe, oid)) - return 0; + return 0; - sks = (Z_SortKeySpec *) nmem_malloc (stream, sizeof(*sks)); + sks = (Z_SortKeySpec *) nmem_malloc(stream, sizeof(*sks)); sks->sortElement = (Z_SortElement *) - nmem_malloc (stream, sizeof(*sks->sortElement)); + nmem_malloc(stream, sizeof(*sks->sortElement)); sks->sortElement->which = Z_SortElement_generic; sk = sks->sortElement->u.generic = (Z_SortKey *) - nmem_malloc (stream, sizeof(*sk)); + nmem_malloc(stream, sizeof(*sk)); sk->which = Z_SortKey_sortAttributes; sk->u.sortAttributes = (Z_SortAttributes *) - nmem_malloc (stream, sizeof(*sk->u.sortAttributes)); + nmem_malloc(stream, sizeof(*sk->u.sortAttributes)); sk->u.sortAttributes->id = oid; sk->u.sortAttributes->list = (Z_AttributeList *) - nmem_malloc (stream, sizeof(*sk->u.sortAttributes->list)); + nmem_malloc(stream, sizeof(*sk->u.sortAttributes->list)); sk->u.sortAttributes->list->num_attributes = 1; sk->u.sortAttributes->list->attributes = (Z_AttributeElement **) - nmem_malloc (stream, sizeof(*sk->u.sortAttributes->list->attributes)); + nmem_malloc(stream, sizeof(*sk->u.sortAttributes->list->attributes)); ae = *sk->u.sortAttributes->list->attributes = (Z_AttributeElement *) - nmem_malloc (stream, sizeof(**sk->u.sortAttributes->list->attributes)); + nmem_malloc(stream, sizeof(**sk->u.sortAttributes->list->attributes)); ae->attributeSet = 0; - ae->attributeType = (int *) - nmem_malloc (stream, sizeof(*ae->attributeType)); + ae->attributeType = (int *) + nmem_malloc(stream, sizeof(*ae->attributeType)); *ae->attributeType = 1; ae->which = Z_AttributeValue_numeric; ae->value.numeric = (int *) - nmem_malloc (stream, sizeof(*ae->value.numeric)); + nmem_malloc(stream, sizeof(*ae->value.numeric)); *ae->value.numeric = use_value; sks->sortRelation = (int *) - nmem_malloc (stream, sizeof(*sks->sortRelation)); + nmem_malloc(stream, sizeof(*sks->sortRelation)); if (sort_relation_value == 1) - *sks->sortRelation = Z_SortRelation_ascending; + *sks->sortRelation = Z_SortKeySpec_ascending; else if (sort_relation_value == 2) - *sks->sortRelation = Z_SortRelation_descending; + *sks->sortRelation = Z_SortKeySpec_descending; else - *sks->sortRelation = Z_SortRelation_ascending; + *sks->sortRelation = Z_SortKeySpec_ascending; sks->caseSensitivity = (int *) - nmem_malloc (stream, sizeof(*sks->caseSensitivity)); + nmem_malloc(stream, sizeof(*sks->caseSensitivity)); *sks->caseSensitivity = 0; -#ifdef ASN_COMPILED sks->which = Z_SortKeySpec_null; sks->u.null = odr_nullval (); -#else - sks->missingValueAction = 0; -#endif + sort_sequence->specs[i] = sks; + return rsnull_create (NULL,key_it_ctrl); + /* FIXME - nmem?? */ +} + + +static int parse_xpath(ZebraHandle zh, Z_AttributesPlusTerm *zapt, + oid_value attributeSet, + struct xpath_location_step *xpath, int max, NMEM mem) +{ + oid_value curAttributeSet = attributeSet; + AttrType use; + const char *use_string = 0; + + attr_init (&use, zapt, 1); + attr_find_ex (&use, &curAttributeSet, &use_string); - sort_sequence->specs[i] = sks; + if (!use_string || *use_string != '/') + return -1; + + return zebra_parse_xpath_str(use_string, xpath, max, mem); +} + + - parms.rset_term = rset_term_create (termz, -1, rank_type); - return rset_create (rset_kind_null, &parms); +static RSET xpath_trunc(ZebraHandle zh, NMEM stream, + int reg_type, const char *term, int use, + oid_value curAttributeSet, NMEM rset_nmem) +{ + RSET rset; + struct grep_info grep_info; + char term_dict[2048]; + char ord_buf[32]; + int prefix_len = 0; + int ord = zebraExplain_lookup_attr_su(zh->reg->zei, curAttributeSet, use); + int ord_len, i, r, max_pos; + int term_type = Z_Term_characterString; + const char *flags = "void"; + + if (grep_info_prepare (zh, 0 /* zapt */, &grep_info, '0', stream)) + return rsnull_create (rset_nmem,key_it_ctrl); + + if (ord < 0) + return rsnull_create (rset_nmem,key_it_ctrl); + if (prefix_len) + term_dict[prefix_len++] = '|'; + else + term_dict[prefix_len++] = '('; + + ord_len = key_SU_encode (ord, ord_buf); + for (i = 0; ireg->dict, term_dict, 0, + &grep_info, &max_pos, 0, grep_handle); + yaz_log (YLOG_DEBUG, "%s %d positions", term, + grep_info.isam_p_indx); + rset = rset_trunc(zh, grep_info.isam_p_buf, + grep_info.isam_p_indx, term, strlen(term), + flags, 1, term_type,rset_nmem, + key_it_ctrl, key_it_ctrl->scope); + grep_info_delete (&grep_info); + return rset; } -static RSET rpn_search_xpath (ZebraHandle zh, Z_AttributesPlusTerm *zapt, +static RSET rpn_search_xpath (ZebraHandle zh, oid_value attributeSet, int num_bases, char **basenames, - NMEM stream, const char *rank_type, RSET rset) + NMEM stream, const char *rank_type, RSET rset, + int xpath_len, struct xpath_location_step *xpath, + NMEM rset_nmem) { - AttrType use; - const char *use_string = 0; oid_value curAttributeSet = attributeSet; - char term_dict[2048]; int base_no; - struct grep_info grep_info; - struct rpn_char_map_info rcmi; - - yaz_log (LOG_LOG, "rpn_search_xpath 1"); - attr_init (&use, zapt, 1); - attr_find_ex (&use, &curAttributeSet, &use_string); + int i; - if (curAttributeSet != VAL_IDXPATH) - { - yaz_log (LOG_LOG, "rpn_search_xpath - not 1"); + if (xpath_len < 0) return rset; - } - if (!use_string) + + yaz_log (YLOG_DEBUG, "xpath len=%d", xpath_len); + for (i = 0; ireg, '0', &rcmi); + curAttributeSet = VAL_IDXPATH; - if (grep_info_prepare (zh, zapt, &grep_info, '0', stream)) - return 0; + /* + //a -> a/.* + //a/b -> b/a/.* + /a -> a/ + /a/b -> b/a/ + + / -> none + + a[@attr = value]/b[@other = othervalue] + + /e/@a val range(e/,range(@a,freetext(w,1015,val),@a),e/) + /a/b val range(b/a/,freetext(w,1016,val),b/a/) + /a/b/@c val range(b/a/,range(@c,freetext(w,1016,val),@c),b/a/) + /a/b[@c = y] val range(b/a/,freetext(w,1016,val),b/a/,@c = y) + /a[@c = y]/b val range(a/,range(b/a/,freetext(w,1016,val),b/a/),a/,@c = y) + /a[@c = x]/b[@c = y] range(a/,range(b/a/,freetext(w,1016,val),b/a/,@c = y),a/,@c = x) + + */ + + dict_grep_cmap (zh->reg->dict, 0, 0); - yaz_log (LOG_LOG, "rpn_search_xpath 2"); for (base_no = 0; base_no < num_bases; base_no++) { - rset_between_parms parms; - RSET rset_start_tag, rset_end_tag; - int ord, ord_len, i, r, max_pos; - int prefix_len ; - char ord_buf[32]; + int level = xpath_len; + int first_path = 1; + if (zebraExplain_curDatabase (zh->reg->zei, basenames[base_no])) { zh->errCode = 109; /* Database unavailable */ zh->errString = basenames[base_no]; return rset; } - - prefix_len = 0; - ord = zebraExplain_lookupSU (zh->reg->zei, curAttributeSet, 1); - if (ord < 0) - continue; - if (prefix_len) - term_dict[prefix_len++] = '|'; - else - term_dict[prefix_len++] = '('; - - ord_len = key_SU_encode (ord, ord_buf); - for (i = 0; ireg->dict, term_dict, 0, - &grep_info, &max_pos, 0, grep_handle); - yaz_log (LOG_LOG, "%s %d positions", use_string, - grep_info.isam_p_indx); - rset_start_tag = - rset_trunc (zh, grep_info.isam_p_buf, - grep_info.isam_p_indx, use_string, strlen(use_string), - rank_type, 1); - - prefix_len = 0; - ord = zebraExplain_lookupSU (zh->reg->zei, curAttributeSet, 2); - if (ord < 0) - continue; - if (prefix_len) - term_dict[prefix_len++] = '|'; - else - term_dict[prefix_len++] = '('; - - ord_len = key_SU_encode (ord, ord_buf); - for (i = 0; i= 0) { - term_dict[prefix_len++] = 1; - term_dict[prefix_len++] = ord_buf[i]; + char xpath_rev[128]; + int i, len; + RSET rset_start_tag = 0, rset_end_tag = 0, rset_attr = 0; + + *xpath_rev = 0; + len = 0; + for (i = level; i >= 1; --i) + { + const char *cp = xpath[i].part; + if (*cp) + { + for (;*cp; cp++) + if (*cp == '*') + { + memcpy (xpath_rev + len, "[^/]*", 5); + len += 5; + } + else if (*cp == ' ') + { + + xpath_rev[len++] = 1; + xpath_rev[len++] = ' '; + } + + else + xpath_rev[len++] = *cp; + xpath_rev[len++] = '/'; + } + else if (i == 1) /* // case */ + { + xpath_rev[len++] = '.'; + xpath_rev[len++] = '*'; + } + } + xpath_rev[len] = 0; + + if (xpath[level].predicate && + xpath[level].predicate->which == XPATH_PREDICATE_RELATION && + xpath[level].predicate->u.relation.name[0]) + { + WRBUF wbuf = wrbuf_alloc(); + wrbuf_puts(wbuf, xpath[level].predicate->u.relation.name+1); + if (xpath[level].predicate->u.relation.value) + { + const char *cp = xpath[level].predicate->u.relation.value; + wrbuf_putc(wbuf, '='); + + while (*cp) + { + if (strchr(REGEX_CHARS, *cp)) + wrbuf_putc(wbuf, '\\'); + wrbuf_putc(wbuf, *cp); + cp++; + } + } + wrbuf_puts(wbuf, ""); + rset_attr = xpath_trunc( + zh, stream, '0', wrbuf_buf(wbuf), 3, + curAttributeSet,rset_nmem); + wrbuf_free(wbuf, 1); + } + else + { + if (!first_path) + continue; + } + yaz_log (log_level_rpn, "xpath_rev (%d) = %s", level, xpath_rev); + if (strlen(xpath_rev)) + { + rset_start_tag = xpath_trunc(zh, stream, '0', + xpath_rev, 1, curAttributeSet, rset_nmem); + + rset_end_tag = xpath_trunc(zh, stream, '0', + xpath_rev, 2, curAttributeSet, rset_nmem); + + rset = rsbetween_create(rset_nmem, key_it_ctrl, + key_it_ctrl->scope, + rset_start_tag, rset, + rset_end_tag, rset_attr); + } + first_path = 0; } - term_dict[prefix_len++] = ')'; - term_dict[prefix_len++] = 1; - term_dict[prefix_len++] = '0'; - strcpy (term_dict+prefix_len, use_string); - grep_info.isam_p_indx = 0; - r = dict_lookup_grep (zh->reg->dict, term_dict, 0, - &grep_info, &max_pos, 0, grep_handle); - - yaz_log (LOG_LOG, "%s %d positions", use_string, - grep_info.isam_p_indx); - rset_end_tag = - rset_trunc (zh, grep_info.isam_p_buf, - grep_info.isam_p_indx, use_string, strlen(use_string), - rank_type, 1); - - parms.key_size = sizeof(struct it_key); - parms.cmp = key_compare_it; - parms.rset_l = rset_start_tag; - parms.rset_m = rset; - parms.rset_r = rset_end_tag; - parms.printer = key_print_it; - yaz_log (LOG_LOG, "rpn_search_xpath 4"); - rset = rset_create (rset_kind_between, &parms); } - grep_info_delete (&grep_info); return rset; } -static RSET rpn_search_APT (ZebraHandle zh, Z_AttributesPlusTerm *zapt, - oid_value attributeSet, NMEM stream, - Z_SortKeySpecList *sort_sequence, - int num_bases, char **basenames) +static RSET rpn_search_APT(ZebraHandle zh, Z_AttributesPlusTerm *zapt, + oid_value attributeSet, NMEM stream, + Z_SortKeySpecList *sort_sequence, + int num_bases, char **basenames, + NMEM rset_nmem) { unsigned reg_id; char *search_type = NULL; @@ -2057,151 +2246,151 @@ static RSET rpn_search_APT (ZebraHandle zh, Z_AttributesPlusTerm *zapt, int sort_flag; char termz[IT_MAX_WORD+1]; RSET rset = 0; + int xpath_len; + int xpath_use = 0; + struct xpath_location_step xpath[10]; - zebra_maps_attr (zh->reg->zebra_maps, zapt, ®_id, &search_type, - rank_type, &complete_flag, &sort_flag); - - logf (LOG_DEBUG, "reg_id=%c", reg_id); - logf (LOG_DEBUG, "complete_flag=%d", complete_flag); - logf (LOG_DEBUG, "search_type=%s", search_type); - logf (LOG_DEBUG, "rank_type=%s", rank_type); - - if (zapt->term->which != Z_Term_general) + if (!log_level_set) { - zh->errCode = 124; - return NULL; + log_level_rpn = yaz_log_module_level("rpn"); + log_level_set = 1; } - trans_term (zh, zapt, termz); + zebra_maps_attr(zh->reg->zebra_maps, zapt, ®_id, &search_type, + rank_type, &complete_flag, &sort_flag); + + yaz_log(YLOG_DEBUG, "reg_id=%c", reg_id); + yaz_log(YLOG_DEBUG, "complete_flag=%d", complete_flag); + yaz_log(YLOG_DEBUG, "search_type=%s", search_type); + yaz_log(YLOG_DEBUG, "rank_type=%s", rank_type); + + if (zapt_term_to_utf8(zh, zapt, termz)) + return 0; if (sort_flag) - return rpn_sort_spec (zh, zapt, attributeSet, stream, sort_sequence, - rank_type); + return rpn_sort_spec(zh, zapt, attributeSet, stream, sort_sequence, + rank_type); + xpath_len = parse_xpath(zh, zapt, attributeSet, xpath, 10, stream); + if (xpath_len >= 0) + { + xpath_use = 1016; + if (xpath[xpath_len-1].part[0] == '@') + xpath_use = 1015; + } if (!strcmp (search_type, "phrase")) { - rset = rpn_search_APT_phrase (zh, zapt, termz, attributeSet, stream, - reg_id, complete_flag, rank_type, - num_bases, basenames); + rset = rpn_search_APT_phrase(zh, zapt, termz, attributeSet, stream, + reg_id, complete_flag, rank_type, + xpath_use, + num_bases, basenames, rset_nmem); } else if (!strcmp (search_type, "and-list")) { - rset = rpn_search_APT_and_list (zh, zapt, termz, attributeSet, stream, - reg_id, complete_flag, rank_type, - num_bases, basenames); + rset = rpn_search_APT_and_list(zh, zapt, termz, attributeSet, stream, + reg_id, complete_flag, rank_type, + xpath_use, + num_bases, basenames, rset_nmem); } else if (!strcmp (search_type, "or-list")) { - rset = rpn_search_APT_or_list (zh, zapt, termz, attributeSet, stream, - reg_id, complete_flag, rank_type, - num_bases, basenames); + rset = rpn_search_APT_or_list(zh, zapt, termz, attributeSet, stream, + reg_id, complete_flag, rank_type, + xpath_use, + num_bases, basenames, rset_nmem); } else if (!strcmp (search_type, "local")) { - rset = rpn_search_APT_local (zh, zapt, termz, attributeSet, stream, - rank_type); + rset = rpn_search_APT_local(zh, zapt, termz, attributeSet, stream, + rank_type, rset_nmem); } else if (!strcmp (search_type, "numeric")) { - rset = rpn_search_APT_numeric (zh, zapt, termz, attributeSet, stream, - reg_id, complete_flag, rank_type, - num_bases, basenames); + rset = rpn_search_APT_numeric(zh, zapt, termz, attributeSet, stream, + reg_id, complete_flag, rank_type, + xpath_use, + num_bases, basenames, rset_nmem); + } + else if (!strcmp (search_type, "always")) + { + rset = 0; } else + { zh->errCode = 118; - return rpn_search_xpath (zh, zapt, attributeSet, num_bases, basenames, - stream, rank_type, rset); + return 0; + } + return rpn_search_xpath (zh, attributeSet, num_bases, basenames, + stream, rank_type, rset, + xpath_len, xpath, rset_nmem); } -static RSET rpn_search_structure (ZebraHandle zh, Z_RPNStructure *zs, - oid_value attributeSet, NMEM stream, - Z_SortKeySpecList *sort_sequence, - int num_bases, char **basenames) +RSET rpn_search_structure(ZebraHandle zh, Z_RPNStructure *zs, + oid_value attributeSet, + NMEM stream, NMEM rset_nmem, + Z_SortKeySpecList *sort_sequence, + int num_bases, char **basenames) { RSET r = NULL; if (zs->which == Z_RPNStructure_complex) { Z_Operator *zop = zs->u.complex->roperator; - rset_bool_parms bool_parms; + RSET rsets[2]; /* l and r argument */ - bool_parms.rset_l = rpn_search_structure (zh, zs->u.complex->s1, - attributeSet, stream, - sort_sequence, - num_bases, basenames); - if (bool_parms.rset_l == NULL) + rsets[0] = rpn_search_structure(zh, zs->u.complex->s1, + attributeSet, stream, rset_nmem, + sort_sequence, + num_bases, basenames); + if (rsets[0] == NULL) return NULL; - bool_parms.rset_r = rpn_search_structure (zh, zs->u.complex->s2, - attributeSet, stream, - sort_sequence, - num_bases, basenames); - if (bool_parms.rset_r == NULL) + rsets[1] = rpn_search_structure(zh, zs->u.complex->s2, + attributeSet, stream, rset_nmem, + sort_sequence, + num_bases, basenames); + if (rsets[1] == NULL) { - rset_delete (bool_parms.rset_l); + rset_delete(rsets[0]); return NULL; } - bool_parms.key_size = sizeof(struct it_key); - bool_parms.cmp = key_compare_it; switch (zop->which) { case Z_Operator_and: - r = rset_create (rset_kind_and, &bool_parms); + r = rsmulti_and_create(rset_nmem, key_it_ctrl, key_it_ctrl->scope, + 2, rsets); break; case Z_Operator_or: - r = rset_create (rset_kind_or, &bool_parms); + r = rsmulti_or_create(rset_nmem, key_it_ctrl, key_it_ctrl->scope, + 2, rsets); break; case Z_Operator_and_not: - r = rset_create (rset_kind_not, &bool_parms); + r = rsbool_create_not(rset_nmem,key_it_ctrl, key_it_ctrl->scope, + rsets[0], rsets[1]); break; case Z_Operator_prox: -#ifdef ASN_COMPILED if (zop->u.prox->which != Z_ProximityOperator_known) { zh->errCode = 132; return NULL; } -#else - if (zop->u.prox->which != Z_ProxCode_known) - { - zh->errCode = 132; - return NULL; - } -#endif - -#ifdef ASN_COMPILED if (*zop->u.prox->u.known != Z_ProxUnit_word) { - char *val = (char *) nmem_malloc (stream, 16); + char *val = (char *) nmem_malloc(stream, 16); zh->errCode = 132; zh->errString = val; sprintf (val, "%d", *zop->u.prox->u.known); return NULL; } -#else - if (*zop->u.prox->proximityUnitCode != Z_ProxUnit_word) + else { - char *val = (char *) nmem_malloc (stream, 16); - zh->errCode = 132; - zh->errString = val; - sprintf (val, "%d", *zop->u.prox->proximityUnitCode); - return NULL; + /* new / old prox */ + r = rsprox_create(rset_nmem,key_it_ctrl,key_it_ctrl->scope, + 2, rsets, + *zop->u.prox->ordered, + (!zop->u.prox->exclusion ? + 0 : *zop->u.prox->exclusion), + *zop->u.prox->relationType, + *zop->u.prox->distance ); } -#endif - else - { - RSET rsets[2]; - - rsets[0] = bool_parms.rset_l; - rsets[1] = bool_parms.rset_r; - - r = rpn_prox (zh, rsets, 2, - *zop->u.prox->ordered, - (!zop->u.prox->exclusion ? 0 : - *zop->u.prox->exclusion), - *zop->u.prox->relationType, - *zop->u.prox->distance); - rset_delete (rsets[0]); - rset_delete (rsets[1]); - } break; default: zh->errCode = 110; @@ -2212,23 +2401,23 @@ static RSET rpn_search_structure (ZebraHandle zh, Z_RPNStructure *zs, { if (zs->u.simple->which == Z_Operand_APT) { - logf (LOG_DEBUG, "rpn_search_APT"); - r = rpn_search_APT (zh, zs->u.simple->u.attributesPlusTerm, - attributeSet, stream, sort_sequence, - num_bases, basenames); + yaz_log(YLOG_DEBUG, "rpn_search_APT"); + r = rpn_search_APT(zh, zs->u.simple->u.attributesPlusTerm, + attributeSet, stream, sort_sequence, + num_bases, basenames,rset_nmem); } else if (zs->u.simple->which == Z_Operand_resultSetId) { - logf (LOG_DEBUG, "rpn_search_ref"); + yaz_log(YLOG_DEBUG, "rpn_search_ref"); r = resultSetRef (zh, zs->u.simple->u.resultSetId); - if (!r) - { - r = rset_create (rset_kind_null, NULL); - zh->errCode = 30; - zh->errString = - nmem_strdup (stream, zs->u.simple->u.resultSetId); - return 0; - } + if (!r) + { + zh->errCode = 30; + zh->errString = + nmem_strdup (stream, zs->u.simple->u.resultSetId); + return 0; + } + rset_dup(r); } else { @@ -2244,62 +2433,9 @@ static RSET rpn_search_structure (ZebraHandle zh, Z_RPNStructure *zs, return r; } - -RSET rpn_search (ZebraHandle zh, NMEM nmem, - Z_RPNQuery *rpn, int num_bases, char **basenames, - const char *setname, - ZebraSet sset) -{ - RSET rset; - oident *attrset; - oid_value attributeSet; - Z_SortKeySpecList *sort_sequence; - int sort_status, i; - - zh->errCode = 0; - zh->errString = NULL; - zh->hits = 0; - - sort_sequence = (Z_SortKeySpecList *) - nmem_malloc (nmem, sizeof(*sort_sequence)); - sort_sequence->num_specs = 10; - sort_sequence->specs = (Z_SortKeySpec **) - nmem_malloc (nmem, sort_sequence->num_specs * - sizeof(*sort_sequence->specs)); - for (i = 0; inum_specs; i++) - sort_sequence->specs[i] = 0; - - attrset = oid_getentbyoid (rpn->attributeSetId); - attributeSet = attrset->value; - rset = rpn_search_structure (zh, rpn->RPNStructure, attributeSet, - nmem, sort_sequence, num_bases, basenames); - if (!rset) - return 0; - - if (zh->errCode) - logf (LOG_DEBUG, "search error: %d", zh->errCode); - - for (i = 0; sort_sequence->specs[i]; i++) - ; - sort_sequence->num_specs = i; - if (!i) - resultSetRank (zh, sset, rset); - else - { - logf (LOG_DEBUG, "resultSetSortSingle in rpn_search"); - resultSetSortSingle (zh, nmem, sset, rset, - sort_sequence, &sort_status); - if (zh->errCode) - { - logf (LOG_DEBUG, "resultSetSortSingle status = %d", zh->errCode); - } - } - return rset; -} - struct scan_info_entry { char *term; - ISAMS_P isam_p; + ISAM_P isam_p; }; struct scan_info { @@ -2321,54 +2457,74 @@ static int scan_handle (char *name, const char *info, int pos, void *client) else idx = - pos - 1; scan_info->list[idx].term = (char *) - odr_malloc (scan_info->odr, strlen(name + len_prefix)+1); - strcpy (scan_info->list[idx].term, name + len_prefix); - assert (*info == sizeof(ISAMS_P)); - memcpy (&scan_info->list[idx].isam_p, info+1, sizeof(ISAMS_P)); + odr_malloc(scan_info->odr, strlen(name + len_prefix)+1); + strcpy(scan_info->list[idx].term, name + len_prefix); + assert (*info == sizeof(ISAM_P)); + memcpy (&scan_info->list[idx].isam_p, info+1, sizeof(ISAM_P)); return 0; } static void scan_term_untrans (ZebraHandle zh, NMEM stream, int reg_type, - char **dst, const char *src) + char **dst, const char *src) { - char term_dst[1024]; - - term_untrans (zh, reg_type, term_dst, src); + char term_src[IT_MAX_WORD]; + char term_dst[IT_MAX_WORD]; - *dst = (char *) nmem_malloc (stream, strlen(term_dst)+1); - strcpy (*dst, term_dst); + term_untrans (zh, reg_type, term_src, src); + + if (zh->iconv_from_utf8 != 0) + { + int len; + char *inbuf = term_src; + size_t inleft = strlen(term_src); + char *outbuf = term_dst; + size_t outleft = sizeof(term_dst)-1; + size_t ret; + + ret = yaz_iconv (zh->iconv_from_utf8, &inbuf, &inleft, + &outbuf, &outleft); + if (ret == (size_t)(-1)) + len = 0; + else + len = outbuf - term_dst; + *dst = nmem_malloc(stream, len + 1); + if (len > 0) + memcpy (*dst, term_dst, len); + (*dst)[len] = '\0'; + } + else + *dst = nmem_strdup(stream, term_src); } static void count_set (RSET r, int *count) { - int psysno = 0; + zint psysno = 0; int kno = 0; struct it_key key; RSFD rfd; - int term_index; - logf (LOG_DEBUG, "count_set"); + yaz_log(YLOG_DEBUG, "count_set"); *count = 0; rfd = rset_open (r, RSETF_READ); - while (rset_read (r, rfd, &key, &term_index)) + while (rset_read (rfd, &key,0 /* never mind terms */)) { - if (key.sysno != psysno) + if (key.mem[0] != psysno) { - psysno = key.sysno; + psysno = key.mem[0]; (*count)++; } kno++; } - rset_close (r, rfd); - logf (LOG_DEBUG, "%d keys, %d records", kno, *count); + rset_close (rfd); + yaz_log(YLOG_DEBUG, "%d keys, %d records", kno, *count); } -void rpn_scan (ZebraHandle zh, ODR stream, Z_AttributesPlusTerm *zapt, - oid_value attributeset, - int num_bases, char **basenames, - int *position, int *num_entries, ZebraScanEntry **list, - int *is_partial) +ZEBRA_RES rpn_scan(ZebraHandle zh, ODR stream, Z_AttributesPlusTerm *zapt, + oid_value attributeset, + int num_bases, char **basenames, + int *position, int *num_entries, ZebraScanEntry **list, + int *is_partial, RSET limit_set, int return_zero) { int i; int pos = *position; @@ -2379,114 +2535,182 @@ void rpn_scan (ZebraHandle zh, ODR stream, Z_AttributesPlusTerm *zapt, char termz[IT_MAX_WORD+20]; AttrType use; int use_value; + const char *use_string = 0; struct scan_info *scan_info_array; ZebraScanEntry *glist; int ords[32], ord_no = 0; int ptr[32]; + int bases_ok = 0; /* no of databases with OK attribute */ + int errCode = 0; /* err code (if any is not OK) */ + char *errString = 0; /* addinfo */ + unsigned reg_id; char *search_type = NULL; char rank_type[128]; int complete_flag; int sort_flag; + NMEM rset_nmem = NULL; + *list = 0; + *is_partial = 0; if (attributeset == VAL_NONE) attributeset = VAL_BIB1; - yaz_log (LOG_DEBUG, "position = %d, num = %d set=%d", + if (!limit_set) + { + AttrType termset; + int termset_value_numeric; + const char *termset_value_string; + attr_init (&termset, zapt, 8); + termset_value_numeric = + attr_find_ex (&termset, NULL, &termset_value_string); + if (termset_value_numeric != -1) + { + char resname[32]; + const char *termset_name = 0; + + if (termset_value_numeric != -2) + { + + sprintf (resname, "%d", termset_value_numeric); + termset_name = resname; + } + else + termset_name = termset_value_string; + + limit_set = resultSetRef (zh, termset_name); + } + } + + yaz_log (YLOG_DEBUG, "position = %d, num = %d set=%d", pos, num, attributeset); attr_init (&use, zapt, 1); - use_value = attr_find (&use, &attributeset); + use_value = attr_find_ex (&use, &attributeset, &use_string); if (zebra_maps_attr (zh->reg->zebra_maps, zapt, ®_id, &search_type, - rank_type, &complete_flag, &sort_flag)) + rank_type, &complete_flag, &sort_flag)) { - *num_entries = 0; - zh->errCode = 113; - return ; + *num_entries = 0; + zh->errCode = 113; + return ZEBRA_FAIL; } - yaz_log (LOG_DEBUG, "use_value = %d", use_value); + yaz_log (YLOG_DEBUG, "use_value = %d", use_value); if (use_value == -1) use_value = 1016; for (base_no = 0; base_no < num_bases && ord_no < 32; base_no++) { - int r; - attent attp; - data1_local_attribute *local_attr; + data1_local_attribute *local_attr; + attent attp; + int ord; - if ((r=att_getentbyatt (zh, &attp, attributeset, use_value))) - { - logf (LOG_DEBUG, "att_getentbyatt fail. set=%d use=%d", - attributeset, use_value); - if (r == -1) - zh->errCode = 114; - else - zh->errCode = 121; - *num_entries = 0; - return; - } - if (zebraExplain_curDatabase (zh->reg->zei, basenames[base_no])) - { - zh->errString = basenames[base_no]; + if (zebraExplain_curDatabase (zh->reg->zei, basenames[base_no])) + { + zh->errString = basenames[base_no]; zh->errCode = 109; /* Database unavailable */ *num_entries = 0; - return; - } - for (local_attr = attp.local_attributes; local_attr && ord_no < 32; - local_attr = local_attr->next) - { - int ord; + return ZEBRA_FAIL; + } - ord = zebraExplain_lookupSU (zh->reg->zei, attp.attset_ordinal, - local_attr->local); - if (ord > 0) - ords[ord_no++] = ord; - } + if (use_string && + (ord = zebraExplain_lookup_attr_str(zh->reg->zei, + use_string)) >= 0) + { + /* we have a match for a raw string attribute */ + if (ord > 0) + ords[ord_no++] = ord; + attp.local_attributes = 0; /* no more attributes */ + } + else + { + int r; + + if ((r = att_getentbyatt (zh, &attp, attributeset, use_value, + use_string))) + { + yaz_log(YLOG_DEBUG, "att_getentbyatt fail. set=%d use=%d", + attributeset, use_value); + if (r == -1) + { + errCode = 114; + if (use_string) + errString = odr_strdup(stream, use_string); + else + { + char val_str[32]; + sprintf (val_str, "%d", use_value); + errString = odr_strdup(stream, val_str); + } + } + else + errCode = 121; + continue; + } + } + bases_ok++; + for (local_attr = attp.local_attributes; local_attr && ord_no < 32; + local_attr = local_attr->next) + { + ord = zebraExplain_lookup_attr_su(zh->reg->zei, + attp.attset_ordinal, + local_attr->local); + if (ord > 0) + ords[ord_no++] = ord; + } + } + if (!bases_ok && errCode) + { + zh->errCode = errCode; + zh->errString = errString; + *num_entries = 0; + return ZEBRA_FAIL; } if (ord_no == 0) { - *num_entries = 0; - zh->errCode = 113; - return; + *num_entries = 0; + return ZEBRA_OK; } /* prepare dictionary scanning */ before = pos-1; after = 1+num-pos; scan_info_array = (struct scan_info *) - odr_malloc (stream, ord_no * sizeof(*scan_info_array)); + odr_malloc(stream, ord_no * sizeof(*scan_info_array)); for (i = 0; i < ord_no; i++) { int j, prefix_len = 0; int before_tmp = before, after_tmp = after; struct scan_info *scan_info = scan_info_array + i; - struct rpn_char_map_info rcmi; + struct rpn_char_map_info rcmi; - rpn_char_map_prepare (zh->reg, reg_id, &rcmi); + rpn_char_map_prepare (zh->reg, reg_id, &rcmi); scan_info->before = before; scan_info->after = after; scan_info->odr = stream; scan_info->list = (struct scan_info_entry *) - odr_malloc (stream, (before+after) * sizeof(*scan_info->list)); + odr_malloc(stream, (before+after) * sizeof(*scan_info->list)); for (j = 0; jlist[j].term = NULL; - prefix_len += key_SU_encode (ords[i], termz + prefix_len); + prefix_len += key_SU_encode (ords[i], termz + prefix_len); termz[prefix_len++] = reg_id; termz[prefix_len] = 0; - strcpy (scan_info->prefix, termz); + strcpy(scan_info->prefix, termz); - trans_scan_term (zh, zapt, termz+prefix_len, reg_id); - - dict_scan (zh->reg->dict, termz, &before_tmp, &after_tmp, - scan_info, scan_handle); + if (trans_scan_term(zh, zapt, termz+prefix_len, reg_id)) + return ZEBRA_FAIL; + + dict_scan(zh->reg->dict, termz, &before_tmp, &after_tmp, + scan_info, scan_handle); } glist = (ZebraScanEntry *) - odr_malloc (stream, (before+after)*sizeof(*glist)); + odr_malloc(stream, (before+after)*sizeof(*glist)); + + rset_nmem = nmem_create(); /* consider terms after main term */ for (i = 0; i < ord_no; i++) @@ -2503,7 +2727,7 @@ void rpn_scan (ZebraHandle zh, ODR stream, Z_AttributesPlusTerm *zapt, for (j = 0; j < ord_no; j++) { if (ptr[j] < before+after && - (tst=scan_info_array[j].list[ptr[j]].term) && + (tst = scan_info_array[j].list[ptr[j]].term) && (!mterm || strcmp (tst, mterm) < 0)) { j0 = j; @@ -2513,105 +2737,118 @@ void rpn_scan (ZebraHandle zh, ODR stream, Z_AttributesPlusTerm *zapt, if (j0 == -1) break; scan_term_untrans (zh, stream->mem, reg_id, - &glist[i+before].term, mterm); - rset = rset_trunc (zh, &scan_info_array[j0].list[ptr[j0]].isam_p, 1, - glist[i+before].term, strlen(glist[i+before].term), - NULL, 0); - + &glist[i+before].term, mterm); + rset = rset_trunc(zh, &scan_info_array[j0].list[ptr[j0]].isam_p, 1, + glist[i+before].term, strlen(glist[i+before].term), + NULL, 0, zapt->term->which, rset_nmem, + key_it_ctrl,key_it_ctrl->scope); ptr[j0]++; for (j = j0+1; jterm->which,rset_nmem, + key_it_ctrl, key_it_ctrl->scope); + rset = rsmulti_or_create(rset_nmem, key_it_ctrl, + 2, key_it_ctrl->scope, rsets); ptr[j]++; } } - count_set (rset, &glist[i+before].occurrences); - rset_delete (rset); + if (limit_set) + { + RSET rsets[2]; + rsets[0] = rset; + rsets[1] = rset_dup(limit_set); + + rset = rsmulti_and_create(rset_nmem, key_it_ctrl, + key_it_ctrl->scope, 2, rsets); + } + count_set(rset, &glist[i+before].occurrences); + rset_delete(rset); } if (i < after) { - *num_entries -= (after-i); - *is_partial = 1; + *num_entries -= (after-i); + *is_partial = 1; } - + /* consider terms before main term */ for (i = 0; i 0)) - { - j0 = j; - mterm = tst; - } - } - if (j0 == -1) - break; - - scan_term_untrans (zh, stream->mem, reg_id, + int j, j0 = -1; + const char *mterm = NULL; + const char *tst; + RSET rset; + + for (j = 0; j 0)) + { + j0 = j; + mterm = tst; + } + } + if (j0 == -1) + break; + + scan_term_untrans (zh, stream->mem, reg_id, &glist[before-1-i].term, mterm); - - rset = rset_trunc - (zh, &scan_info_array[j0].list[before-1-ptr[j0]].isam_p, 1, - glist[before-1-i].term, strlen(glist[before-1-i].term), - NULL, 0); - - ptr[j0]++; - - for (j = j0+1; jterm->which,rset_nmem, + key_it_ctrl,key_it_ctrl->scope); + + ptr[j0]++; + + for (j = j0+1; jterm->which, rset_nmem, + key_it_ctrl, key_it_ctrl->scope); + rset = rsmulti_or_create(rset_nmem, key_it_ctrl, + 2, key_it_ctrl->scope, rsets); + + ptr[j]++; + } + } + if (limit_set) + { + RSET rsets[2]; + rsets[0] = rset; + rsets[1] = rset_dup(limit_set); + + rset = rsmulti_and_create(rset_nmem, key_it_ctrl, + key_it_ctrl->scope, 2, rsets); + } + count_set (rset, &glist[before-1-i].occurrences); + rset_delete (rset); } i = before-i; if (i) @@ -2620,11 +2857,12 @@ void rpn_scan (ZebraHandle zh, ODR stream, Z_AttributesPlusTerm *zapt, *position -= i; *num_entries -= i; } + + nmem_destroy(rset_nmem); *list = glist + i; /* list is set to first 'real' entry */ - logf (LOG_DEBUG, "position = %d, num_entries = %d", - *position, *num_entries); - if (zh->errCode) - logf (LOG_DEBUG, "scan error: %d", zh->errCode); + yaz_log(YLOG_DEBUG, "position = %d, num_entries = %d", + *position, *num_entries); + return ZEBRA_OK; } - +