-/* $Id: rpnsearch.c,v 1.14 2007-05-14 14:05:21 adam Exp $
- Copyright (C) 1995-2007
- Index Data ApS
-
-This file is part of the Zebra server.
+/* This file is part of the Zebra server.
+ Copyright (C) 1994-2010 Index Data
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
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;
- const char **out = zebra_maps_input(p->zm, p->reg_type, from, len, 0);
+ const char **out = zebra_maps_input(p->zm, from, len, 0);
#if 0
if (out && *out)
{
return out;
}
-void rpn_char_map_prepare(struct zebra_register *reg, int reg_type,
+void rpn_char_map_prepare(struct zebra_register *reg, zebra_map_t zm,
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);
+ map_info->zm = zm;
+ if (zebra_maps_is_icu(zm))
+ dict_grep_cmap(reg->dict, 0, 0);
+ else
+ dict_grep_cmap(reg->dict, map_info, rpn_char_map_handler);
}
#define TERM_COUNT
int isam_p_indx;
int trunc_max;
ZebraHandle zh;
- int reg_type;
+ const char *index_type;
ZebraSet termset;
};
if (p->isam_p_buf)
{
memcpy(new_isam_p_buf, p->isam_p_buf,
- p->isam_p_indx * sizeof(*p->isam_p_buf));
+ p->isam_p_indx * sizeof(*p->isam_p_buf));
xfree(p->isam_p_buf);
}
p->isam_p_buf = new_isam_p_buf;
if (p->term_no)
{
memcpy(new_term_no, p->isam_p_buf,
- p->isam_p_indx * sizeof(*p->term_no));
+ p->isam_p_indx * sizeof(*p->term_no));
xfree(p->term_no);
}
p->term_no = new_term_no;
char term_tmp[IT_MAX_WORD];
int ord = 0;
const char *index_name;
- int len = key_SU_decode (&ord, (const unsigned char *) name);
+ int len = key_SU_decode(&ord, (const unsigned char *) name);
- zebra_term_untrans (p->zh, p->reg_type, term_tmp, name+len);
+ zebra_term_untrans (p->zh, p->index_type, term_tmp, name+len);
yaz_log(log_level_rpn, "grep: %d %c %s", ord, name[len], term_tmp);
zebraExplain_lookup_ord(p->zh->reg->zei,
ord, 0 /* index_type */, &db, &index_name);
return add_isam_p(name, info, (struct grep_info *) p);
}
-static int term_pre(ZebraMaps zebra_maps, int reg_type, const char **src,
+static int term_pre(zebra_map_t zm, const char **src,
const char *ct1, const char *ct2, int first)
{
const char *s1, *s0 = *src;
if (ct2 && strchr(ct2, *s0))
break;
s1 = s0;
- map = zebra_maps_input(zebra_maps, reg_type, &s1, strlen(s1), first);
+ map = zebra_maps_input(zm, &s1, strlen(s1), first);
if (**map != *CHR_SPACE)
break;
s0 = s1;
}
}
-#define REGEX_CHARS " []()|.*+?!"
+#define REGEX_CHARS " ^[]()|.*+?!\"$"
static void add_non_space(const char *start, const char *end,
WRBUF term_dict,
- char *dst_term, int *dst_ptr,
+ WRBUF display_term,
const char **map, int q_map_match)
{
size_t sz = end - start;
- memcpy(dst_term + *dst_ptr, start, sz);
- (*dst_ptr) += sz;
+
+ wrbuf_write(display_term, start, sz);
if (!q_map_match)
{
while (start < end)
}
}
+
+static int term_100_icu(zebra_map_t zm,
+ const char **src, WRBUF term_dict, int space_split,
+ WRBUF display_term,
+ int right_trunc)
+{
+ int i;
+ const char *res_buf = 0;
+ size_t res_len = 0;
+ const char *display_buf;
+ size_t display_len;
+ if (!zebra_map_tokenize_next(zm, &res_buf, &res_len,
+ &display_buf, &display_len))
+ {
+ *src += strlen(*src);
+ return 0;
+ }
+ wrbuf_write(display_term, display_buf, display_len);
+ if (right_trunc)
+ {
+ /* ICU sort keys seem to be of the form
+ basechars \x01 accents \x01 length
+ For now we'll just right truncate from basechars . This
+ may give false hits due to accents not being used.
+ */
+ i = res_len;
+ while (--i >= 0 && res_buf[i] != '\x01')
+ ;
+ if (i > 0)
+ {
+ while (--i >= 0 && res_buf[i] != '\x01')
+ ;
+ }
+ if (i == 0)
+ { /* did not find base chars at all. Throw error */
+ return -1;
+ }
+ res_len = i; /* reduce res_len */
+ }
+ for (i = 0; i < res_len; i++)
+ {
+ if (strchr(REGEX_CHARS "\\", res_buf[i]))
+ wrbuf_putc(term_dict, '\\');
+ if (res_buf[i] < 32)
+ wrbuf_putc(term_dict, 1);
+
+ wrbuf_putc(term_dict, res_buf[i]);
+ }
+ if (right_trunc)
+ wrbuf_puts(term_dict, ".*");
+ return 1;
+}
+
/* term_100: handle term, where trunc = none(no operators at all) */
-static int term_100(ZebraMaps zebra_maps, int reg_type,
+static int term_100(zebra_map_t zm,
const char **src, WRBUF term_dict, int space_split,
- char *dst_term)
+ WRBUF display_term)
{
const char *s0;
const char **map;
int i = 0;
- int j = 0;
const char *space_start = 0;
const char *space_end = 0;
- if (!term_pre(zebra_maps, reg_type, src, NULL, NULL, !space_split))
+ if (!term_pre(zm, src, NULL, NULL, !space_split))
return 0;
s0 = *src;
while (*s0)
{
const char *s1 = s0;
int q_map_match = 0;
- map = zebra_maps_search(zebra_maps, reg_type, &s0, strlen(s0),
- &q_map_match);
+ map = zebra_maps_search(zm, &s0, strlen(s0), &q_map_match);
if (space_split)
{
if (**map == *CHR_SPACE)
{
if (strchr(REGEX_CHARS, *space_start))
wrbuf_putc(term_dict, '\\');
- dst_term[j++] = *space_start;
+ wrbuf_putc(display_term, *space_start);
wrbuf_putc(term_dict, *space_start);
space_start++;
}
i++;
- add_non_space(s1, s0, term_dict, dst_term, &j,
- map, q_map_match);
+ add_non_space(s1, s0, term_dict, display_term, map, q_map_match);
}
- dst_term[j] = '\0';
*src = s0;
return i;
}
/* term_101: handle term, where trunc = Process # */
-static int term_101(ZebraMaps zebra_maps, int reg_type,
+static int term_101(zebra_map_t zm,
const char **src, WRBUF term_dict, int space_split,
- char *dst_term)
+ WRBUF display_term)
{
const char *s0;
const char **map;
int i = 0;
- int j = 0;
- if (!term_pre(zebra_maps, reg_type, src, "#", "#", !space_split))
+ if (!term_pre(zm, src, "#", "#", !space_split))
return 0;
s0 = *src;
while (*s0)
{
i++;
wrbuf_puts(term_dict, ".*");
- dst_term[j++] = *s0++;
+ wrbuf_putc(display_term, *s0);
+ s0++;
}
else
{
const char *s1 = s0;
int q_map_match = 0;
- map = zebra_maps_search(zebra_maps, reg_type, &s0, strlen(s0),
- &q_map_match);
+ map = zebra_maps_search(zm, &s0, strlen(s0), &q_map_match);
if (space_split && **map == *CHR_SPACE)
break;
i++;
- add_non_space(s1, s0, term_dict, dst_term, &j,
- map, q_map_match);
+ add_non_space(s1, s0, term_dict, display_term, map, q_map_match);
}
}
- dst_term[j++] = '\0';
*src = s0;
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,
+static int term_103(zebra_map_t zm, const char **src,
WRBUF term_dict, int *errors, int space_split,
- char *dst_term)
+ WRBUF display_term)
{
int i = 0;
- int j = 0;
const char *s0;
const char **map;
- if (!term_pre(zebra_maps, reg_type, src, "^\\()[].*+?|", "(", !space_split))
+ if (!term_pre(zm, src, "^\\()[].*+?|", "(", !space_split))
return 0;
s0 = *src;
if (errors && *s0 == '+' && s0[1] && s0[2] == '+' && s0[3] &&
{
if (strchr("^\\()[].*+?|-", *s0))
{
- dst_term[j++] = *s0;
+ wrbuf_putc(display_term, *s0);
wrbuf_putc(term_dict, *s0);
s0++;
i++;
{
const char *s1 = s0;
int q_map_match = 0;
- map = zebra_maps_search(zebra_maps, reg_type, &s0, strlen(s0),
- &q_map_match);
+ map = zebra_maps_search(zm, &s0, strlen(s0), &q_map_match);
if (space_split && **map == *CHR_SPACE)
break;
i++;
- add_non_space(s1, s0, term_dict, dst_term, &j,
- map, q_map_match);
+ add_non_space(s1, s0, term_dict, display_term, map, q_map_match);
}
}
- dst_term[j] = '\0';
*src = s0;
return i;
}
/* term_103: handle term, where trunc = re-1 (regular expressions) */
-static int term_102(ZebraMaps zebra_maps, int reg_type, const char **src,
- WRBUF term_dict, int space_split, char *dst_term)
+static int term_102(zebra_map_t zm, const char **src,
+ WRBUF term_dict, int space_split, WRBUF display_term)
{
- return term_103(zebra_maps, reg_type, src, term_dict, NULL, space_split,
- dst_term);
+ return term_103(zm, src, term_dict, NULL, space_split, display_term);
}
/* term_104: handle term, process # and ! */
-static int term_104(ZebraMaps zebra_maps, int reg_type,
- const char **src, WRBUF term_dict, int space_split,
- char *dst_term)
+static int term_104(zebra_map_t zm, const char **src,
+ WRBUF term_dict, int space_split, WRBUF display_term)
{
const char *s0;
const char **map;
int i = 0;
- int j = 0;
- if (!term_pre(zebra_maps, reg_type, src, "?*#", "?*#", !space_split))
+ if (!term_pre(zm, src, "?*#", "?*#", !space_split))
return 0;
s0 = *src;
while (*s0)
if (*s0 == '?')
{
i++;
- dst_term[j++] = *s0++;
+ wrbuf_putc(display_term, *s0);
+ s0++;
if (*s0 >= '0' && *s0 <= '9')
{
int limit = 0;
while (*s0 >= '0' && *s0 <= '9')
{
limit = limit * 10 + (*s0 - '0');
- dst_term[j++] = *s0++;
+ wrbuf_putc(display_term, *s0);
+ s0++;
}
if (limit > 20)
limit = 20;
{
i++;
wrbuf_puts(term_dict, ".*");
- dst_term[j++] = *s0++;
+ wrbuf_putc(display_term, *s0);
+ s0++;
}
else if (*s0 == '#')
{
i++;
wrbuf_puts(term_dict, ".");
- dst_term[j++] = *s0++;
+ wrbuf_putc(display_term, *s0);
+ s0++;
}
else
{
const char *s1 = s0;
int q_map_match = 0;
- map = zebra_maps_search(zebra_maps, reg_type, &s0, strlen(s0),
- &q_map_match);
+ map = zebra_maps_search(zm, &s0, strlen(s0), &q_map_match);
if (space_split && **map == *CHR_SPACE)
break;
i++;
- add_non_space(s1, s0, term_dict, dst_term, &j,
- map, q_map_match);
+ add_non_space(s1, s0, term_dict, display_term, map, q_map_match);
}
}
- dst_term[j++] = '\0';
*src = s0;
return i;
}
/* 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, WRBUF term_dict, int space_split,
- char *dst_term, int right_truncate)
+static int term_105(zebra_map_t zm, const char **src,
+ WRBUF term_dict, int space_split,
+ WRBUF display_term, int right_truncate)
{
const char *s0;
const char **map;
int i = 0;
- int j = 0;
- if (!term_pre(zebra_maps, reg_type, src, "*!", "*!", !space_split))
+ if (!term_pre(zm, src, "*!", "*!", !space_split))
return 0;
s0 = *src;
while (*s0)
{
i++;
wrbuf_puts(term_dict, ".*");
- dst_term[j++] = *s0++;
+ wrbuf_putc(display_term, *s0);
+ s0++;
}
else if (*s0 == '!')
{
i++;
wrbuf_putc(term_dict, '.');
- dst_term[j++] = *s0++;
+ wrbuf_putc(display_term, *s0);
+ s0++;
}
else
{
const char *s1 = s0;
int q_map_match = 0;
- map = zebra_maps_search(zebra_maps, reg_type, &s0, strlen(s0),
- &q_map_match);
+ map = zebra_maps_search(zm, &s0, strlen(s0), &q_map_match);
if (space_split && **map == *CHR_SPACE)
break;
i++;
- add_non_space(s1, s0, term_dict, dst_term, &j,
- map, q_map_match);
+ add_non_space(s1, s0, term_dict, display_term, map, q_map_match);
}
}
if (right_truncate)
wrbuf_puts(term_dict, ".*");
- dst_term[j++] = '\0';
*src = s0;
return i;
}
static int string_relation(ZebraHandle zh, Z_AttributesPlusTerm *zapt,
const char **term_sub, WRBUF term_dict,
const Odr_oid *attributeSet,
- int reg_type, int space_split, char *term_dst,
+ zebra_map_t zm, int space_split,
+ WRBUF display_term,
int *error_code)
{
AttrType relation;
switch (relation_value)
{
case 1:
- if (!term_100(zh->reg->zebra_maps, reg_type,
- term_sub, term_component,
- space_split, term_dst))
+ if (!term_100(zm, term_sub, term_component, space_split, display_term))
{
wrbuf_destroy(term_component);
return 0;
wrbuf_putc(term_dict, ')');
break;
case 2:
- if (!term_100(zh->reg->zebra_maps, reg_type,
- term_sub, term_component,
- space_split, term_dst))
+ if (!term_100(zm, term_sub, term_component, space_split, display_term))
{
wrbuf_destroy(term_component);
return 0;
wrbuf_putc(term_dict, ')');
break;
case 5:
- if (!term_100 (zh->reg->zebra_maps, reg_type,
- term_sub, term_component, space_split, term_dst))
+ if (!term_100(zm, term_sub, term_component, space_split, display_term))
{
wrbuf_destroy(term_component);
return 0;
wrbuf_putc(term_dict, ')');
break;
case 4:
- if (!term_100(zh->reg->zebra_maps, reg_type, term_sub,
- term_component, space_split, term_dst))
+ if (!term_100(zm, term_sub, term_component, space_split, display_term))
{
wrbuf_destroy(term_component);
return 0;
if (!**term_sub)
return 1;
yaz_log(log_level_rpn, "Relation =");
- if (!term_100(zh->reg->zebra_maps, reg_type, term_sub,
- term_component, space_split, term_dst))
+ if (!term_100(zm, term_sub, term_component, space_split, display_term))
{
wrbuf_destroy(term_component);
return 0;
WRBUF term_dict,
const Odr_oid *attributeSet, NMEM stream,
struct grep_info *grep_info,
- int reg_type, int complete_flag,
- int num_bases, char **basenames,
- char *term_dst,
+ const char *index_type, int complete_flag,
+ WRBUF display_term,
const char *xpath_use,
- struct ord_list **ol);
-
-static ZEBRA_RES term_limits_APT(ZebraHandle zh,
- Z_AttributesPlusTerm *zapt,
- zint *hits_limit_value,
- const char **term_ref_id_str,
- NMEM nmem)
+ struct ord_list **ol,
+ zebra_map_t zm);
+
+ZEBRA_RES zebra_term_limits_APT(ZebraHandle zh,
+ Z_AttributesPlusTerm *zapt,
+ zint *hits_limit_value,
+ const char **term_ref_id_str,
+ NMEM nmem)
{
AttrType term_ref_id_attr;
AttrType hits_limit_attr;
int term_ref_id_int;
+ zint hits_limit_from_attr;
attr_init_APT(&hits_limit_attr, zapt, 11);
- *hits_limit_value = attr_find(&hits_limit_attr, NULL);
+ hits_limit_from_attr = attr_find(&hits_limit_attr, NULL);
attr_init_APT(&term_ref_id_attr, zapt, 10);
term_ref_id_int = attr_find_ex(&term_ref_id_attr, NULL, term_ref_id_str);
sprintf(res, "%d", term_ref_id_int);
*term_ref_id_str = res;
}
+ if (hits_limit_from_attr != -1)
+ *hits_limit_value = hits_limit_from_attr;
- /* no limit given ? */
- if (*hits_limit_value == -1)
- {
- if (*term_ref_id_str)
- {
- /* use global if term_ref is present */
- *hits_limit_value = zh->approx_limit;
- }
- else
- {
- /* no counting if term_ref is not present */
- *hits_limit_value = 0;
- }
- }
- else if (*hits_limit_value == 0)
- {
- /* 0 is the same as global limit */
- *hits_limit_value = zh->approx_limit;
- }
yaz_log(YLOG_DEBUG, "term_limits_APT ref_id=%s limit=" ZINT_FORMAT,
*term_ref_id_str ? *term_ref_id_str : "none",
*hits_limit_value);
return ZEBRA_OK;
}
-static ZEBRA_RES term_trunc(ZebraHandle zh,
- Z_AttributesPlusTerm *zapt,
- const char **term_sub,
- const Odr_oid *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,
- const char *xpath_use,
- NMEM rset_nmem,
- RSET *rset,
- struct rset_key_control *kc)
+/** \brief search for term (which may be truncated)
+ */
+static ZEBRA_RES search_term(ZebraHandle zh,
+ Z_AttributesPlusTerm *zapt,
+ const char **term_sub,
+ const Odr_oid *attributeSet,
+ zint hits_limit, NMEM stream,
+ struct grep_info *grep_info,
+ const char *index_type, int complete_flag,
+ const char *rank_type,
+ const char *xpath_use,
+ NMEM rset_nmem,
+ RSET *rset,
+ struct rset_key_control *kc,
+ zebra_map_t zm)
{
ZEBRA_RES res;
struct ord_list *ol;
- zint hits_limit_value;
+ zint hits_limit_value = hits_limit;
const char *term_ref_id_str = 0;
WRBUF term_dict = wrbuf_alloc();
-
+ WRBUF display_term = wrbuf_alloc();
*rset = 0;
- term_limits_APT(zh, zapt, &hits_limit_value, &term_ref_id_str, stream);
+ zebra_term_limits_APT(zh, zapt, &hits_limit_value, &term_ref_id_str,
+ stream);
grep_info->isam_p_indx = 0;
res = string_term(zh, zapt, term_sub, term_dict,
attributeSet, stream, grep_info,
- reg_type, complete_flag, num_bases, basenames,
- term_dst, xpath_use, &ol);
+ index_type, complete_flag,
+ display_term, xpath_use, &ol, zm);
wrbuf_destroy(term_dict);
- if (res != ZEBRA_OK)
- return res;
- if (!*term_sub) /* no more terms ? */
- return res;
- yaz_log(log_level_rpn, "term: %s", term_dst);
- *rset = 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,
- kc, kc->scope, ol, reg_type, hits_limit_value,
- term_ref_id_str);
- if (!*rset)
- return ZEBRA_FAIL;
- return ZEBRA_OK;
+ if (res == ZEBRA_OK && *term_sub)
+ {
+ yaz_log(log_level_rpn, "term: %s", wrbuf_cstr(display_term));
+ *rset = rset_trunc(zh, grep_info->isam_p_buf,
+ grep_info->isam_p_indx, wrbuf_buf(display_term),
+ wrbuf_len(display_term), rank_type,
+ 1 /* preserve pos */,
+ zapt->term->which, rset_nmem,
+ kc, kc->scope, ol, index_type, hits_limit_value,
+ term_ref_id_str);
+ if (!*rset)
+ res = ZEBRA_FAIL;
+ }
+ wrbuf_destroy(display_term);
+ return res;
}
static ZEBRA_RES string_term(ZebraHandle zh, Z_AttributesPlusTerm *zapt,
WRBUF term_dict,
const Odr_oid *attributeSet, NMEM stream,
struct grep_info *grep_info,
- int reg_type, int complete_flag,
- int num_bases, char **basenames,
- char *term_dst,
+ const char *index_type, int complete_flag,
+ WRBUF display_term,
const char *xpath_use,
- struct ord_list **ol)
+ struct ord_list **ol,
+ zebra_map_t zm)
{
- int r, base_no;
+ int r;
AttrType truncation;
int truncation_value;
const char *termp;
struct rpn_char_map_info rcmi;
- int space_split = complete_flag ? 0 : 1;
- int bases_ok = 0; /* no of databases with OK attribute */
+ int space_split = complete_flag ? 0 : 1;
+ int ord = -1;
+ int regex_range = 0;
+ int max_pos, prefix_len = 0;
+ int relation_error;
+ char ord_buf[32];
+ int ord_len, i;
*ol = ord_list_create(stream);
- rpn_char_map_prepare (zh->reg, reg_type, &rcmi);
+ rpn_char_map_prepare(zh->reg, zm, &rcmi);
attr_init_APT(&truncation, zapt, 5);
truncation_value = attr_find(&truncation, NULL);
yaz_log(log_level_rpn, "truncation value %d", truncation_value);
- for (base_no = 0; base_no < num_bases; base_no++)
+ termp = *term_sub; /* start of term for each database */
+
+ if (zebra_apt_get_ord(zh, zapt, index_type, xpath_use,
+ attributeSet, &ord) != ZEBRA_OK)
{
- int ord = -1;
- int regex_range = 0;
- int max_pos, prefix_len = 0;
- int relation_error;
- char ord_buf[32];
- int ord_len, i;
-
- termp = *term_sub; /* start of term for each database */
-
+ *term_sub = 0;
+ return ZEBRA_FAIL;
+ }
+
+ wrbuf_rewind(term_dict); /* new dictionary regexp term */
+
+ *ol = ord_list_append(stream, *ol, ord);
+ ord_len = key_SU_encode(ord, ord_buf);
+
+ wrbuf_putc(term_dict, '(');
+
+ for (i = 0; i<ord_len; i++)
+ {
+ wrbuf_putc(term_dict, 1); /* our internal regexp escape char */
+ wrbuf_putc(term_dict, ord_buf[i]);
+ }
+ wrbuf_putc(term_dict, ')');
+
+ prefix_len = wrbuf_len(term_dict);
- if (zebraExplain_curDatabase (zh->reg->zei, basenames[base_no]))
+ if (zebra_maps_is_icu(zm))
+ {
+ int relation_value;
+ AttrType relation;
+
+ attr_init_APT(&relation, zapt, 2);
+ relation_value = attr_find(&relation, NULL);
+ if (relation_value == 103) /* always matches */
+ termp += strlen(termp); /* move to end of term */
+ else if (relation_value == 3 || relation_value == 102 || relation_value == -1)
{
- zebra_setError(zh, YAZ_BIB1_DATABASE_UNAVAILABLE,
- basenames[base_no]);
- return ZEBRA_FAIL;
+ /* ICU case */
+ switch (truncation_value)
+ {
+ case -1: /* not specified */
+ case 100: /* do not truncate */
+ if (!term_100_icu(zm, &termp, term_dict, space_split, display_term, 0))
+ {
+ *term_sub = 0;
+ return ZEBRA_OK;
+ }
+ break;
+ case 1: /* right truncation */
+ if (!term_100_icu(zm, &termp, term_dict, space_split, display_term, 1))
+ {
+ *term_sub = 0;
+ return ZEBRA_OK;
+ }
+ break;
+ default:
+ zebra_setError_zint(zh,
+ YAZ_BIB1_UNSUPP_TRUNCATION_ATTRIBUTE,
+ truncation_value);
+ return ZEBRA_FAIL;
+ }
}
-
- if (zebra_apt_get_ord(zh, zapt, reg_type, xpath_use,
- attributeSet, &ord) != ZEBRA_OK)
- continue;
-
-
- wrbuf_rewind(term_dict); /* new dictionary regexp term */
-
- bases_ok++;
-
- *ol = ord_list_append(stream, *ol, ord);
- ord_len = key_SU_encode (ord, ord_buf);
-
- wrbuf_putc(term_dict, '(');
-
- for (i = 0; i<ord_len; i++)
+ else
{
- wrbuf_putc(term_dict, 1); /* our internal regexp escape char */
- wrbuf_putc(term_dict, ord_buf[i]);
+ zebra_setError_zint(zh,
+ YAZ_BIB1_UNSUPP_RELATION_ATTRIBUTE,
+ relation_value);
+ return ZEBRA_FAIL;
}
- wrbuf_putc(term_dict, ')');
-
- prefix_len = wrbuf_len(term_dict);
-
+ }
+ else
+ {
+ /* non-ICU case. using string.chr and friends */
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)
- {
- zebra_setError(zh, relation_error, 0);
- return ZEBRA_FAIL;
- }
- *term_sub = 0;
- return ZEBRA_OK;
- }
- break;
+ zm, space_split, display_term,
+ &relation_error))
+ {
+ if (relation_error)
+ {
+ zebra_setError(zh, relation_error, 0);
+ return ZEBRA_FAIL;
+ }
+ *term_sub = 0;
+ return ZEBRA_OK;
+ }
+ break;
case 1: /* right truncation */
wrbuf_putc(term_dict, '(');
- if (!term_100(zh->reg->zebra_maps, reg_type,
- &termp, term_dict, space_split, term_dst))
- {
- *term_sub = 0;
- return ZEBRA_OK;
- }
+ if (!term_100(zm, &termp, term_dict, space_split, display_term))
+ {
+ *term_sub = 0;
+ return ZEBRA_OK;
+ }
wrbuf_puts(term_dict, ".*)");
break;
- case 2: /* keft truncation */
+ case 2: /* left truncation */
wrbuf_puts(term_dict, "(.*");
- if (!term_100(zh->reg->zebra_maps, reg_type,
- &termp, term_dict, space_split, term_dst))
- {
- *term_sub = 0;
- return ZEBRA_OK;
- }
+ if (!term_100(zm, &termp, term_dict, space_split, display_term))
+ {
+ *term_sub = 0;
+ return ZEBRA_OK;
+ }
wrbuf_putc(term_dict, ')');
break;
case 3: /* left&right truncation */
wrbuf_puts(term_dict, "(.*");
- if (!term_100(zh->reg->zebra_maps, reg_type,
- &termp, term_dict, space_split, term_dst))
- {
- *term_sub = 0;
- return ZEBRA_OK;
- }
+ if (!term_100(zm, &termp, term_dict, space_split, display_term))
+ {
+ *term_sub = 0;
+ return ZEBRA_OK;
+ }
wrbuf_puts(term_dict, ".*)");
break;
case 101: /* process # in term */
wrbuf_putc(term_dict, '(');
- if (!term_101(zh->reg->zebra_maps, reg_type,
- &termp, term_dict, space_split, term_dst))
- {
- *term_sub = 0;
+ if (!term_101(zm, &termp, term_dict, space_split, display_term))
+ {
+ *term_sub = 0;
return ZEBRA_OK;
- }
+ }
wrbuf_puts(term_dict, ")");
break;
case 102: /* Regexp-1 */
wrbuf_putc(term_dict, '(');
- if (!term_102(zh->reg->zebra_maps, reg_type,
- &termp, term_dict, space_split, term_dst))
- {
- *term_sub = 0;
+ if (!term_102(zm, &termp, term_dict, space_split, display_term))
+ {
+ *term_sub = 0;
return ZEBRA_OK;
- }
+ }
wrbuf_putc(term_dict, ')');
break;
case 103: /* Regexp-2 */
regex_range = 1;
wrbuf_putc(term_dict, '(');
- if (!term_103(zh->reg->zebra_maps, reg_type,
- &termp, term_dict, ®ex_range,
- space_split, term_dst))
- {
- *term_sub = 0;
+ if (!term_103(zm, &termp, term_dict, ®ex_range,
+ space_split, display_term))
+ {
+ *term_sub = 0;
return ZEBRA_OK;
- }
+ }
wrbuf_putc(term_dict, ')');
- break;
+ break;
case 104: /* process # and ! in term */
wrbuf_putc(term_dict, '(');
- if (!term_104(zh->reg->zebra_maps, reg_type,
- &termp, term_dict, space_split, term_dst))
- {
- *term_sub = 0;
+ if (!term_104(zm, &termp, term_dict, space_split, display_term))
+ {
+ *term_sub = 0;
return ZEBRA_OK;
- }
+ }
wrbuf_putc(term_dict, ')');
break;
case 105: /* process * and ! in term */
wrbuf_putc(term_dict, '(');
- if (!term_105(zh->reg->zebra_maps, reg_type,
- &termp, term_dict, space_split, term_dst, 1))
- {
- *term_sub = 0;
+ if (!term_105(zm, &termp, term_dict, space_split, display_term, 1))
+ {
+ *term_sub = 0;
return ZEBRA_OK;
- }
+ }
wrbuf_putc(term_dict, ')');
break;
case 106: /* process * and ! in term */
wrbuf_putc(term_dict, '(');
- if (!term_105(zh->reg->zebra_maps, reg_type,
- &termp, term_dict, space_split, term_dst, 0))
- {
- *term_sub = 0;
+ if (!term_105(zm, &termp, term_dict, space_split, display_term, 0))
+ {
+ *term_sub = 0;
return ZEBRA_OK;
- }
+ }
wrbuf_putc(term_dict, ')');
break;
- default:
- zebra_setError_zint(zh,
- YAZ_BIB1_UNSUPP_TRUNCATION_ATTRIBUTE,
- truncation_value);
- return ZEBRA_FAIL;
+ default:
+ zebra_setError_zint(zh,
+ YAZ_BIB1_UNSUPP_TRUNCATION_ATTRIBUTE,
+ truncation_value);
+ return ZEBRA_FAIL;
}
- if (1)
- {
- char buf[1000];
- const char *input = wrbuf_cstr(term_dict) + prefix_len;
- esc_str(buf, sizeof(buf), input, strlen(input));
- }
- yaz_log(log_level_rpn, "dict_lookup_grep: %s",
- wrbuf_cstr(term_dict) + prefix_len);
- r = dict_lookup_grep(zh->reg->dict, wrbuf_cstr(term_dict), regex_range,
- grep_info, &max_pos,
- ord_len /* number of "exact" chars */,
- grep_handle);
- if (r == 1)
- zebra_set_partial_result(zh);
- else if (r)
- yaz_log(YLOG_WARN, "dict_lookup_grep fail %d", r);
- }
- if (!bases_ok)
- return ZEBRA_FAIL;
+ }
+ if (1)
+ {
+ char buf[1000];
+ const char *input = wrbuf_cstr(term_dict) + prefix_len;
+ esc_str(buf, sizeof(buf), input, strlen(input));
+ }
+ {
+ WRBUF pr_wr = wrbuf_alloc();
+
+ wrbuf_write_escaped(pr_wr, wrbuf_buf(term_dict), wrbuf_len(term_dict));
+ yaz_log(YLOG_LOG, "dict_lookup_grep: %s", wrbuf_cstr(pr_wr));
+ wrbuf_destroy(pr_wr);
+ }
+ r = dict_lookup_grep(zh->reg->dict, wrbuf_cstr(term_dict), regex_range,
+ grep_info, &max_pos,
+ ord_len /* number of "exact" chars */,
+ grep_handle);
+ if (r == 1)
+ zebra_set_partial_result(zh);
+ else if (r)
+ yaz_log(YLOG_WARN, "dict_lookup_grep fail %d", r);
*term_sub = termp;
yaz_log(YLOG_DEBUG, "%d positions", grep_info->isam_p_indx);
return ZEBRA_OK;
static ZEBRA_RES grep_info_prepare(ZebraHandle zh,
Z_AttributesPlusTerm *zapt,
struct grep_info *grep_info,
- int reg_type)
+ const char *index_type)
{
#ifdef TERM_COUNT
grep_info->term_no = 0;
grep_info->isam_p_size = 0;
grep_info->isam_p_buf = NULL;
grep_info->zh = zh;
- grep_info->reg_type = reg_type;
+ grep_info->index_type = index_type;
grep_info->termset = 0;
if (zapt)
{
termset_name = resname;
}
else
- termset_name = termset_value_string;
+ 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)
}
return ZEBRA_OK;
}
-
-/**
- \brief Create result set(s) for list of terms
- \param zh Zebra Handle
- \param zapt Attributes Plust Term (RPN leaf)
- \param termz term as used in query but converted to UTF-8
- \param attributeSet default attribute set
- \param stream memory for result
- \param reg_type register type ('w', 'p',..)
- \param complete_flag whether it's phrases or not
- \param rank_type term flags for ranking
- \param xpath_use use attribute for X-Path (-1 for no X-path)
- \param num_bases number of databases
- \param basenames array of databases
- \param rset_nmem memory for result sets
- \param result_sets output result set for each term in list (output)
- \param num_result_sets number of output result sets
- \param kc rset key control to be used for created result sets
-*/
-static ZEBRA_RES term_list_trunc(ZebraHandle zh,
- Z_AttributesPlusTerm *zapt,
- const char *termz,
- const Odr_oid *attributeSet,
- NMEM stream,
- int reg_type, int complete_flag,
- const char *rank_type,
- const char *xpath_use,
- int num_bases, char **basenames,
- NMEM rset_nmem,
- RSET **result_sets, int *num_result_sets,
- struct rset_key_control *kc)
+
+static ZEBRA_RES search_terms_chrmap(ZebraHandle zh,
+ Z_AttributesPlusTerm *zapt,
+ const char *termz,
+ const Odr_oid *attributeSet,
+ zint hits_limit,
+ NMEM stream,
+ const char *index_type, int complete_flag,
+ const char *rank_type,
+ const char *xpath_use,
+ NMEM rset_nmem,
+ RSET **result_sets, int *num_result_sets,
+ struct rset_key_control *kc,
+ zebra_map_t zm)
{
- char term_dst[IT_MAX_WORD+1];
struct grep_info grep_info;
const char *termp = termz;
int alloc_sets = 0;
-
+
*num_result_sets = 0;
- *term_dst = 0;
- if (grep_info_prepare(zh, zapt, &grep_info, reg_type) == ZEBRA_FAIL)
+ if (grep_info_prepare(zh, zapt, &grep_info, index_type) == ZEBRA_FAIL)
return ZEBRA_FAIL;
- while(1)
+ while (1)
{
ZEBRA_RES res;
alloc_sets = alloc_sets + add;
*result_sets = rnew;
}
- res = term_trunc(zh, zapt, &termp, attributeSet,
- stream, &grep_info,
- reg_type, complete_flag,
- num_bases, basenames,
- term_dst, rank_type,
- xpath_use, rset_nmem,
- &(*result_sets)[*num_result_sets],
- kc);
+ res = search_term(zh, zapt, &termp, attributeSet, hits_limit,
+ stream, &grep_info,
+ index_type, complete_flag,
+ rank_type,
+ xpath_use, rset_nmem,
+ &(*result_sets)[*num_result_sets],
+ kc, zm);
if (res != ZEBRA_OK)
{
int i;
for (i = 0; i < *num_result_sets; i++)
rset_delete((*result_sets)[i]);
- grep_info_delete (&grep_info);
+ grep_info_delete(&grep_info);
return res;
}
if ((*result_sets)[*num_result_sets] == 0)
grep_info_delete(&grep_info);
return ZEBRA_OK;
}
+
+/**
+ \brief Create result set(s) for list of terms
+ \param zh Zebra Handle
+ \param zapt Attributes Plust Term (RPN leaf)
+ \param termz term as used in query but converted to UTF-8
+ \param attributeSet default attribute set
+ \param stream memory for result
+ \param index_type register type ("w", "p",..)
+ \param complete_flag whether it's phrases or not
+ \param rank_type term flags for ranking
+ \param xpath_use use attribute for X-Path (-1 for no X-path)
+ \param rset_nmem memory for result sets
+ \param result_sets output result set for each term in list (output)
+ \param num_result_sets number of output result sets
+ \param kc rset key control to be used for created result sets
+*/
+static ZEBRA_RES search_terms_list(ZebraHandle zh,
+ Z_AttributesPlusTerm *zapt,
+ const char *termz,
+ const Odr_oid *attributeSet,
+ zint hits_limit,
+ NMEM stream,
+ const char *index_type, int complete_flag,
+ const char *rank_type,
+ const char *xpath_use,
+ NMEM rset_nmem,
+ RSET **result_sets, int *num_result_sets,
+ struct rset_key_control *kc)
+{
+ zebra_map_t zm = zebra_map_get_or_add(zh->reg->zebra_maps, index_type);
+ if (zebra_maps_is_icu(zm))
+ zebra_map_tokenize_start(zm, termz, strlen(termz));
+ return search_terms_chrmap(zh, zapt, termz, attributeSet, hits_limit,
+ stream, index_type, complete_flag,
+ rank_type, xpath_use,
+ rset_nmem, result_sets, num_result_sets,
+ kc, zm);
+}
-static ZEBRA_RES rpn_search_APT_position(ZebraHandle zh,
- Z_AttributesPlusTerm *zapt,
- const Odr_oid *attributeSet,
- int reg_type,
- int num_bases, char **basenames,
- NMEM rset_nmem,
- RSET *rset,
- struct rset_key_control *kc)
+
+/** \brief limit a search by position - returns result set
+ */
+static ZEBRA_RES search_position(ZebraHandle zh,
+ Z_AttributesPlusTerm *zapt,
+ const Odr_oid *attributeSet,
+ const char *index_type,
+ NMEM rset_nmem,
+ RSET *rset,
+ struct rset_key_control *kc)
{
- RSET *f_set;
- int base_no;
int position_value;
- int num_sets = 0;
AttrType position;
-
+ int ord = -1;
+ char ord_buf[32];
+ char term_dict[100];
+ int ord_len;
+ char *val;
+ ISAM_P isam_p;
+ zebra_map_t zm = zebra_map_get_or_add(zh->reg->zebra_maps, index_type);
+
attr_init_APT(&position, zapt, 3);
position_value = attr_find(&position, NULL);
switch(position_value)
return ZEBRA_FAIL;
}
- if (!zebra_maps_is_first_in_field(zh->reg->zebra_maps, reg_type))
+
+ if (!zebra_maps_is_first_in_field(zm))
{
zebra_setError_zint(zh, YAZ_BIB1_UNSUPP_POSITION_ATTRIBUTE,
position_value);
return ZEBRA_FAIL;
}
- if (!zh->reg->isamb && !zh->reg->isamc)
+ if (zebra_apt_get_ord(zh, zapt, index_type, 0,
+ attributeSet, &ord) != ZEBRA_OK)
{
- zebra_setError_zint(zh, YAZ_BIB1_UNSUPP_POSITION_ATTRIBUTE,
- position_value);
return ZEBRA_FAIL;
}
- f_set = xmalloc(sizeof(RSET) * num_bases);
- for (base_no = 0; base_no < num_bases; base_no++)
+ ord_len = key_SU_encode(ord, ord_buf);
+ memcpy(term_dict, ord_buf, ord_len);
+ strcpy(term_dict+ord_len, FIRST_IN_FIELD_STR);
+ val = dict_lookup(zh->reg->dict, term_dict);
+ if (val)
{
- int ord = -1;
- char ord_buf[32];
- char term_dict[100];
- int ord_len;
- char *val;
- ISAM_P isam_p;
-
- if (zebraExplain_curDatabase (zh->reg->zei, basenames[base_no]))
- {
- zebra_setError(zh, YAZ_BIB1_DATABASE_UNAVAILABLE,
- basenames[base_no]);
- return ZEBRA_FAIL;
- }
-
- if (zebra_apt_get_ord(zh, zapt, reg_type, 0,
- attributeSet, &ord) != ZEBRA_OK)
- continue;
-
- ord_len = key_SU_encode (ord, ord_buf);
- memcpy(term_dict, ord_buf, ord_len);
- strcpy(term_dict+ord_len, FIRST_IN_FIELD_STR);
- val = dict_lookup(zh->reg->dict, term_dict);
- if (!val)
- continue;
assert(*val == sizeof(ISAM_P));
memcpy(&isam_p, val+1, sizeof(isam_p));
-
- if (zh->reg->isamb)
- f_set[num_sets++] = rsisamb_create(rset_nmem, kc, kc->scope,
- zh->reg->isamb, isam_p, 0);
- else if (zh->reg->isamc)
- f_set[num_sets++] = rsisamc_create(rset_nmem, kc, kc->scope,
- zh->reg->isamc, isam_p, 0);
+ *rset = zebra_create_rset_isam(zh, rset_nmem, kc, kc->scope,
+ isam_p, 0);
}
- if (num_sets)
- {
- *rset = rset_create_or(rset_nmem, kc, kc->scope,
- 0 /* termid */, num_sets, f_set);
- }
- xfree(f_set);
return ZEBRA_OK;
}
-
+
+/** \brief returns result set for phrase search
+ */
static ZEBRA_RES rpn_search_APT_phrase(ZebraHandle zh,
Z_AttributesPlusTerm *zapt,
const char *termz_org,
const Odr_oid *attributeSet,
+ zint hits_limit,
NMEM stream,
- int reg_type, int complete_flag,
+ const char *index_type,
+ int complete_flag,
const char *rank_type,
const char *xpath_use,
- int num_bases, char **basenames,
NMEM rset_nmem,
RSET *rset,
struct rset_key_control *kc)
RSET *result_sets = 0;
int num_result_sets = 0;
ZEBRA_RES res =
- term_list_trunc(zh, zapt, termz_org, attributeSet,
- stream, reg_type, complete_flag,
- rank_type, xpath_use,
- num_bases, basenames,
- rset_nmem,
- &result_sets, &num_result_sets, kc);
-
+ search_terms_list(zh, zapt, termz_org, attributeSet, hits_limit,
+ stream, index_type, complete_flag,
+ rank_type, xpath_use,
+ rset_nmem,
+ &result_sets, &num_result_sets, kc);
+
if (res != ZEBRA_OK)
return res;
if (num_result_sets > 0)
{
RSET first_set = 0;
- res = rpn_search_APT_position(zh, zapt, attributeSet,
- reg_type,
- num_bases, basenames,
- rset_nmem, &first_set,
- kc);
+ res = search_position(zh, zapt, attributeSet,
+ index_type,
+ rset_nmem, &first_set,
+ kc);
if (res != ZEBRA_OK)
+ {
+ int i;
+ for (i = 0; i<num_result_sets; i++)
+ rset_delete(result_sets[i]);
return res;
+ }
if (first_set)
{
RSET *nsets = nmem_malloc(stream,
return ZEBRA_OK;
}
+/** \brief returns result set for or-list search
+ */
static ZEBRA_RES rpn_search_APT_or_list(ZebraHandle zh,
Z_AttributesPlusTerm *zapt,
const char *termz_org,
const Odr_oid *attributeSet,
+ zint hits_limit,
NMEM stream,
- int reg_type, int complete_flag,
+ const char *index_type,
+ int complete_flag,
const char *rank_type,
const char *xpath_use,
- int num_bases, char **basenames,
NMEM rset_nmem,
RSET *rset,
struct rset_key_control *kc)
int num_result_sets = 0;
int i;
ZEBRA_RES res =
- term_list_trunc(zh, zapt, termz_org, attributeSet,
- stream, reg_type, complete_flag,
- rank_type, xpath_use,
- num_bases, basenames,
- rset_nmem,
- &result_sets, &num_result_sets, kc);
+ search_terms_list(zh, zapt, termz_org, attributeSet, hits_limit,
+ stream, index_type, complete_flag,
+ rank_type, xpath_use,
+ rset_nmem,
+ &result_sets, &num_result_sets, kc);
if (res != ZEBRA_OK)
return res;
for (i = 0; i<num_result_sets; i++)
{
RSET first_set = 0;
- res = rpn_search_APT_position(zh, zapt, attributeSet,
- reg_type,
- num_bases, basenames,
- rset_nmem, &first_set,
- kc);
+ res = search_position(zh, zapt, attributeSet,
+ index_type,
+ rset_nmem, &first_set,
+ kc);
if (res != ZEBRA_OK)
{
for (i = 0; i<num_result_sets; i++)
return ZEBRA_OK;
}
+/** \brief returns result set for and-list search
+ */
static ZEBRA_RES rpn_search_APT_and_list(ZebraHandle zh,
Z_AttributesPlusTerm *zapt,
const char *termz_org,
const Odr_oid *attributeSet,
+ zint hits_limit,
NMEM stream,
- int reg_type, int complete_flag,
+ const char *index_type,
+ int complete_flag,
const char *rank_type,
const char *xpath_use,
- int num_bases, char **basenames,
NMEM rset_nmem,
RSET *rset,
struct rset_key_control *kc)
int num_result_sets = 0;
int i;
ZEBRA_RES res =
- term_list_trunc(zh, zapt, termz_org, attributeSet,
- stream, reg_type, complete_flag,
- rank_type, xpath_use,
- num_bases, basenames,
- rset_nmem,
- &result_sets, &num_result_sets,
- kc);
+ search_terms_list(zh, zapt, termz_org, attributeSet, hits_limit,
+ stream, index_type, complete_flag,
+ rank_type, xpath_use,
+ rset_nmem,
+ &result_sets, &num_result_sets,
+ kc);
if (res != ZEBRA_OK)
return res;
for (i = 0; i<num_result_sets; i++)
{
RSET first_set = 0;
- res = rpn_search_APT_position(zh, zapt, attributeSet,
- reg_type,
- num_bases, basenames,
- rset_nmem, &first_set,
- kc);
+ res = search_position(zh, zapt, attributeSet,
+ index_type,
+ rset_nmem, &first_set,
+ kc);
if (res != ZEBRA_OK)
{
for (i = 0; i<num_result_sets; i++)
const Odr_oid *attributeSet,
struct grep_info *grep_info,
int *max_pos,
- int reg_type,
- char *term_dst,
+ zebra_map_t zm,
+ WRBUF display_term,
int *error_code)
{
AttrType relation;
{
case 1:
yaz_log(log_level_rpn, "Relation <");
- if (!term_100(zh->reg->zebra_maps, reg_type, term_sub, term_num, 1,
- term_dst))
+ if (!term_100(zm, term_sub, term_num, 1, display_term))
{
wrbuf_destroy(term_num);
return 0;
}
- term_value = atoi (wrbuf_cstr(term_num));
+ term_value = atoi(wrbuf_cstr(term_num));
gen_regular_rel(term_dict, term_value-1, 1);
break;
case 2:
yaz_log(log_level_rpn, "Relation <=");
- if (!term_100(zh->reg->zebra_maps, reg_type, term_sub, term_num, 1,
- term_dst))
+ if (!term_100(zm, term_sub, term_num, 1, display_term))
{
wrbuf_destroy(term_num);
return 0;
}
- term_value = atoi (wrbuf_cstr(term_num));
+ term_value = atoi(wrbuf_cstr(term_num));
gen_regular_rel(term_dict, term_value, 1);
break;
case 4:
yaz_log(log_level_rpn, "Relation >=");
- if (!term_100(zh->reg->zebra_maps, reg_type, term_sub, term_num, 1,
- term_dst))
+ if (!term_100(zm, term_sub, term_num, 1, display_term))
{
wrbuf_destroy(term_num);
return 0;
}
- term_value = atoi (wrbuf_cstr(term_num));
+ term_value = atoi(wrbuf_cstr(term_num));
gen_regular_rel(term_dict, term_value, 0);
break;
case 5:
yaz_log(log_level_rpn, "Relation >");
- if (!term_100(zh->reg->zebra_maps, reg_type, term_sub, term_num, 1,
- term_dst))
+ if (!term_100(zm, term_sub, term_num, 1, display_term))
{
wrbuf_destroy(term_num);
return 0;
}
- term_value = atoi (wrbuf_cstr(term_num));
+ term_value = atoi(wrbuf_cstr(term_num));
gen_regular_rel(term_dict, term_value+1, 0);
break;
case -1:
case 3:
yaz_log(log_level_rpn, "Relation =");
- if (!term_100(zh->reg->zebra_maps, reg_type, term_sub, term_num, 1,
- term_dst))
+ if (!term_100(zm, term_sub, term_num, 1, display_term))
{
wrbuf_destroy(term_num);
return 0;
}
- term_value = atoi (wrbuf_cstr(term_num));
+ term_value = atoi(wrbuf_cstr(term_num));
wrbuf_printf(term_dict, "(0*%d)", term_value);
break;
case 103:
WRBUF term_dict,
const Odr_oid *attributeSet, NMEM stream,
struct grep_info *grep_info,
- int reg_type, int complete_flag,
- int num_bases, char **basenames,
- char *term_dst,
+ const char *index_type, int complete_flag,
+ WRBUF display_term,
const char *xpath_use,
struct ord_list **ol)
{
- int base_no;
const char *termp;
struct rpn_char_map_info rcmi;
-
- int bases_ok = 0; /* no of databases with OK attribute */
-
+ int max_pos;
+ int relation_error = 0;
+ int ord, ord_len, i;
+ char ord_buf[32];
+ zebra_map_t zm = zebra_map_get_or_add(zh->reg->zebra_maps, index_type);
+
*ol = ord_list_create(stream);
- rpn_char_map_prepare (zh->reg, reg_type, &rcmi);
+ rpn_char_map_prepare(zh->reg, zm, &rcmi);
- for (base_no = 0; base_no < num_bases; base_no++)
+ termp = *term_sub;
+
+ if (zebra_apt_get_ord(zh, zapt, index_type, xpath_use,
+ attributeSet, &ord) != ZEBRA_OK)
{
- int max_pos;
- int relation_error = 0;
- int ord, ord_len, i;
- char ord_buf[32];
-
- termp = *term_sub;
-
- if (zebraExplain_curDatabase (zh->reg->zei, basenames[base_no]))
+ return ZEBRA_FAIL;
+ }
+
+ wrbuf_rewind(term_dict);
+
+ *ol = ord_list_append(stream, *ol, ord);
+
+ ord_len = key_SU_encode(ord, ord_buf);
+
+ wrbuf_putc(term_dict, '(');
+ for (i = 0; i < ord_len; i++)
+ {
+ wrbuf_putc(term_dict, 1);
+ wrbuf_putc(term_dict, ord_buf[i]);
+ }
+ wrbuf_putc(term_dict, ')');
+
+ if (!numeric_relation(zh, zapt, &termp, term_dict,
+ attributeSet, grep_info, &max_pos, zm,
+ display_term, &relation_error))
+ {
+ if (relation_error)
{
- zebra_setError(zh, YAZ_BIB1_DATABASE_UNAVAILABLE,
- basenames[base_no]);
+ zebra_setError(zh, relation_error, 0);
return ZEBRA_FAIL;
}
-
- if (zebra_apt_get_ord(zh, zapt, reg_type, xpath_use,
- attributeSet, &ord) != ZEBRA_OK)
- continue;
- bases_ok++;
-
- wrbuf_rewind(term_dict);
-
- *ol = ord_list_append(stream, *ol, ord);
-
- ord_len = key_SU_encode (ord, ord_buf);
-
- wrbuf_putc(term_dict, '(');
- for (i = 0; i < ord_len; i++)
- {
- wrbuf_putc(term_dict, 1);
- wrbuf_putc(term_dict, ord_buf[i]);
- }
- wrbuf_putc(term_dict, ')');
-
- if (!numeric_relation(zh, zapt, &termp, term_dict,
- attributeSet, grep_info, &max_pos, reg_type,
- term_dst, &relation_error))
- {
- if (relation_error)
- {
- zebra_setError(zh, relation_error, 0);
- return ZEBRA_FAIL;
- }
- *term_sub = 0;
- return ZEBRA_OK;
- }
+ *term_sub = 0;
+ return ZEBRA_OK;
}
- if (!bases_ok)
- return ZEBRA_FAIL;
*term_sub = termp;
yaz_log(YLOG_DEBUG, "%d positions", grep_info->isam_p_indx);
return ZEBRA_OK;
const char *termz,
const Odr_oid *attributeSet,
NMEM stream,
- int reg_type, int complete_flag,
+ const char *index_type,
+ int complete_flag,
const char *rank_type,
const char *xpath_use,
- int num_bases, char **basenames,
NMEM rset_nmem,
RSET *rset,
struct rset_key_control *kc)
{
- char term_dst[IT_MAX_WORD+1];
const char *termp = termz;
RSET *result_sets = 0;
int num_result_sets = 0;
zint hits_limit_value;
const char *term_ref_id_str = 0;
- term_limits_APT(zh, zapt, &hits_limit_value, &term_ref_id_str, stream);
+ zebra_term_limits_APT(zh, zapt, &hits_limit_value, &term_ref_id_str,
+ stream);
yaz_log(log_level_rpn, "APT_numeric t='%s'", termz);
- if (grep_info_prepare(zh, zapt, &grep_info, reg_type) == ZEBRA_FAIL)
+ if (grep_info_prepare(zh, zapt, &grep_info, index_type) == ZEBRA_FAIL)
return ZEBRA_FAIL;
while (1)
{
struct ord_list *ol;
WRBUF term_dict = wrbuf_alloc();
+ WRBUF display_term = wrbuf_alloc();
if (alloc_sets == num_result_sets)
{
int add = 10;
grep_info.isam_p_indx = 0;
res = numeric_term(zh, zapt, &termp, term_dict,
attributeSet, stream, &grep_info,
- reg_type, complete_flag, num_bases, basenames,
- term_dst, xpath_use, &ol);
+ index_type, complete_flag,
+ display_term, xpath_use, &ol);
wrbuf_destroy(term_dict);
if (res == ZEBRA_FAIL || termp == 0)
+ {
+ wrbuf_destroy(display_term);
break;
- yaz_log(YLOG_DEBUG, "term: %s", term_dst);
+ }
+ yaz_log(YLOG_DEBUG, "term: %s", wrbuf_cstr(display_term));
result_sets[num_result_sets] =
rset_trunc(zh, grep_info.isam_p_buf,
- grep_info.isam_p_indx, term_dst,
- strlen(term_dst), rank_type,
+ grep_info.isam_p_indx, wrbuf_buf(display_term),
+ wrbuf_len(display_term), rank_type,
0 /* preserve position */,
zapt->term->which, rset_nmem,
- kc, kc->scope, ol, reg_type,
+ kc, kc->scope, ol, index_type,
hits_limit_value,
term_ref_id_str);
+ wrbuf_destroy(display_term);
if (!result_sets[num_result_sets])
break;
num_result_sets++;
sort_sequence->num_specs = 10;
sort_sequence->specs = (Z_SortKeySpec **)
nmem_malloc(stream, sort_sequence->num_specs *
- sizeof(*sort_sequence->specs));
+ sizeof(*sort_sequence->specs));
for (i = 0; i<sort_sequence->num_specs; i++)
sort_sequence->specs[i] = 0;
}
if (zapt->term->which != Z_Term_general)
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;
sprintf(termz, "%d", i);
sk->u.sortAttributes->id = odr_oiddup_nmem(stream, attributeSet);
sk->u.sortAttributes->list = zapt->attributes;
- sks->sortRelation = (int *)
+ sks->sortRelation = (Odr_int *)
nmem_malloc(stream, sizeof(*sks->sortRelation));
if (sort_relation_value == 1)
*sks->sortRelation = Z_SortKeySpec_ascending;
else
*sks->sortRelation = Z_SortKeySpec_ascending;
- sks->caseSensitivity = (int *)
+ sks->caseSensitivity = (Odr_int *)
nmem_malloc(stream, sizeof(*sks->caseSensitivity));
*sks->caseSensitivity = 0;
static RSET xpath_trunc(ZebraHandle zh, NMEM stream,
- int reg_type, const char *term,
+ const char *index_type, const char *term,
const char *xpath_use,
NMEM rset_nmem,
struct rset_key_control *kc)
struct grep_info grep_info;
int ord = zebraExplain_lookup_attr_str(zh->reg->zei,
zinfo_index_category_index,
- reg_type, xpath_use);
- if (grep_info_prepare(zh, 0 /* zapt */, &grep_info, '0') == ZEBRA_FAIL)
+ index_type, xpath_use);
+ if (grep_info_prepare(zh, 0 /* zapt */, &grep_info, "0") == ZEBRA_FAIL)
return rset_create_null(rset_nmem, kc, 0);
if (ord < 0)
char ord_buf[32];
RSET rset;
WRBUF term_dict = wrbuf_alloc();
- int ord_len = key_SU_encode (ord, ord_buf);
+ int ord_len = key_SU_encode(ord, ord_buf);
int term_type = Z_Term_characterString;
const char *flags = "void";
rset = rset_trunc(zh, grep_info.isam_p_buf,
grep_info.isam_p_indx, term, strlen(term),
flags, 1, term_type, rset_nmem,
- kc, kc->scope, 0, reg_type, 0 /* hits_limit */,
+ kc, kc->scope, 0, index_type, 0 /* hits_limit */,
0 /* term_ref_id_str */);
grep_info_delete(&grep_info);
wrbuf_destroy(term_dict);
static
ZEBRA_RES rpn_search_xpath(ZebraHandle zh,
- int num_bases, char **basenames,
NMEM stream, const char *rank_type, RSET rset,
int xpath_len, struct xpath_location_step *xpath,
NMEM rset_nmem,
RSET *rset_out,
struct rset_key_control *kc)
{
- int base_no;
int i;
int always_matches = rset ? 0 : 1;
}
/*
- //a -> a/.*
- //a/b -> b/a/.*
- /a -> a/
- /a/b -> b/a/
+ //a -> a/.*
+ //a/b -> b/a/.*
+ /a -> a/
+ /a/b -> b/a/
- / -> none
+ / -> none
- a[@attr = value]/b[@other = othervalue]
+ 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)
+ /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);
-
- for (base_no = 0; base_no < num_bases; base_no++)
+ dict_grep_cmap(zh->reg->dict, 0, 0);
+
{
int level = xpath_len;
int first_path = 1;
- if (zebraExplain_curDatabase (zh->reg->zei, basenames[base_no]))
- {
- zebra_setError(zh, YAZ_BIB1_DATABASE_UNAVAILABLE,
- basenames[base_no]);
- *rset_out = rset;
- return ZEBRA_FAIL;
- }
while (--level >= 0)
{
WRBUF xpath_rev = wrbuf_alloc();
}
}
rset_attr = xpath_trunc(
- zh, stream, '0', wrbuf_cstr(wbuf), ZEBRA_XPATH_ATTR_NAME,
+ zh, stream, "0", wrbuf_cstr(wbuf),
+ ZEBRA_XPATH_ATTR_NAME,
rset_nmem, kc);
wrbuf_destroy(wbuf);
}
wrbuf_cstr(xpath_rev));
if (wrbuf_len(xpath_rev))
{
- rset_start_tag = xpath_trunc(zh, stream, '0',
+ rset_start_tag = xpath_trunc(zh, stream, "0",
wrbuf_cstr(xpath_rev),
ZEBRA_XPATH_ELM_BEGIN,
rset_nmem, kc);
rset = rset_start_tag;
else
{
- rset_end_tag = xpath_trunc(zh, stream, '0',
+ rset_end_tag = xpath_trunc(zh, stream, "0",
wrbuf_cstr(xpath_rev),
ZEBRA_XPATH_ELM_END,
rset_nmem, kc);
#define MAX_XPATH_STEPS 10
+static ZEBRA_RES rpn_search_database(ZebraHandle zh,
+ Z_AttributesPlusTerm *zapt,
+ const Odr_oid *attributeSet,
+ zint hits_limit, NMEM stream,
+ Z_SortKeySpecList *sort_sequence,
+ NMEM rset_nmem,
+ RSET *rset,
+ struct rset_key_control *kc);
+
static ZEBRA_RES rpn_search_APT(ZebraHandle zh, Z_AttributesPlusTerm *zapt,
- const Odr_oid *attributeSet, NMEM stream,
+ const Odr_oid *attributeSet,
+ zint hits_limit, NMEM stream,
Z_SortKeySpecList *sort_sequence,
- int num_bases, char **basenames,
+ int num_bases, const char **basenames,
NMEM rset_nmem,
RSET *rset,
struct rset_key_control *kc)
{
+ RSET *rsets = nmem_malloc(stream, num_bases * sizeof(*rsets));
+ ZEBRA_RES res = ZEBRA_OK;
+ int i;
+ for (i = 0; i < num_bases; i++)
+ {
+
+ if (zebraExplain_curDatabase(zh->reg->zei, basenames[i]))
+ {
+ zebra_setError(zh, YAZ_BIB1_DATABASE_UNAVAILABLE,
+ basenames[i]);
+ res = ZEBRA_FAIL;
+ break;
+ }
+ res = rpn_search_database(zh, zapt, attributeSet, hits_limit, stream,
+ sort_sequence,
+ rset_nmem, rsets+i, kc);
+ if (res != ZEBRA_OK)
+ break;
+ }
+ if (res != ZEBRA_OK)
+ { /* must clean up the already created sets */
+ while (--i >= 0)
+ rset_delete(rsets[i]);
+ *rset = 0;
+ }
+ else
+ {
+ if (num_bases == 1)
+ *rset = rsets[0];
+ else if (num_bases == 0)
+ *rset = rset_create_null(rset_nmem, kc, 0);
+ else
+ *rset = rset_create_or(rset_nmem, kc, kc->scope, 0 /* TERMID */,
+ num_bases, rsets);
+ }
+ return res;
+}
+
+static ZEBRA_RES rpn_search_database(ZebraHandle zh,
+ Z_AttributesPlusTerm *zapt,
+ const Odr_oid *attributeSet,
+ zint hits_limit, NMEM stream,
+ Z_SortKeySpecList *sort_sequence,
+ NMEM rset_nmem,
+ RSET *rset,
+ struct rset_key_control *kc)
+{
ZEBRA_RES res = ZEBRA_OK;
- unsigned reg_id;
+ const char *index_type;
char *search_type = NULL;
char rank_type[128];
int complete_flag;
log_level_rpn = yaz_log_module_level("rpn");
log_level_set = 1;
}
- zebra_maps_attr(zh->reg->zebra_maps, zapt, ®_id, &search_type,
+ zebra_maps_attr(zh->reg->zebra_maps, zapt, &index_type, &search_type,
rank_type, &complete_flag, &sort_flag);
- yaz_log(YLOG_DEBUG, "reg_id=%c", reg_id);
+ yaz_log(YLOG_DEBUG, "index_type=%s", index_type);
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 (relation_value == 103) /* alwaysmatches */
{
*rset = 0; /* signal no "term" set */
- return rpn_search_xpath(zh, num_bases, basenames,
- stream, rank_type, *rset,
+ return rpn_search_xpath(zh, stream, rank_type, *rset,
xpath_len, xpath, rset_nmem, rset, kc);
}
}
*/
if (!strcmp(search_type, "phrase"))
{
- res = rpn_search_APT_phrase(zh, zapt, termz, attributeSet, stream,
- reg_id, complete_flag, rank_type,
+ res = rpn_search_APT_phrase(zh, zapt, termz, attributeSet, hits_limit,
+ stream,
+ index_type, complete_flag, rank_type,
xpath_use,
- num_bases, basenames, rset_nmem,
+ rset_nmem,
rset, kc);
}
else if (!strcmp(search_type, "and-list"))
{
- res = rpn_search_APT_and_list(zh, zapt, termz, attributeSet, stream,
- reg_id, complete_flag, rank_type,
+ res = rpn_search_APT_and_list(zh, zapt, termz, attributeSet, hits_limit,
+ stream,
+ index_type, complete_flag, rank_type,
xpath_use,
- num_bases, basenames, rset_nmem,
+ rset_nmem,
rset, kc);
}
else if (!strcmp(search_type, "or-list"))
{
- res = rpn_search_APT_or_list(zh, zapt, termz, attributeSet, stream,
- reg_id, complete_flag, rank_type,
+ res = rpn_search_APT_or_list(zh, zapt, termz, attributeSet, hits_limit,
+ stream,
+ index_type, complete_flag, rank_type,
xpath_use,
- num_bases, basenames, rset_nmem,
+ rset_nmem,
rset, kc);
}
else if (!strcmp(search_type, "local"))
else if (!strcmp(search_type, "numeric"))
{
res = rpn_search_APT_numeric(zh, zapt, termz, attributeSet, stream,
- reg_id, complete_flag, rank_type,
+ index_type, complete_flag, rank_type,
xpath_use,
- num_bases, basenames, rset_nmem,
+ rset_nmem,
rset, kc);
}
else
return res;
if (!*rset)
return ZEBRA_FAIL;
- return rpn_search_xpath(zh, num_bases, basenames,
- stream, rank_type, *rset,
+ return rpn_search_xpath(zh, stream, rank_type, *rset,
xpath_len, xpath, rset_nmem, rset, kc);
}
static ZEBRA_RES rpn_search_structure(ZebraHandle zh, Z_RPNStructure *zs,
- const Odr_oid *attributeSet,
+ const Odr_oid *attributeSet,
+ zint hits_limit,
NMEM stream, NMEM rset_nmem,
Z_SortKeySpecList *sort_sequence,
- int num_bases, char **basenames,
+ int num_bases, const char **basenames,
RSET **result_sets, int *num_result_sets,
Z_Operator *parent_op,
struct rset_key_control *kc);
ZEBRA_RES rpn_search_top(ZebraHandle zh, Z_RPNStructure *zs,
const Odr_oid *attributeSet,
+ zint hits_limit,
NMEM stream, NMEM rset_nmem,
Z_SortKeySpecList *sort_sequence,
- int num_bases, char **basenames,
+ int num_bases, const char **basenames,
RSET *result_set)
{
RSET *result_sets = 0;
ZEBRA_RES res;
struct rset_key_control *kc = zebra_key_control_create(zh);
- res = rpn_search_structure(zh, zs, attributeSet,
+ res = rpn_search_structure(zh, zs, attributeSet, hits_limit,
stream, rset_nmem,
sort_sequence,
num_bases, basenames,
}
ZEBRA_RES rpn_search_structure(ZebraHandle zh, Z_RPNStructure *zs,
- const Odr_oid *attributeSet,
+ const Odr_oid *attributeSet, zint hits_limit,
NMEM stream, NMEM rset_nmem,
Z_SortKeySpecList *sort_sequence,
- int num_bases, char **basenames,
+ int num_bases, const char **basenames,
RSET **result_sets, int *num_result_sets,
Z_Operator *parent_op,
struct rset_key_control *kc)
int num_result_sets_r = 0;
res = rpn_search_structure(zh, zs->u.complex->s1,
- attributeSet, stream, rset_nmem,
+ attributeSet, hits_limit, stream, rset_nmem,
sort_sequence,
num_bases, basenames,
&result_sets_l, &num_result_sets_l,
return res;
}
res = rpn_search_structure(zh, zs->u.complex->s2,
- attributeSet, stream, rset_nmem,
+ attributeSet, hits_limit, stream, rset_nmem,
sort_sequence,
num_bases, basenames,
&result_sets_r, &num_result_sets_r,
{
yaz_log(YLOG_DEBUG, "rpn_search_APT");
res = rpn_search_APT(zh, zs->u.simple->u.attributesPlusTerm,
- attributeSet, stream, sort_sequence,
+ attributeSet, hits_limit,
+ stream, sort_sequence,
num_bases, basenames, rset_nmem, &rset,
kc);
if (res != ZEBRA_OK)
/*
* Local variables:
* c-basic-offset: 4
+ * c-file-style: "Stroustrup"
* indent-tabs-mode: nil
* End:
* vim: shiftwidth=4 tabstop=8 expandtab