* Sebastian Hammer, Adam Dickmeiss
*
* $Log: zrpn.c,v $
- * Revision 1.47 1996-05-15 18:36:28 adam
+ * Revision 1.53 1996-06-26 09:21:43 adam
+ * Bug fix: local attribute set wasn't obeyed in scan.
+ *
+ * Revision 1.52 1996/06/17 14:26:20 adam
+ * Function gen_regular_rel changed to handle negative numbers.
+ *
+ * Revision 1.51 1996/06/11 10:54:15 quinn
+ * Relevance work
+ *
+ * Revision 1.50 1996/06/07 08:51:53 adam
+ * Bug fix: Character mapping was broken (introducued by last revision).
+ *
+ * Revision 1.49 1996/06/04 10:18:11 adam
+ * Search/scan uses character mapping module.
+ *
+ * Revision 1.48 1996/05/28 15:15:01 adam
+ * Bug fix: Didn't handle unknown database correctly.
+ *
+ * Revision 1.47 1996/05/15 18:36:28 adam
* Function trans_term transforms unsearchable characters to blanks.
*
* Revision 1.46 1996/05/15 11:57:56 adam
#include "zserver.h"
#include "attribute.h"
+#include <charmap.h>
#include <rsisam.h>
#include <rstemp.h>
#include <rsnull.h>
#include <rsbool.h>
#include <rsrel.h>
-int index_word_prefix_map (char *string, oid_value attrSet, int attrUse,
- char *basename)
-{
- attent *attp;
-
- logf (LOG_DEBUG, "oid_value attrSet = %d, attrUse = %d", attrSet, attrUse);
- attp = att_getentbyatt (attrSet, attrUse);
- if (!attp)
- return -1;
- logf (LOG_DEBUG, "ord=%d", attp->attset_ordinal);
- return index_word_prefix (string, attp->attset_ordinal,
- attp->local_attributes->local, basename);
-}
-
typedef struct {
int type;
int major;
return rset_trunc_r (isam, isam_p, 0, no, 100);
}
-struct grep_info {
- ISAM_P *isam_p_buf;
- int isam_p_size;
- int isam_p_indx;
-};
+#define TERM_COUNT
+
+struct grep_info {
+#ifdef TERM_COUNT
+ int *term_no;
+#endif
+ ISAM_P *isam_p_buf;
+ int isam_p_size;
+ int isam_p_indx;
+};
static void add_isam_p (const char *info, struct grep_info *p)
{
if (p->isam_p_indx == p->isam_p_size)
{
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 = xmalloc (sizeof(*new_isam_p_buf) *
xfree (p->isam_p_buf);
}
p->isam_p_buf = new_isam_p_buf;
+
+#ifdef TERM_COUNT
+ new_term_no = xmalloc (sizeof(*new_term_no) *
+ p->isam_p_size);
+ if (p->term_no)
+ {
+ memcpy (new_term_no, p->isam_p_buf,
+ p->isam_p_indx * sizeof(*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));
static int grep_handle (char *name, const char *info, void *p)
{
- logf (LOG_DEBUG, "dict name: %s", name);
add_isam_p (info, p);
return 0;
}
+/* gen_regular_rel - generate regular expression from relation
+ * val: border value (inclusive)
+ * islt: 1 if <=; 0 if >=.
+ */
static void gen_regular_rel (char *dst, int val, int islt)
{
- int dst_p = 1;
+ int dst_p;
int w, d, i;
int pos = 0;
char numstr[20];
- *dst = '(';
+ logf (LOG_DEBUG, "gen_regular_rel. val=%d, islt=%d", val, islt);
+ if (val >= 0)
+ {
+ if (islt)
+ strcpy (dst, "(-[0-9]+|");
+ else
+ strcpy (dst, "(");
+ }
+ else
+ {
+ if (!islt)
+ {
+ strcpy (dst, "([0-9]+|-");
+ dst_p = strlen (dst);
+ islt = 1;
+ }
+ else
+ {
+ strcpy (dst, "(-");
+ islt = 0;
+ }
+ val = -val;
+ }
+ dst_p = strlen (dst);
sprintf (numstr, "%d", val);
for (w = strlen(numstr); --w >= 0; pos++)
{
{
zi->errCode = 109; /* Database unavailable */
zi->errString = basenames[base_no];
+ return -1;
}
for (local_attr = attp->local_attributes; local_attr;
local_attr = local_attr->next)
term_dict[prefix_len++] = '|';
else
term_dict[prefix_len++] = '(';
- if ((ord >= 'A' && ord <= 'Z') || (ord >= 'a' && ord <= 'z'))
- term_dict[prefix_len++] = ord;
- else
- {
- term_dict[prefix_len++] = '\\';
- term_dict[prefix_len++] = ord;
- }
+ term_dict[prefix_len++] = 1;
+ term_dict[prefix_len++] = ord;
}
if (!prefix_len)
{
zi->errCode = 114;
return -1;
}
- term_dict[prefix_len++] = ')';
+ term_dict[prefix_len++] = ')';
+ term_dict[prefix_len++] = 1;
term_dict[prefix_len++] = regType;
term_dict[prefix_len] = '\0';
if (!relational_term (zi, zapt, term_sub, term_dict,
static void trans_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt,
char *termz)
{
- size_t i, sizez;
- int sep = 0;
+ size_t sizez;
Z_Term *term = zapt->term;
sizez = term->u.general->len;
if (sizez > IT_MAX_WORD-1)
sizez = IT_MAX_WORD-1;
- for (i = 0; i < sizez; i++)
+ memcpy (termz, term->u.general->buf, sizez);
+ termz[sizez] = '\0';
+}
+
+static void trans_scan_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt,
+ char *termz)
+{
+ Z_Term *term = zapt->term;
+ 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;
+ int prev_space = 0;
+ int len;
+
+ while ((len = (cp_end - cp)) > 0)
{
- if (!isalnum(term->u.general->buf[i]))
- sep = 1;
- else
+ map = map_chrs_input (&cp, len);
+ if (**map == *CHR_SPACE)
{
- if (sep)
- *termz++ = ' ';
- *termz++ = index_char_cvt (term->u.general->buf[i]);
- sep = 0;
- }
+ if (prev_space)
+ continue;
+ prev_space = 1;
+ }
+ else
+ prev_space = 0;
+ for (src = *map; *src; src++)
+ termz[i++] = *src;
}
- if (sep)
- *termz++ = ' ';
- *termz = '\0';
+ termz[i] = '\0';
}
static RSET rpn_search_APT_relevance (ZServerInfo *zi,
char termz[IT_MAX_WORD+1];
char term_sub[IT_MAX_WORD+1];
struct grep_info grep_info;
- char *p0 = termz, *p1 = NULL;
+ char *p0 = termz;
RSET result;
+ int term_index = 0;
parms.key_size = sizeof(struct it_key);
parms.max_rec = 100;
parms.cmp = key_compare;
parms.is = zi->wordIsam;
+ parms.no_terms = 0;
if (zapt->term->which != Z_Term_general)
{
}
trans_term (zi, zapt, termz);
+#ifdef TERM_COUNT
+ grep_info.term_no = 0;
+#endif
grep_info.isam_p_indx = 0;
grep_info.isam_p_size = 0;
grep_info.isam_p_buf = NULL;
while (1)
{
- if ((p1 = strchr (p0, ' ')))
+ char **map;
+ char *p2, *p1;
+
+ p1 = p0;
+ while (*(p0 = p1))
{
- memcpy (term_sub, p0, p1-p0);
- term_sub[p1-p0] = '\0';
+ map = map_chrs_input (&p1, strlen(p1));
+ if (**map != *CHR_SPACE)
+ break;
}
- else
- strcpy (term_sub, p0);
+ if (!*p0)
+ break;
+
+ p1 = p0;
+ while (*(p2 = p1))
+ {
+ map = map_chrs_input (&p1, strlen(p1));
+ if (**map == *CHR_SPACE)
+ break;
+ }
+ if (p2 == p0)
+ break;
+ memcpy (term_sub, p0, p2-p0);
+ term_sub[p2-p0] = '\0';
+ p0 = p2;
if (field_term (zi, zapt, term_sub, 'w', attributeSet, &grep_info,
num_bases, basenames))
return NULL;
- if (!p1)
- break;
- p0 = p1;
- while (*++p0 == ' ')
- ;
+#ifdef TERM_COUNT
+ for (; term_index < grep_info.isam_p_indx; term_index++)
+ grep_info.term_no[term_index] = parms.no_terms;
+ parms.no_terms++;
+#endif
}
+ parms.term_no = grep_info.term_no;
parms.isam_positions = grep_info.isam_p_buf;
parms.no_isam_positions = grep_info.isam_p_indx;
if (grep_info.isam_p_indx > 0)
result = rset_create (rset_kind_relevance, &parms);
else
result = rset_create (rset_kind_null, NULL);
+#ifdef TERM_COUNT
+ xfree(grep_info.term_no);
+#endif
xfree (grep_info.isam_p_buf);
return result;
}
}
trans_term (zi, zapt, termz);
+#ifdef TERM_COUNT
+ grep_info.term_no = 0;
+#endif
grep_info.isam_p_indx = 0;
grep_info.isam_p_size = 0;
grep_info.isam_p_buf = NULL;
else
result = rset_trunc (zi->wordIsam, grep_info.isam_p_buf,
grep_info.isam_p_indx);
- xfree (grep_info.isam_p_buf);
- return result;
-}
-static RSET rpn_search_APT_word (ZServerInfo *zi,
- Z_AttributesPlusTerm *zapt,
- oid_value attributeSet,
- int num_bases, char **basenames)
-{
- rset_isam_parms parms;
- char termz[IT_MAX_WORD+1];
- struct grep_info grep_info;
- RSET result;
-
- if (zapt->term->which != Z_Term_general)
- {
- zi->errCode = 124;
- return NULL;
- }
- trans_term (zi, zapt, termz);
-
- grep_info.isam_p_indx = 0;
- grep_info.isam_p_size = 0;
- grep_info.isam_p_buf = NULL;
-
- if (field_term (zi, zapt, termz, 'w', attributeSet, &grep_info,
- num_bases, basenames))
- return NULL;
- if (grep_info.isam_p_indx < 1)
- result = rset_create (rset_kind_null, NULL);
- else if (grep_info.isam_p_indx == 1)
- {
- parms.is = zi->wordIsam;
- parms.pos = *grep_info.isam_p_buf;
- result = rset_create (rset_kind_isam, &parms);
- }
- else
- result = rset_trunc (zi->wordIsam, grep_info.isam_p_buf,
- grep_info.isam_p_indx);
+#ifdef TERM_COUNT
+ xfree(grep_info.term_no);
+#endif
xfree (grep_info.isam_p_buf);
return result;
}
{
char termz[IT_MAX_WORD+1];
char term_sub[IT_MAX_WORD+1];
- char *p0 = termz, *p1 = NULL;
+ char *p0 = termz;
RSET rset[60], result;
int i, rset_no = 0;
struct grep_info grep_info;
}
trans_term (zi, zapt, termz);
+#ifdef TERM_COUNT
+ grep_info.term_no = 0;
+#endif
grep_info.isam_p_size = 0;
grep_info.isam_p_buf = NULL;
while (1)
{
- if ((p1 = strchr (p0, ' ')))
+ char **map;
+ char *p2, *p1;
+
+ p1 = p0;
+ while (*(p0 = p1))
{
- memcpy (term_sub, p0, p1-p0);
- term_sub[p1-p0] = '\0';
+ map = map_chrs_input (&p1, strlen(p1));
+ if (**map != *CHR_SPACE)
+ break;
}
- else
- strcpy (term_sub, p0);
+ if (!*p0)
+ break;
+
+ p1 = p0;
+ while (*(p2 = p1))
+ {
+ map = map_chrs_input (&p1, strlen(p1));
+ if (**map == *CHR_SPACE)
+ break;
+ }
+ if (p2 == p0)
+ break;
+
+ memcpy (term_sub, p0, p2-p0);
+ term_sub[p2-p0] = '\0';
+ p0 = p2;
grep_info.isam_p_indx = 0;
if (field_term (zi, zapt, term_sub, 'w', attributeSet, &grep_info,
assert (rset[rset_no]);
if (++rset_no >= sizeof(rset)/sizeof(*rset))
break;
- if (!p1)
- break;
- p0 = p1;
- while (*++p0 == ' ')
- ;
}
+#ifdef TERM_COUNT
+ xfree(grep_info.term_no);
+#endif
xfree (grep_info.isam_p_buf);
if (rset_no == 0)
return rset_create (rset_kind_null, NULL);
if (completeness_value == 2 || completeness_value == 3)
return rpn_search_APT_cphrase (zi, zapt, attributeSet,
num_bases, basenames);
- return rpn_search_APT_word (zi, zapt, attributeSet,
- num_bases, basenames);
+ return rpn_search_APT_phrase (zi, zapt, attributeSet,
+ num_bases, basenames);
case 3: /* key */
break;
case 4: /* year */
case 107: /* local-number */
return rpn_search_APT_local (zi, zapt, attributeSet);
case 108: /* string */
- return rpn_search_APT_word (zi, zapt, attributeSet,
- num_bases, basenames);
+ return rpn_search_APT_phrase (zi, zapt, attributeSet,
+ num_bases, basenames);
case 109: /* numeric string */
break;
}
oident *attrset;
oid_value attributeSet;
+ dict_grep_cmap (zi->wordDict, map_chrs_input);
zlog_rpn (rpn);
zi->errCode = 0;
}
+static void scan_term_untrans (ODR odr, char **dstp, const char *src)
+{
+ char *dst = odr_malloc (odr, strlen(src)*2+1);
+ *dstp = dst;
+
+ while (*src)
+ {
+ const char *cp = map_chrs_output (&src);
+ while (*cp)
+ *dst++ = *cp++;
+ }
+ *dst = '\0';
+}
+
int rpn_scan (ZServerInfo *zi, Z_AttributesPlusTerm *zapt,
oid_value attributeset,
int num_bases, char **basenames,
logf (LOG_DEBUG, "scan, position = %d, num = %d", pos, num);
+ if (attributeset == VAL_NONE)
+ attributeset = VAL_BIB1;
+
attr_init (&use, zapt, 1);
- use_value = attr_find (&use, NULL);
+ use_value = attr_find (&use, &attributeset);
logf (LOG_DEBUG, "use value %d", use_value);
attr_init (&completeness, zapt, 6);
completeness_value = attr_find (&completeness, NULL);
logf (LOG_DEBUG, "completeness value %d", completeness_value);
- if (attributeset == VAL_NONE)
- attributeset = VAL_BIB1;
-
if (use_value == -1)
use_value = 1016;
for (base_no = 0; base_no < num_bases && ord_no < 32; base_no++)
(completeness_value==2 || completeness_value==3) ? 'p': 'w';
termz[prefix_len] = 0;
strcpy (scan_info->prefix, termz);
- trans_term (zi, zapt, termz+prefix_len);
+
+ trans_scan_term (zi, zapt, termz+prefix_len);
+
dict_scan (zi->wordDict, termz, &before_tmp, &after_tmp, scan_info,
scan_handle);
}
}
if (j0 == -1)
break;
- glist[i+before].term = odr_malloc (zi->odr, strlen(mterm)+1);
- strcpy (glist[i+before].term, mterm);
-
+ scan_term_untrans (zi->odr, &glist[i+before].term, mterm);
parms.is = zi->wordIsam;
parms.pos = scan_info_array[j0].list[ptr[j0]].isam_p;
rset = rset_create (rset_kind_isam, &parms);
}
if (j0 == -1)
break;
- glist[before-1-i].term = odr_malloc (zi->odr, strlen(mterm)+1);
- strcpy (glist[before-1-i].term, mterm);
+
+ scan_term_untrans (zi->odr, &glist[before-1-i].term, mterm);
parms.is = zi->wordIsam;
parms.pos = scan_info_array[j0].list[before-1-ptr[j0]].isam_p;
projects / idzebra-moved-to-github.git / blobdiff