* Sebastian Hammer, Adam Dickmeiss
*
* $Log: zrpn.c,v $
- * Revision 1.30 1995-10-16 09:32:38 adam
+ * Revision 1.36 1995-12-06 12:41:27 adam
+ * New command 'stat' for the index program.
+ * Filenames can be read from stdin by specifying '-'.
+ * Bug fix/enhancement of the transformation from terms to regular
+ * expressons in the search engine.
+ *
+ * Revision 1.35 1995/11/27 09:29:00 adam
+ * Bug fixes regarding conversion to regular expressions.
+ *
+ * Revision 1.34 1995/11/16 17:00:56 adam
+ * Better logging of rpn query.
+ *
+ * Revision 1.33 1995/11/01 13:58:28 quinn
+ * Moving data1 to yaz/retrieval
+ *
+ * Revision 1.32 1995/10/27 14:00:11 adam
+ * Implemented detection of database availability.
+ *
+ * Revision 1.31 1995/10/17 18:02:10 adam
+ * New feature: databases. Implemented as prefix to words in dictionary.
+ *
+ * Revision 1.30 1995/10/16 09:32:38 adam
* More work on relational op.
*
* Revision 1.29 1995/10/13 16:01:49 adam
#include <stdio.h>
#include <assert.h>
#include <unistd.h>
+#include <ctype.h>
#include "zserver.h"
-#include <attribute.h>
+#include "attribute.h"
#include <rsisam.h>
#include <rstemp.h>
#include <rsbool.h>
#include <rsrel.h>
-int index_word_prefix_map (char *string, oid_value attrSet, int attrUse)
+int index_word_prefix_map (char *string, oid_value attrSet, int attrUse,
+ char *basename)
{
attent *attp;
return -1;
logf (LOG_DEBUG, "ord=%d", attp->attset_ordinal);
return index_word_prefix (string, attp->attset_ordinal,
- attp->local_attribute);
-}
-
-/*
- * attr_print: log attributes
- */
-static void attr_print (Z_AttributesPlusTerm *t)
-{
- int of, i;
- for (of = 0; of < t->num_attributes; of++)
- {
- Z_AttributeElement *element;
- element = t->attributeList[of];
-
- switch (element->which)
- {
- case Z_AttributeValue_numeric:
- logf (LOG_DEBUG, "attributeType=%d value=%d",
- *element->attributeType,
- *element->value.numeric);
- break;
- case Z_AttributeValue_complex:
- logf (LOG_DEBUG, "attributeType=%d complex",
- *element->attributeType);
- for (i = 0; i<element->value.complex->num_list; i++)
- {
- if (element->value.complex->list[i]->which ==
- Z_StringOrNumeric_string)
- logf (LOG_DEBUG, " string: '%s'",
- element->value.complex->list[i]->u.string);
- else if (element->value.complex->list[i]->which ==
- Z_StringOrNumeric_numeric)
- logf (LOG_DEBUG, " numeric: '%d'",
- *element->value.complex->list[i]->u.numeric);
- }
- break;
- default:
- assert (0);
- }
- }
+ attp->local_attribute, basename);
}
typedef struct {
return 0;
}
-static void gen_regular_ge (char *dst, int val)
+static void gen_regular_rel (char *dst, int val, int islt)
{
int dst_p = 1;
int w, d, i;
char numstr[20];
*dst = '(';
- if (val < 0)
- val = 0;
sprintf (numstr, "%d", val);
for (w = strlen(numstr); --w >= 0; pos++)
{
d = numstr[w];
if (pos > 0)
{
- if (d == '9')
- continue;
- d++;
+ if (islt)
+ {
+ if (d == '0')
+ continue;
+ d--;
+ }
+ else
+ {
+ if (d == '9')
+ continue;
+ d++;
+ }
}
strcpy (dst + dst_p, numstr);
dst_p = strlen(dst) - pos - 1;
- if (d <= '8')
- {
- dst[dst_p++] = '[';
- dst[dst_p++] = d;
- dst[dst_p++] = '-';
- dst[dst_p++] = '9';
- dst[dst_p++] = ']';
+ if (islt)
+ {
+ if (d != '0')
+ {
+ dst[dst_p++] = '[';
+ dst[dst_p++] = '0';
+ dst[dst_p++] = '-';
+ dst[dst_p++] = d;
+ dst[dst_p++] = ']';
+ }
+ else
+ dst[dst_p++] = d;
}
else
- dst[dst_p++] = d;
-
+ {
+ if (d != '9')
+ {
+ dst[dst_p++] = '[';
+ dst[dst_p++] = d;
+ dst[dst_p++] = '-';
+ dst[dst_p++] = '9';
+ dst[dst_p++] = ']';
+ }
+ else
+ dst[dst_p++] = d;
+ }
for (i = 0; i<pos; i++)
{
dst[dst_p++] = '[';
dst[dst_p++] = '|';
}
dst[dst_p] = '\0';
- for (i = 0; i <= pos; i++)
- strcat (dst, "[0-9]");
- strcat (dst, "[0-9]*)");
+ if (islt)
+ {
+ for (i=1; i<pos; i++)
+ strcat (dst, "[0-9]?");
+ }
+ else
+ {
+ for (i = 0; i <= pos; i++)
+ strcat (dst, "[0-9]");
+ strcat (dst, "[0-9]*");
+ }
+ strcat (dst, ")");
+}
+
+static int relational_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt,
+ const char *term_sub,
+ char *term_dict,
+ oid_value attributeSet,
+ struct grep_info *grep_info,
+ int *max_pos)
+{
+ AttrType relation;
+ int relation_value;
+ int term_value;
+ int r;
+
+ attr_init (&relation, zapt, 2);
+ relation_value = attr_find (&relation, NULL);
+ term_value = atoi (term_sub);
+
+ switch (relation_value)
+ {
+ case 1:
+ if (term_value <= 0)
+ return 1;
+ logf (LOG_DEBUG, "Relation <");
+ gen_regular_rel (term_dict + strlen(term_dict), term_value-1, 1);
+ break;
+ case 2:
+ if (term_value < 0)
+ return 1;
+ logf (LOG_DEBUG, "Relation <=");
+ gen_regular_rel (term_dict + strlen(term_dict), term_value, 1);
+ break;
+ case 4:
+ if (term_value < 0)
+ term_value = 0;
+ logf (LOG_DEBUG, "Relation >=");
+ gen_regular_rel (term_dict + strlen(term_dict), term_value, 0);
+ break;
+ case 5:
+ if (term_value < 0)
+ term_value = 0;
+ logf (LOG_DEBUG, "Relation >");
+ gen_regular_rel (term_dict + strlen(term_dict), term_value+1, 0);
+ break;
+ default:
+ return 0;
+ }
+ logf (LOG_DEBUG, "dict_lookup_grep: %s", term_dict);
+ r = dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info, max_pos,
+ 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);
+ return 1;
+}
+
+static void verbatim_char (int ch, int *indx, char *dst)
+{
+ if (!isalnum (ch))
+ dst[(*indx)++] = '\\';
+ dst[(*indx)++] = ch;
}
static int trunc_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt,
const char *term_sub,
- oid_value attributeSet, struct grep_info *grep_info)
+ oid_value attributeSet, struct grep_info *grep_info,
+ int num_bases, char **basenames)
{
- char term_dict[10*IT_MAX_WORD+2];
- int i, j, r;
- const char *info;
+ char term_dict[2*IT_MAX_WORD+2];
+ int i, j, r, base_no;
AttrType truncation;
int truncation_value;
AttrType use;
int use_value;
- AttrType relation;
- int relation_value;
oid_value curAttributeSet = attributeSet;
attr_init (&use, zapt, 1);
truncation_value = attr_find (&truncation, NULL);
logf (LOG_DEBUG, "truncation value %d", truncation_value);
- attr_init (&relation, zapt, 2);
- relation_value = attr_find (&relation, NULL);
-
if (use_value == -1)
use_value = 1016;
- i = index_word_prefix_map (term_dict, curAttributeSet, use_value);
- if (i < 0)
- {
- zi->errCode = 114;
- return -1;
- }
- switch (relation_value)
- {
- case 1:
- case 2:
- break;
- case 4:
- logf (LOG_LOG, "Relation ge");
- gen_regular_ge (term_dict + strlen(term_dict), atoi(term_sub));
- logf (LOG_LOG, "dict_lookup_grep: %s", term_dict);
- r = dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info,
- grep_handle);
- if (r)
- logf (LOG_WARN, "dict_lookup_grep fail, rel=ge: %d", r);
- logf (LOG_LOG, "%d positions", grep_info->isam_p_indx);
- return 0;
- case 5:
- logf (LOG_LOG, "Relation gt");
- gen_regular_ge (term_dict + strlen(term_dict), atoi(term_sub)+1);
- logf (LOG_LOG, "dict_lookup_grep: %s", term_dict);
- r = dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info,
- grep_handle);
- if (r)
- logf (LOG_WARN, "dict_lookup_grep fail, rel=gt: %d", r);
- logf (LOG_LOG, "%d positions", grep_info->isam_p_indx);
- return 0;
- }
- switch (truncation_value)
+
+ for (base_no = 0; base_no < num_bases; base_no++)
{
- case -1: /* not specified */
- case 100: /* do not truncate */
- strcat (term_dict, term_sub);
- logf (LOG_DEBUG, "dict_lookup: %s", term_dict);
- if ((info = dict_lookup (zi->wordDict, term_dict)))
- add_isam_p (info, grep_info);
- break;
- case 1: /* right truncation */
- strcat (term_dict, term_sub);
- strcat (term_dict, ".*");
- dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info, grep_handle);
- break;
- case 2: /* left truncation */
- case 3: /* left&right truncation */
- zi->errCode = 120;
- return -1;
- case 101: /* process # in term */
- for (j = strlen(term_dict), i = 0; term_sub[i] && i < 2; i++)
- term_dict[j++] = term_sub[i];
- for (; term_sub[i]; i++)
- if (term_sub[i] == '#')
+ int max_pos;
+ int prefix_len = index_word_prefix_map (term_dict, curAttributeSet,
+ use_value,
+ basenames[base_no]);
+ if (prefix_len < 0)
+ {
+ zi->errCode = 114;
+ return -1;
+ }
+ if (!relational_term (zi, zapt, term_sub, term_dict,
+ attributeSet, grep_info, &max_pos))
+ {
+ switch (truncation_value)
{
- term_dict[j++] = '.';
- term_dict[j++] = '*';
+ case -1: /* not specified */
+ case 100: /* do not truncate */
+ j = strlen(term_dict);
+ term_dict[j++] = '(';
+ for (i = 0; term_sub[i]; i++)
+ verbatim_char (term_sub[i], &j, term_dict);
+ strcpy (term_dict+j, ")");
+ r = dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info,
+ &max_pos, grep_handle);
+ if (r)
+ logf (LOG_WARN, "dict_lookup_grep err, trunc=none:%d", r);
+ break;
+ case 1: /* right truncation */
+ j = strlen(term_dict);
+ term_dict[j++] = '(';
+ for (i = 0; term_sub[i]; i++)
+ verbatim_char (term_sub[i], &j, term_dict);
+ strcpy (term_dict+j, ".*)");
+ dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info,
+ &max_pos, grep_handle);
+ break;
+ case 2: /* left truncation */
+ case 3: /* left&right truncation */
+ zi->errCode = 120;
+ return -1;
+ case 101: /* process # in term */
+ j = strlen(term_dict);
+ term_dict[j++] = '(';
+ for (i=0; term_sub[i]; i++)
+ if (term_sub[i] == '#' && i > 2)
+ {
+ term_dict[j++] = '.';
+ term_dict[j++] = '*';
+ }
+ else
+ verbatim_char (term_sub[i], &j, term_dict);
+ strcpy (term_dict+j, ")");
+ r = dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info,
+ &max_pos, grep_handle);
+ if (r)
+ logf (LOG_WARN, "dict_lookup_grep err, trunc=#: %d",
+ r);
+ break;
+ case 102: /* regular expression */
+ sprintf (term_dict + strlen(term_dict), "(%s)", term_sub);
+ r = dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info,
+ &max_pos, grep_handle);
+ if (r)
+ logf (LOG_WARN, "dict_lookup_grep err, trunc=regular: %d",
+ r);
+ break;
}
- else
- term_dict[j++] = term_sub[i];
- term_dict[j] = '\0';
- dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info, grep_handle);
- break;
- case 102: /* regular expression */
- strcat (term_dict, "(");
- strcat (term_dict, term_sub);
- strcat (term_dict, ")");
- logf (LOG_LOG, "dict_lookup_grep: %s", term_dict);
- r = dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info,
- grep_handle);
- if (r)
- logf (LOG_WARN, "dict_lookup_grep fail, truncation=regular: %d", r);
- break;
+ }
+ if (max_pos <= strlen(basenames[base_no]))
+ {
+ zi->errCode = 109; /* Database unavailable */
+ zi->errString = basenames[base_no];
+ return -1;
+ }
}
- logf (LOG_LOG, "%d positions", grep_info->isam_p_indx);
+ logf (LOG_DEBUG, "%d positions", grep_info->isam_p_indx);
return 0;
}
static RSET rpn_search_APT_relevance (ZServerInfo *zi,
Z_AttributesPlusTerm *zapt,
- oid_value attributeSet)
+ oid_value attributeSet,
+ int num_bases, char **basenames)
{
rset_relevance_parms parms;
char termz[IT_MAX_WORD+1];
}
else
strcpy (term_sub, p0);
- if (trunc_term (zi, zapt, term_sub, attributeSet, &grep_info))
+ if (trunc_term (zi, zapt, term_sub, attributeSet, &grep_info,
+ num_bases, basenames))
return NULL;
if (!p1)
break;
static RSET rpn_search_APT_word (ZServerInfo *zi,
Z_AttributesPlusTerm *zapt,
- oid_value attributeSet)
+ oid_value attributeSet,
+ int num_bases, char **basenames)
{
rset_isam_parms parms;
char termz[IT_MAX_WORD+1];
grep_info.isam_p_size = 0;
grep_info.isam_p_buf = NULL;
- if (trunc_term (zi, zapt, termz, attributeSet, &grep_info))
+ if (trunc_term (zi, zapt, termz, attributeSet, &grep_info,
+ num_bases, basenames))
return NULL;
if (grep_info.isam_p_indx < 1)
result = rset_create (rset_kind_null, NULL);
static RSET rpn_search_APT_phrase (ZServerInfo *zi,
Z_AttributesPlusTerm *zapt,
- oid_value attributeSet)
+ oid_value attributeSet,
+ int num_bases, char **basenames)
{
char termz[IT_MAX_WORD+1];
char term_sub[IT_MAX_WORD+1];
strcpy (term_sub, p0);
grep_info.isam_p_indx = 0;
- if (trunc_term (zi, zapt, term_sub, attributeSet, &grep_info))
+ if (trunc_term (zi, zapt, term_sub, attributeSet, &grep_info,
+ num_bases, basenames))
return NULL;
if (grep_info.isam_p_indx == 0)
rset[rset_no] = rset_create (rset_kind_null, NULL);
}
static RSET rpn_search_APT (ZServerInfo *zi, Z_AttributesPlusTerm *zapt,
- oid_value attributeSet)
+ oid_value attributeSet,
+ int num_bases, char **basenames)
{
AttrType relation;
AttrType structure;
{
case -1:
if (relation_value == 102) /* relevance relation */
- return rpn_search_APT_relevance (zi, zapt, attributeSet);
- return rpn_search_APT_phrase (zi, zapt, attributeSet);
+ return rpn_search_APT_relevance (zi, zapt, attributeSet,
+ num_bases, basenames);
+ return rpn_search_APT_phrase (zi, zapt, attributeSet,
+ num_bases, basenames);
case 1: /* phrase */
if (relation_value == 102) /* relevance relation */
- return rpn_search_APT_relevance (zi, zapt, attributeSet);
- return rpn_search_APT_phrase (zi, zapt, attributeSet);
+ return rpn_search_APT_relevance (zi, zapt, attributeSet,
+ num_bases, basenames);
+ return rpn_search_APT_phrase (zi, zapt, attributeSet,
+ num_bases, basenames);
break;
case 2: /* word */
if (relation_value == 102) /* relevance relation */
- return rpn_search_APT_relevance (zi, zapt, attributeSet);
- return rpn_search_APT_word (zi, zapt, attributeSet);
+ return rpn_search_APT_relevance (zi, zapt, attributeSet,
+ num_bases, basenames);
+ return rpn_search_APT_word (zi, zapt, attributeSet,
+ num_bases, basenames);
case 3: /* key */
break;
case 4: /* year */
case 5: /* date - normalized */
break;
case 6: /* word list */
- return rpn_search_APT_relevance (zi, zapt, attributeSet);
+ return rpn_search_APT_relevance (zi, zapt, attributeSet,
+ num_bases, basenames);
case 100: /* date - un-normalized */
break;
case 101: /* name - normalized */
case 104: /* urx */
break;
case 105: /* free-form-text */
- return rpn_search_APT_relevance (zi, zapt, attributeSet);
+ return rpn_search_APT_relevance (zi, zapt, attributeSet,
+ num_bases, basenames);
case 106: /* document-text */
- return rpn_search_APT_relevance (zi, zapt, attributeSet);
+ return rpn_search_APT_relevance (zi, zapt, attributeSet,
+ num_bases, basenames);
case 107: /* local-number */
return rpn_search_APT_local (zi, zapt, attributeSet);
case 108: /* string */
- return rpn_search_APT_word (zi, zapt, attributeSet);
+ return rpn_search_APT_word (zi, zapt, attributeSet,
+ num_bases, basenames);
case 109: /* numeric string */
break;
}
}
static RSET rpn_search_structure (ZServerInfo *zi, Z_RPNStructure *zs,
- oid_value attributeSet)
+ oid_value attributeSet,
+ int num_bases, char **basenames)
{
RSET r = NULL;
if (zs->which == Z_RPNStructure_complex)
rset_bool_parms bool_parms;
bool_parms.rset_l = rpn_search_structure (zi, zs->u.complex->s1,
- attributeSet);
+ attributeSet,
+ num_bases, basenames);
if (bool_parms.rset_l == NULL)
return NULL;
bool_parms.rset_r = rpn_search_structure (zi, zs->u.complex->s2,
- attributeSet);
+ attributeSet,
+ num_bases, basenames);
if (bool_parms.rset_r == NULL)
{
rset_delete (bool_parms.rset_l);
{
logf (LOG_DEBUG, "rpn_search_APT");
r = rpn_search_APT (zi, zs->u.simple->u.attributesPlusTerm,
- attributeSet);
+ attributeSet, num_bases, basenames);
}
else if (zs->u.simple->which == Z_Operand_resultSetId)
{
oident *attrset;
oid_value attributeSet;
+ zlog_rpn (rpn);
+
zi->errCode = 0;
zi->errString = NULL;
-
+
attrset = oid_getentbyoid (rpn->attributeSetId);
attributeSet = attrset->value;
-
- rset = rpn_search_structure (zi, rpn->RPNStructure, attributeSet);
+ rset = rpn_search_structure (zi, rpn->RPNStructure, attributeSet,
+ num_bases, basenames);
if (!rset)
return zi->errCode;
count_set (rset, hits);
return 0;
}
-int rpn_scan (ZServerInfo *zi, ODR odr, Z_AttributesPlusTerm *zapt,
+
+static int dummy_handle (Dict_char *name, const char *info, void *p)
+{
+ return 0;
+}
+
+int rpn_scan (ZServerInfo *zi, Z_AttributesPlusTerm *zapt,
+ int num_bases, char **basenames,
int *position, int *num_entries, struct scan_entry **list,
int *status)
{
- int i, j, sizez;
+ int i, j, sizez, max_pos;
int pos = *position;
int num = *num_entries;
int before;
struct scan_info scan_info;
logf (LOG_DEBUG, "scan, position = %d, num = %d", pos, num);
+
+ if (num_bases != 1)
+ return 111;
scan_info.before = before = pos-1;
scan_info.after = after = 1+num-pos;
- scan_info.odr = odr;
+ scan_info.odr = zi->odr;
logf (LOG_DEBUG, "scan, before = %d, after = %d", before, after);
scan_info.isam = zi->wordIsam;
- scan_info.list = odr_malloc (odr, (before+after)*sizeof(*scan_info.list));
+ scan_info.list = odr_malloc (zi->odr, (before+after)*
+ sizeof(*scan_info.list));
for (j = 0; j<before+after; j++)
scan_info.list[j].term = NULL;
attr_init (&use, zapt, 1);
if (use_value == -1)
use_value = 1016;
- i = index_word_prefix (termz, 1, use_value);
+ i = index_word_prefix (termz, 1, use_value, *basenames);
+
+ dict_lookup_grep (zi->wordDict, termz, 0, NULL, &max_pos,
+ dummy_handle);
+ if (max_pos <= strlen(*basenames))
+ {
+ zi->errString = *basenames;
+ return zi->errCode = 109; /* Database unavailable */
+ }
strcpy (scan_info.prefix, termz);
sizez = term->u.general->len;
if (sizez > IT_MAX_WORD)