-/*
- * Copyright (C) 1994, Index Data I/S
- * All rights reserved.
- * Sebastian Hammer, Adam Dickmeiss
- *
- * $Log: lookgrep.c,v $
- * Revision 1.3 1994-10-05 12:16:50 adam
- * Pagesize is a resource now.
- *
- * Revision 1.2 1994/10/04 12:08:07 adam
- * Some bug fixes and some optimizations.
- *
- * Revision 1.1 1994/10/03 17:23:04 adam
- * First version of dictionary lookup with regular expressions and errors.
- *
- */
+/* $Id: lookgrep.c,v 1.31 2006-05-10 08:13:18 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 <stdlib.h>
#include <string.h>
#include <assert.h>
#include <dfa.h>
-#include <dict.h>
+#include "dict-p.h"
typedef unsigned MatchWord;
#define WORD_BITS 32
#define MAX_LENGTH 1024
+/* This code is based
+ * Sun Wu and Udi Manber: Fast Text Searching Allowing Errors.
+ * Communications of the ACM, pp. 83-91, Vol. 35, No. 10, Oct. 1992, USA.
+ * PostScript version of the paper in its submitted form: agrep1.ps)
+ * recommended reading to understand AGREP !
+ *
+ * http://www.tgries.de/agrep/#AGREP1PS
+ * http://www.tgries.de/agrep/doc/agrep1ps.zip
+ */
+
typedef struct {
int n; /* no of MatchWord needed */
int range; /* max no. of errors */
return m[mc->n * ch + wno] & (1<<off);
}
-static MatchContext *mk_MatchContext (DFA_states *dfas, int range)
+static MatchContext *mk_MatchContext (struct DFA *dfa, int range)
{
- MatchContext *mc = xmalloc (sizeof(*mc));
+ MatchContext *mc = (MatchContext *) xmalloc (sizeof(*mc));
int s;
- mc->n = (dfas->no+WORD_BITS) / WORD_BITS;
+ mc->n = (dfa->no_states+WORD_BITS) / WORD_BITS;
mc->range = range;
mc->fact = (range+1)*mc->n;
- mc->match_mask = xcalloc (mc->n, sizeof(*mc->match_mask));
+ mc->match_mask = (MatchWord *) xcalloc (mc->n, sizeof(*mc->match_mask));
- for (s = 0; s<dfas->no; s++)
- if (dfas->sortarray[s]->rule_no)
+ for (s = 0; s<dfa->no_states; s++)
+ if (dfa->states[s]->rule_no)
set_bit (mc, mc->match_mask, 0, s);
return mc;
}
}
static void mask_shift (MatchContext *mc, MatchWord *Rdst, MatchWord *Rsrc,
- DFA_states *dfas, int ch)
+ struct DFA *dfa, int ch)
{
int j, s = 0;
MatchWord *Rsrc_p = Rsrc, mask;
{
if (mask & 1)
{
- DFA_state *state = dfas->sortarray[s];
+ struct DFA_state *state = dfa->states[s];
int i = state->tran_no;
while (--i >= 0)
if (ch >= state->trans[i].ch[0] &&
}
if (mask & 2)
{
- DFA_state *state = dfas->sortarray[s+1];
+ struct DFA_state *state = dfa->states[s+1];
int i = state->tran_no;
while (--i >= 0)
if (ch >= state->trans[i].ch[0] &&
}
if (mask & 4)
{
- DFA_state *state = dfas->sortarray[s+2];
+ struct DFA_state *state = dfa->states[s+2];
int i = state->tran_no;
while (--i >= 0)
if (ch >= state->trans[i].ch[0] &&
}
if (mask & 8)
{
- DFA_state *state = dfas->sortarray[s+3];
+ struct DFA_state *state = dfa->states[s+3];
int i = state->tran_no;
while (--i >= 0)
if (ch >= state->trans[i].ch[0] &&
}
}
s += 4;
- if (s >= dfas->no)
+ if (s >= dfa->no_states)
return;
mask >>= 4;
}
}
static void shift (MatchContext *mc, MatchWord *Rdst, MatchWord *Rsrc,
- DFA_states *dfas)
+ struct DFA *dfa)
{
int j, s = 0;
MatchWord *Rsrc_p = Rsrc, mask;
{
if (mask & 1)
{
- DFA_state *state = dfas->sortarray[s];
+ struct DFA_state *state = dfa->states[s];
int i = state->tran_no;
while (--i >= 0)
set_bit (mc, Rdst, 0, state->trans[i].to);
}
if (mask & 2)
{
- DFA_state *state = dfas->sortarray[s+1];
+ struct DFA_state *state = dfa->states[s+1];
int i = state->tran_no;
while (--i >= 0)
set_bit (mc, Rdst, 0, state->trans[i].to);
}
if (mask & 4)
{
- DFA_state *state = dfas->sortarray[s+2];
+ struct DFA_state *state = dfa->states[s+2];
int i = state->tran_no;
while (--i >= 0)
set_bit (mc, Rdst, 0, state->trans[i].to);
}
if (mask & 8)
{
- DFA_state *state = dfas->sortarray[s+3];
+ struct DFA_state *state = dfa->states[s+3];
int i = state->tran_no;
while (--i >= 0)
set_bit (mc, Rdst, 0, state->trans[i].to);
}
}
s += 4;
- if (s >= dfas->no)
+ if (s >= dfa->no_states)
return;
mask >>= 4;
}
}
static INLINE int move (MatchContext *mc, MatchWord *Rj1, MatchWord *Rj,
- Dict_char ch, DFA_states *dfas, MatchWord *Rtmp)
+ Dict_char ch, struct DFA *dfa, MatchWord *Rtmp,
+ int range)
{
int d;
MatchWord *Rtmp_2 = Rtmp + mc->n;
- mask_shift (mc, Rj1, Rj, dfas, ch);
+ mask_shift (mc, Rj1, Rj, dfa, ch);
for (d = 1; d <= mc->range; d++)
{
or (mc, Rtmp, Rj, Rj1); /* 2,3 */
- shift (mc, Rtmp_2, Rtmp, dfas);
+ shift (mc, Rtmp_2, Rtmp, dfa);
- mask_shift (mc, Rtmp, Rj+mc->n, dfas, ch); /* 1 */
+ mask_shift (mc, Rtmp, Rj+mc->n, dfa, ch); /* 1 */
or (mc, Rtmp, Rtmp_2, Rtmp); /* 1,2,3*/
static int dict_grep (Dict dict, Dict_ptr ptr, MatchContext *mc,
- MatchWord *Rj, int pos,
- int (*userfunc)(Dict_char *name, char *info),
- Dict_char *prefix, DFA_states *dfas)
+ MatchWord *Rj, int pos, void *client,
+ int (*userfunc)(char *, const char *, void *),
+ Dict_char *prefix, struct DFA *dfa,
+ int *max_pos, int init_pos)
{
int lo, hi, d;
void *p;
dict_bf_readp (dict->dbf, ptr, &p);
lo = 0;
hi = DICT_nodir(p)-1;
- indxp = (short*) ((char*) p+DICT_pagesize(dict)-sizeof(short));
+ indxp = (short*) ((char*) p+DICT_bsize(p)-sizeof(short));
while (lo <= hi)
{
MatchWord *Rj0 = Rj + j *mc->fact;
MatchWord *Rj1 = Rj + (j+1)*mc->fact;
MatchWord *Rj_tmp = Rj + (j+2)*mc->fact;
+ int range;
memcpy (&ch, info+j*sizeof(Dict_char), sizeof(Dict_char));
prefix[pos+j] = ch;
+ if (pos+j > *max_pos)
+ *max_pos = pos+j;
if (ch == DICT_EOS)
{
if (was_match)
- (*userfunc)(prefix, info+(j+1)*sizeof(Dict_char));
+ if ((*userfunc)((char*) prefix,
+ info+(j+1)*sizeof(Dict_char), client))
+ return 1;
break;
}
- move (mc, Rj1, Rj0, ch, dfas, Rj_tmp);
+ if (pos+j >= init_pos)
+ range = mc->range;
+ else
+ range = 0;
+ move (mc, Rj1, Rj0, ch, dfa, Rj_tmp, range);
for (d = mc->n; --d >= 0; )
- if (Rj1[mc->range*mc->n + d])
+ if (Rj1[range*mc->n + d])
break;
if (d < 0)
break;
was_match = 0;
for (d = mc->n; --d >= 0; )
- if (Rj1[mc->range*mc->n + d] & mc->match_mask[d])
+ if (Rj1[range*mc->n + d] & mc->match_mask[d])
{
was_match = 1;
break;
MatchWord *Rj1 = Rj+ mc->fact;
MatchWord *Rj_tmp = Rj+2*mc->fact;
Dict_char ch;
+ int range;
/* Dict_ptr subptr */
/* Dict_char sub char */
memcpy (&ch, info+sizeof(Dict_ptr), sizeof(Dict_char));
prefix[pos] = ch;
- move (mc, Rj1, Rj, ch, dfas, Rj_tmp);
+ if (pos > *max_pos)
+ *max_pos = pos;
+ if (pos >= init_pos)
+ range = mc->range;
+ else
+ range = 0;
+ move (mc, Rj1, Rj, ch, dfa, Rj_tmp, range);
for (d = mc->n; --d >= 0; )
- if (Rj1[mc->range*mc->n + d])
+ if (Rj1[range*mc->n + d])
break;
if (d >= 0)
{
if (info[sizeof(Dict_ptr)+sizeof(Dict_char)])
{
for (d = mc->n; --d >= 0; )
- if (Rj1[mc->range*mc->n + d] & mc->match_mask[d])
+ if (Rj1[range*mc->n + d] & mc->match_mask[d])
{
prefix[pos+1] = DICT_EOS;
- (*userfunc)(prefix, info+sizeof(Dict_ptr)+
- sizeof(Dict_char));
+ if ((*userfunc)((char*) prefix,
+ info+sizeof(Dict_ptr)+
+ sizeof(Dict_char), client))
+ return 1;
break;
}
}
memcpy (&subptr, info, sizeof(Dict_ptr));
if (subptr)
{
- dict_grep (dict, subptr, mc, Rj1, pos+1,
- userfunc, prefix, dfas);
+ if (dict_grep (dict, subptr, mc, Rj1, pos+1,
+ client, userfunc, prefix, dfa, max_pos,
+ init_pos))
+ return 1;
dict_bf_readp (dict->dbf, ptr, &p);
- indxp = (short*) ((char*) p+DICT_pagesize(dict)
- -sizeof(short));
+ indxp = (short*) ((char*) p+DICT_bsize(p)-sizeof(short));
}
}
}
return 0;
}
-int dict_lookup_grep (Dict dict, Dict_char *pattern, int range,
- int (*userfunc)(Dict_char *name, char *info))
+int dict_lookup_grep (Dict dict, const char *pattern, int range, void *client,
+ int *max_pos, int init_pos,
+ int (*userfunc)(char *name, const char *info,
+ void *client))
{
MatchWord *Rj;
Dict_char prefix[MAX_LENGTH+1];
- char *this_pattern = pattern;
- DFA_states *dfas;
+ const char *this_pattern = pattern;
MatchContext *mc;
- DFA *dfa = init_dfa();
+ struct DFA *dfa = dfa_init();
int i, d;
- i = parse_dfa (dfa, &this_pattern, dfa_thompson_chars);
+#if 0
+ debug_dfa_trav = 1;
+ debug_dfa_tran = 1;
+ debug_dfa_followpos = 1;
+ dfa_verbose = 1;
+#endif
+
+ yaz_log (YLOG_DEBUG, "dict_lookup_grep range=%d", range);
+ for (i = 0; pattern[i]; i++)
+ {
+ yaz_log (YLOG_DEBUG, " %3d %c", pattern[i],
+ (pattern[i] > ' ' && pattern[i] < 127) ? pattern[i] : '?');
+ }
+
+ dfa_set_cmap (dfa, dict->grep_cmap_data, dict->grep_cmap);
+
+ i = dfa_parse (dfa, &this_pattern);
if (i || *this_pattern)
{
- rm_dfa (&dfa);
+ yaz_log (YLOG_WARN, "dfa_parse fail=%d", i);
+ dfa_delete (&dfa);
return -1;
}
- dfa->root = dfa->top;
- dfas = mk_dfas (dfa, 50);
- rm_dfa (&dfa);
+ dfa_mkstate (dfa);
- mc = mk_MatchContext (dfas, range);
+ mc = mk_MatchContext (dfa, range);
- Rj = xcalloc ((MAX_LENGTH+1) * mc->n, sizeof(*Rj));
+ Rj = (MatchWord *) xcalloc ((MAX_LENGTH+1) * mc->n, sizeof(*Rj));
set_bit (mc, Rj, 0, 0);
for (d = 1; d<=mc->range; d++)
{
int s;
memcpy (Rj + mc->n * d, Rj + mc->n * (d-1), mc->n * sizeof(*Rj));
- for (s = 0; s<dfas->no; s++)
+ for (s = 0; s<dfa->no_states; s++)
{
if (get_bit (mc, Rj, d-1, s))
{
- DFA_state *state = dfas->sortarray[s];
+ struct DFA_state *state = dfa->states[s];
int i = state->tran_no;
while (--i >= 0)
set_bit (mc, Rj, d, state->trans[i].to);
}
}
}
- i = dict_grep (dict, 1, mc, Rj, 0, userfunc, prefix, dfas);
-
- rm_dfas (&dfas);
+ *max_pos = 0;
+ if (dict->head.root)
+ i = dict_grep (dict, dict->head.root, mc, Rj, 0, client,
+ userfunc, prefix,
+ dfa, max_pos, init_pos);
+ else
+ i = 0;
+ yaz_log (YLOG_DEBUG, "max_pos = %d", *max_pos);
+ dfa_delete (&dfa);
xfree (Rj);
rm_MatchContext (&mc);
return i;
}
+
+void dict_grep_cmap (Dict dict, void *vp,
+ const char **(*cmap)(void *vp, const char **from, int len))
+{
+ dict->grep_cmap = cmap;
+ dict->grep_cmap_data = vp;
+}
+/*
+ * Local variables:
+ * c-basic-offset: 4
+ * indent-tabs-mode: nil
+ * End:
+ * vim: shiftwidth=4 tabstop=8 expandtab
+ */
+