X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=dict%2Flookgrep.c;h=03193ed839155bd65f7e8288525e842b2e1c9fd3;hb=07daef4fb241a9106fb995bb87a823708babcb10;hp=2fd3533200d612af0f510c533923c035275aa1cf;hpb=b31a8654c5b7a597dd4a351f452ac91e5ef60798;p=idzebra-moved-to-github.git diff --git a/dict/lookgrep.c b/dict/lookgrep.c index 2fd3533..03193ed 100644 --- a/dict/lookgrep.c +++ b/dict/lookgrep.c @@ -1,55 +1,26 @@ -/* - * Copyright (C) 1994, Index Data I/S - * All rights reserved. - * Sebastian Hammer, Adam Dickmeiss - * - * $Log: lookgrep.c,v $ - * Revision 1.14 1996-02-02 13:43:51 adam - * The public functions simply use char instead of Dict_char to represent - * search strings. Dict_char is used internally only. - * - * Revision 1.13 1996/01/08 09:09:30 adam - * Function dfa_parse got 'const' string argument. - * - * Revision 1.12 1995/12/11 09:04:48 adam - * Bug fix: the lookup/scan/lookgrep didn't handle empty dictionary. - * - * Revision 1.11 1995/12/06 14:43:02 adam - * New function: dict_delete. - * - * Revision 1.10 1995/11/16 17:00:44 adam - * Changed stupid log. - * - * Revision 1.9 1995/10/27 13:58:09 adam - * Makes 'Database unavailable' diagnostic. - * - * Revision 1.8 1995/10/19 14:57:21 adam - * New feature: grep lookup saves length of longest prefix match. - * - * Revision 1.7 1995/10/17 18:01:22 adam - * Userfunc may return non-zero in which case the the grepping stops - * immediately. - * - * Revision 1.6 1995/10/09 16:18:32 adam - * Function dict_lookup_grep got extra client data parameter. - * - * Revision 1.5 1995/09/14 11:52:59 adam - * Grep handle function parameter info is const now. - * - * Revision 1.4 1995/01/24 16:01:02 adam - * Added -ansi to CFLAGS. - * Use new API of dfa module. - * - * 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 1.30 2005-11-01 10:16:08 adam Exp $ + Copyright (C) 1995-2005 + Index Data ApS + +This file is part of the Zebra server. + +Zebra is free software; you can redistribute it and/or modify it under +the terms of the GNU General Public License as published by the Free +Software Foundation; either version 2, or (at your option) any later +version. + +Zebra is distributed in the hope that it will be useful, but WITHOUT ANY +WARRANTY; without even the implied warranty of MERCHANTABILITY or +FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License +for more details. + +You should have received a copy of the GNU General Public License +along with Zebra; see the file LICENSE.zebra. If not, write to the +Free Software Foundation, 59 Temple Place - Suite 330, Boston, MA +02111-1307, USA. +*/ + + #include #include @@ -57,12 +28,22 @@ #include #include -#include +#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 */ @@ -91,13 +72,13 @@ static INLINE MatchWord get_bit (MatchContext *mc, MatchWord *m, int ch, static MatchContext *mk_MatchContext (struct DFA *dfa, int range) { - MatchContext *mc = xmalloc (sizeof(*mc)); + MatchContext *mc = (MatchContext *) xmalloc (sizeof(*mc)); int s; 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; sno_states; s++) if (dfa->states[s]->rule_no) @@ -232,7 +213,8 @@ static void or (MatchContext *mc, MatchWord *Rdst, } static INLINE int move (MatchContext *mc, MatchWord *Rj1, MatchWord *Rj, - Dict_char ch, struct DFA *dfa, MatchWord *Rtmp) + Dict_char ch, struct DFA *dfa, MatchWord *Rtmp, + int range) { int d; MatchWord *Rtmp_2 = Rtmp + mc->n; @@ -263,7 +245,7 @@ static int dict_grep (Dict dict, Dict_ptr ptr, MatchContext *mc, MatchWord *Rj, int pos, void *client, int (*userfunc)(char *, const char *, void *), Dict_char *prefix, struct DFA *dfa, - int *max_pos) + int *max_pos, int init_pos) { int lo, hi, d; void *p; @@ -273,7 +255,7 @@ static int dict_grep (Dict dict, Dict_ptr ptr, MatchContext *mc, 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) { @@ -291,6 +273,7 @@ static int dict_grep (Dict dict, Dict_ptr ptr, MatchContext *mc, 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; @@ -304,15 +287,19 @@ static int dict_grep (Dict dict, Dict_ptr ptr, MatchContext *mc, return 1; break; } - move (mc, Rj1, Rj0, ch, dfa, 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; @@ -324,6 +311,7 @@ static int dict_grep (Dict dict, Dict_ptr ptr, MatchContext *mc, MatchWord *Rj1 = Rj+ mc->fact; MatchWord *Rj_tmp = Rj+2*mc->fact; Dict_char ch; + int range; /* Dict_ptr subptr */ /* Dict_char sub char */ @@ -335,9 +323,13 @@ static int dict_grep (Dict dict, Dict_ptr ptr, MatchContext *mc, if (pos > *max_pos) *max_pos = pos; - move (mc, Rj1, Rj, ch, dfa, Rj_tmp); + 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) { @@ -345,7 +337,7 @@ static int dict_grep (Dict dict, Dict_ptr ptr, MatchContext *mc, 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; if ((*userfunc)((char*) prefix, @@ -359,11 +351,11 @@ static int dict_grep (Dict dict, Dict_ptr ptr, MatchContext *mc, if (subptr) { if (dict_grep (dict, subptr, mc, Rj1, pos+1, - client, userfunc, prefix, dfa, max_pos)) + 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)); } } } @@ -373,7 +365,7 @@ static int dict_grep (Dict dict, Dict_ptr ptr, MatchContext *mc, } int dict_lookup_grep (Dict dict, const char *pattern, int range, void *client, - int *max_pos, + int *max_pos, int init_pos, int (*userfunc)(char *name, const char *info, void *client)) { @@ -384,10 +376,26 @@ int dict_lookup_grep (Dict dict, const char *pattern, int range, void *client, struct DFA *dfa = dfa_init(); int i, d; - logf (LOG_DEBUG, "dict_lookup_grep '%s' range=%d", pattern, range); +#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) { + yaz_log (YLOG_WARN, "dfa_parse fail=%d", i); dfa_delete (&dfa); return -1; } @@ -395,7 +403,7 @@ int dict_lookup_grep (Dict dict, const char *pattern, int range, void *client, 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++) @@ -414,14 +422,22 @@ int dict_lookup_grep (Dict dict, const char *pattern, int range, void *client, } } *max_pos = 0; - if (dict->head.last > 1) - i = dict_grep (dict, 1, mc, Rj, 0, client, userfunc, prefix, - dfa, max_pos); + 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; - logf (LOG_DEBUG, "max_pos = %d", *max_pos); + 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; +}