X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=index%2Frecgrs.c;h=46805ab947f2a66b92a04e11caa4f52aef86ba10;hb=b04a1cdb1a0345252a78caa8f7731089df314ad0;hp=4d621efddbfad14eac5cd30373a21f17584d0b06;hpb=f5fb1bd0a0df8a28471285cdf622897f4c2391b3;p=idzebra-moved-to-github.git diff --git a/index/recgrs.c b/index/recgrs.c index 4d621ef..46805ab 100644 --- a/index/recgrs.c +++ b/index/recgrs.c @@ -1,4 +1,4 @@ -/* $Id: recgrs.c,v 1.7 2006-09-29 10:02:47 adam Exp $ +/* $Id: recgrs.c,v 1.12 2006-12-22 13:57:27 adam Exp $ Copyright (C) 1995-2006 Index Data ApS @@ -245,7 +245,7 @@ static int sp_expr(struct source_parser *sp, data1_node *n, RecWord *wrd) return 1; } -static struct source_parser *source_parser_create() +static struct source_parser *source_parser_create(void) { struct source_parser *sp = xmalloc(sizeof(*sp)); @@ -295,8 +295,6 @@ int d1_check_xpath_predicate(data1_node *n, struct xpath_predicate *p) res = 0; /* looking for the attribute with a specified name */ for (attr = n->u.tag.attributes; attr; attr = attr->next) { - yaz_log(YLOG_DEBUG," - attribute %s <-> %s", attname, attr->name ); - if (!strcmp(attr->name, attname)) { if (p->u.relation.op[0]) { if (*p->u.relation.op != '=') { @@ -305,8 +303,6 @@ int d1_check_xpath_predicate(data1_node *n, struct xpath_predicate *p) yaz_log(YLOG_WARN, "predicate %s ignored", p->u.relation.name); res = 1; break; } else { - yaz_log(YLOG_DEBUG," - value %s <-> %s", - p->u.relation.value, attr->value ); if (!strcmp(attr->value, p->u.relation.value)) { res = 1; break; } @@ -317,7 +313,6 @@ int d1_check_xpath_predicate(data1_node *n, struct xpath_predicate *p) } } } - yaz_log(YLOG_DEBUG, "return %d", res); return res; } else { return 1; @@ -436,12 +431,8 @@ data1_termlist *xpath_termlist_by_tagpath(char *tagpath, data1_node *n) backwards trough xpath location steps ... */ for (i = xpe->xpath_len - 1; i>0; i--) { - yaz_log(YLOG_DEBUG, "Checking step %d: %s on tag %s", - i, xp[i].part, nn->u.tag.tag); - if (!d1_check_xpath_predicate(nn, xp[i].predicate)) { - yaz_log(YLOG_DEBUG, " Predicates didn't match"); ok = 0; break; } @@ -534,6 +525,25 @@ static void mk_tag_path_full(char *tag_path_full, size_t max, data1_node *n) } +static void index_staticrank(struct recExtractCtrl *p, + RecWord *wrd, + data1_absyn *absyn) +{ + const char *staticrank_index = data1_absyn_get_staticrank(absyn); + + if (staticrank_index && !strcmp(wrd->index_name, staticrank_index)) + { + char valz[20]; + size_t len = wrd->term_len; + + if (len > sizeof(valz)-1) + len = sizeof(valz)-1; + memcpy(valz, wrd->term_buf, len); + valz[len] = '\0'; + p->staticrank = atozint(valz); + } +} + static void index_xpath(struct source_parser *sp, data1_node *n, struct recExtractCtrl *p, int level, RecWord *wrd, @@ -546,11 +556,15 @@ static void index_xpath(struct source_parser *sp, data1_node *n, int termlist_only = 1; data1_termlist *tl; int xpdone = 0; - if ((!n->root->u.root.absyn) || - (n->root->u.root.absyn->xpath_indexing == DATA1_XPATH_INDEXING_ENABLE)) { + + if (!n->root->u.root.absyn + || + n->root->u.root.absyn->xpath_indexing == DATA1_XPATH_INDEXING_ENABLE) + { termlist_only = 0; } + switch (n->which) { case DATA1N_data: @@ -564,6 +578,7 @@ static void index_xpath(struct source_parser *sp, data1_node *n, if (n->root->u.root.absyn && (tl = xpath_termlist_by_tagpath(tag_path_full, n))) { + zint max_seqno = 0; for (; tl; tl = tl->next) { /* need to copy recword because it may be changed */ @@ -572,46 +587,34 @@ static void index_xpath(struct source_parser *sp, data1_node *n, memcpy (&wrd_tl, wrd, sizeof(*wrd)); if (tl->source) sp_parse(sp, n, &wrd_tl, tl->source); - if (!tl->index_name) - { - /* this is the ! case, so structure is for the xpath index */ - wrd_tl.index_name = xpath_index; - if (p->flagShowRecords) - { - int i; - printf("%*sXPath index", (level + 1) * 4, ""); - printf (" XData:\""); - for (i = 0; i 40) - printf (" ..."); - fputc ('\n', stdout); - } - else - (*p->tokenAdd)(&wrd_tl); - xpdone = 1; - } else { - /* this is just the old fashioned attribute based index */ - wrd_tl.index_name = tl->index_name; - if (p->flagShowRecords) - { - int i; - printf("%*sIdx: [%s]", (level + 1) * 4, "", - tl->structure); - printf("%s %s", tl->index_name, tl->source); - printf (" XData:\""); - for (i = 0; i 40) - printf (" ..."); - fputc ('\n', stdout); - } - else - (*p->tokenAdd)(&wrd_tl); - } + + /* this is just the old fashioned attribute based index */ + wrd_tl.index_name = tl->index_name; + if (p->flagShowRecords) + { + int i; + printf("%*sIdx: [%s]", (level + 1) * 4, "", + tl->structure); + printf("%s %s", tl->index_name, tl->source); + printf (" XData:\""); + for (i = 0; i 40) + printf (" ..."); + fputc ('\n', stdout); + } + else + { + (*p->tokenAdd)(&wrd_tl); + index_staticrank(p, &wrd_tl, n->root->u.root.absyn); + } + if (wrd_tl.seqno > max_seqno) + max_seqno = wrd_tl.seqno; } + if (max_seqno) + wrd->seqno = max_seqno; + } /* xpath indexing is done, if there was no termlist given, or no ! in the termlist, and default indexing is enabled... */ @@ -623,6 +626,8 @@ static void index_xpath(struct source_parser *sp, data1_node *n, } break; case DATA1N_tag: + if (termlist_only) + return; mk_tag_path_full(tag_path_full, sizeof(tag_path_full), n); wrd->index_type = '0'; @@ -641,27 +646,8 @@ static void index_xpath(struct source_parser *sp, data1_node *n, else { data1_xattr *xp; - data1_termlist *tl; - int do_xpindex; - - /* Add tag start/end xpath index, only when there is a ! in - the apropriate xelm directive, or default xpath indexing - is enabled - */ - if (!(do_xpindex = 1 - termlist_only)) - { - if ((tl = xpath_termlist_by_tagpath(tag_path_full, n))) - { - for (; tl; tl = tl->next) - { - if (!tl->index_name) - do_xpindex = 1; - } - } - } - if (do_xpindex) { - (*p->tokenAdd)(wrd); /* index element pag (AKA tag path) */ - } + + (*p->tokenAdd)(wrd); /* index element pag (AKA tag path) */ if (xpath_is_start == 1) /* only for the starting tag... */ { @@ -669,17 +655,8 @@ static void index_xpath(struct source_parser *sp, data1_node *n, data1_termlist *tll[MAX_ATTR_COUNT]; int i = 0; - - /* get termlists for attributes, and find out, if we have to do xpath indexing */ - for (xp = n->u.tag.attributes; xp; xp = xp->next) { - i++; - } - - i = 0; for (xp = n->u.tag.attributes; xp; xp = xp->next) { char comb[512]; - int do_xpindex = 1 - termlist_only; - data1_termlist *tl; char attr_tag_path_full[1024]; /* this could be cached as well */ @@ -688,44 +665,31 @@ static void index_xpath(struct source_parser *sp, data1_node *n, tll[i] = xpath_termlist_by_tagpath(attr_tag_path_full,n); - /* if there is a ! in the xelm termlist, or default indexing is on, - proceed with xpath idx */ - if ((tl = tll[i])) - { - for (; tl; tl = tl->next) - { - if (!tl->index_name) - do_xpindex = 1; - } - } + /* attribute (no value) */ + wrd->index_type = '0'; + wrd->index_name = ZEBRA_XPATH_ATTR_NAME; + wrd->term_buf = xp->name; + wrd->term_len = strlen(xp->name); - if (do_xpindex) { + wrd->seqno--; + (*p->tokenAdd)(wrd); + + if (xp->value + && + strlen(xp->name) + strlen(xp->value) < sizeof(comb)-2) + { + /* attribute value exact */ + strcpy (comb, xp->name); + strcat (comb, "="); + strcat (comb, xp->value); - /* attribute (no value) */ + wrd->index_name = ZEBRA_XPATH_ATTR_NAME; wrd->index_type = '0'; - wrd->index_name = ZEBRA_XPATH_ATTR_NAME; - wrd->term_buf = xp->name; - wrd->term_len = strlen(xp->name); - + wrd->term_buf = comb; + wrd->term_len = strlen(comb); wrd->seqno--; - (*p->tokenAdd)(wrd); - if (xp->value && - strlen(xp->name) + strlen(xp->value) < sizeof(comb)-2) { - - /* attribute value exact */ - strcpy (comb, xp->name); - strcat (comb, "="); - strcat (comb, xp->value); - - wrd->index_name = ZEBRA_XPATH_ATTR_NAME; - wrd->index_type = '0'; - wrd->term_buf = comb; - wrd->term_len = strlen(comb); - wrd->seqno--; - - (*p->tokenAdd)(wrd); - } + (*p->tokenAdd)(wrd); } i++; } @@ -760,6 +724,8 @@ static void index_xpath(struct source_parser *sp, data1_node *n, wrd->term_buf = xp->value; wrd->term_len = strlen(xp->value); (*p->tokenAdd)(wrd); + index_staticrank(p, wrd, + n->root->u.root.absyn); } } } @@ -827,6 +793,7 @@ static void index_termlist (struct source_parser *sp, data1_node *par, { wrd->index_type = *tlist->structure; wrd->index_name = tlist->index_name; + index_staticrank(p, wrd, n->root->u.root.absyn); (*p->tokenAdd)(wrd); } }