X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=recctrl%2Frecgrs.c;h=644cf3f41341521c7b7c58afde44bd95cf62dbeb;hb=f964910eb92b5671e5d999aab72301457413c89c;hp=b832911adf338ed90708317d03f93ccbadbc7771;hpb=a749ce4ff4739420474e7d40f400321a86f67684;p=idzebra-moved-to-github.git diff --git a/recctrl/recgrs.c b/recctrl/recgrs.c index b832911..644cf3f 100644 --- a/recctrl/recgrs.c +++ b/recctrl/recgrs.c @@ -1,9 +1,24 @@ -/* - * Copyright (C) 1994-2002, Index Data - * All rights reserved. - * - * $Id: recgrs.c,v 1.54 2002-07-05 16:07:02 adam Exp $ - */ +/* $Id: recgrs.c,v 1.66 2002-09-02 11:19:33 adam Exp $ + Copyright (C) 1995,1996,1997,1998,1999,2000,2001,2002 + 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 @@ -84,7 +99,7 @@ static void *grs_init(RecType recType) grs_add_handler (h, recTypeGrs_tcl); #endif grs_add_handler (h, recTypeGrs_marc); -#if YAZ_HAVE_EXPAT +#if HAVE_EXPAT_H grs_add_handler (h, recTypeGrs_xml); #endif return h; @@ -105,6 +120,16 @@ static void grs_destroy(void *clientData) xfree (h); } +/* use + 1 start element (tag) + 2 end element + 3 start attr (and attr-exact) + 4 end attr + + 1016 cdata + 1015 attr data +*/ + static void index_xpath (data1_node *n, struct recExtractCtrl *p, int level, RecWord *wrd, int use) { @@ -164,9 +189,70 @@ static void index_xpath (data1_node *n, struct recExtractCtrl *p, } else { - (*p->tokenAdd)(wrd); + data1_xattr *xp; + (*p->tokenAdd)(wrd); /* index element pag (AKA tag path) */ + if (use == 1) + { + for (xp = n->u.tag.attributes; xp; xp = xp->next) + { + char comb[512]; + /* attribute (no value) */ + wrd->reg_type = '0'; + wrd->attrUse = 3; + wrd->string = xp->name; + wrd->length = strlen(xp->name); + + 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->attrUse = 3; + wrd->reg_type = '0'; + wrd->string = comb; + wrd->length = strlen(comb); + wrd->seqno--; + + (*p->tokenAdd)(wrd); + } + } + for (xp = n->u.tag.attributes; xp; xp = xp->next) + { + char attr_tag_path_full[1024]; + int int_len = flen; + + sprintf (attr_tag_path_full, "@%s/%.*s", + xp->name, int_len, tag_path_full); + + wrd->reg_type = '0'; + wrd->attrUse = 1; + wrd->string = attr_tag_path_full; + wrd->length = strlen(attr_tag_path_full); + (*p->tokenAdd)(wrd); + + if (xp->value) + { + wrd->attrUse = 1015; + wrd->reg_type = 'w'; + wrd->string = xp->value; + wrd->length = strlen(xp->value); + (*p->tokenAdd)(wrd); + } + + wrd->reg_type = '0'; + wrd->attrUse = 2; + wrd->string = attr_tag_path_full; + wrd->length = strlen(attr_tag_path_full); + (*p->tokenAdd)(wrd); + } + } } - break; } } @@ -293,8 +379,12 @@ static int dumpkeys(data1_node *n, struct recExtractCtrl *p, int level, { index_termlist (n, n, p, level, wrd); /* index start tag */ + assert (n->root->u.root.absyn); + if (!n->root->u.root.absyn) index_xpath (n, p, level, wrd, 1); + else if (n->root->u.root.absyn->enable_xpath_indexing) + index_xpath (n, p, level, wrd, 1); } if (n->child) @@ -310,8 +400,8 @@ static int dumpkeys(data1_node *n, struct recExtractCtrl *p, int level, { printf("%*s", level * 4, ""); printf("Data: "); - if (n->u.data.len > 32) - printf("'%.24s ... %.6s'\n", n->u.data.data, + if (n->u.data.len > 256) + printf("'%.240s ... %.6s'\n", n->u.data.data, n->u.data.data + n->u.data.len-6); else if (n->u.data.len > 0) printf("'%.*s'\n", n->u.data.len, n->u.data.data); @@ -323,7 +413,8 @@ static int dumpkeys(data1_node *n, struct recExtractCtrl *p, int level, index_termlist (par, n, p, level, wrd); if (!n->root->u.root.absyn) index_xpath (n, p, level, wrd, 1016); - + else if (n->root->u.root.absyn->enable_xpath_indexing) + index_xpath (n, p, level, wrd, 1016); } if (n->which == DATA1N_tag) @@ -331,9 +422,10 @@ static int dumpkeys(data1_node *n, struct recExtractCtrl *p, int level, /* index end tag */ if (!n->root->u.root.absyn) index_xpath (n, p, level, wrd, 2); + else if (n->root->u.root.absyn->enable_xpath_indexing) + index_xpath (n, p, level, wrd, 2); } - if (p->flagShowRecords && n->which == DATA1N_root) { printf("%*s-------------\n\n", level * 4, ""); @@ -381,7 +473,7 @@ static int grs_extract_sub(struct grs_handlers *h, struct recExtractCtrl *p, gri.dh = p->dh; if (read_grs_type (h, &gri, p->subType, &n)) - return RECCTRL_EXTRACT_ERROR; + return RECCTRL_EXTRACT_ERROR_NO_SUCH_FILTER; if (!n) return RECCTRL_EXTRACT_EOF; oe.proto = PROTO_Z3950; @@ -396,16 +488,19 @@ static int grs_extract_sub(struct grs_handlers *h, struct recExtractCtrl *p, if ((oid_ent_to_oid (&oe, oidtmp))) (*p->schemaAdd)(p, oidtmp); } + + /* ensure our data1 tree is UTF-8 */ + data1_iconv (p->dh, mem, n, "UTF-8", data1_get_encoding(p->dh, n)); + #if 0 data1_pr_tree (p->dh, n, stdout); #endif - data1_iconv (p->dh, mem, n, "ISO-8859-1", "UTF-8"); (*p->init)(p, &wrd); if (dumpkeys(n, p, 0, &wrd) < 0) { data1_free_tree(p->dh, n); - return RECCTRL_EXTRACT_ERROR; + return RECCTRL_EXTRACT_ERROR_GENERIC; } data1_free_tree(p->dh, n); return RECCTRL_EXTRACT_OK; @@ -497,6 +592,17 @@ static int process_comp(data1_handle dh, data1_node *n, Z_RecordComposition *c) } } +static void add_nice_whitespace (struct recRetrieveCtrl *p, data1_node *top, + NMEM mem) +{ + data1_node *n = top->child; + while (n && n->which == DATA1N_data && n->u.data.what == DATA1I_text) + { + data1_mk_text_n(p->dh, mem, n->u.data.data, n->u.data.len, top); + n = n->next; + } +} + static void add_idzebra_info (struct recRetrieveCtrl *p, data1_node *top, NMEM mem) { @@ -508,17 +614,24 @@ static void add_idzebra_info (struct recRetrieveCtrl *p, data1_node *top, data1_tag_add_attr (p->dh, mem, top, idzebra_ns); + add_nice_whitespace (p, top, mem); data1_mk_tag_data_int (p->dh, top, "idzebra:size", p->recordSize, mem); if (p->score != -1) + { + add_nice_whitespace (p, top, mem); data1_mk_tag_data_int (p->dh, top, "idzebra:score", p->score, mem); - + } + add_nice_whitespace (p, top, mem); data1_mk_tag_data_int (p->dh, top, "idzebra:localnumber", p->localno, mem); if (p->fname) + { + add_nice_whitespace (p, top, mem); data1_mk_tag_data_text(p->dh, top, "idzebra:filename", p->fname, mem); + } } static int grs_retrieve(void *clientData, struct recRetrieveCtrl *p) @@ -558,6 +671,9 @@ static int grs_retrieve(void *clientData, struct recRetrieveCtrl *p) nmem_destroy (mem); return 0; } + /* ensure our data1 tree is UTF-8 */ + data1_iconv (p->dh, mem, node, "UTF-8", data1_get_encoding(p->dh, node)); + #if 0 data1_pr_tree (p->dh, node, stdout); #endif @@ -711,17 +827,16 @@ static int grs_retrieve(void *clientData, struct recRetrieveCtrl *p) #if 0 data1_pr_tree (p->dh, node, stdout); #endif -#if 1 - data1_iconv (p->dh, mem, node, "ISO-8859-1", "UTF-8"); -#endif logf (LOG_DEBUG, "grs_retrieve: transfer syntax mapping"); switch (p->output_format = (p->input_format != VAL_NONE ? p->input_format : VAL_SUTRS)) { - case VAL_TEXT_XML: add_idzebra_info (p, top, mem); + if (p->encoding) + data1_iconv (p->dh, mem, node, p->encoding, "UTF-8"); + if (!(p->rec_buf = data1_nodetoidsgml(p->dh, node, selected, &p->rec_len))) p->diagnostic = 238; @@ -755,6 +870,8 @@ static int grs_retrieve(void *clientData, struct recRetrieveCtrl *p) p->rec_len = (size_t) (-1); break; case VAL_SUTRS: + if (p->encoding) + data1_iconv (p->dh, mem, node, p->encoding, "UTF-8"); if (!(p->rec_buf = data1_nodetobuf(p->dh, node, selected, &p->rec_len))) p->diagnostic = 238; @@ -791,6 +908,8 @@ static int grs_retrieve(void *clientData, struct recRetrieveCtrl *p) p->diagnostic = 238; break; } + if (p->encoding) + data1_iconv (p->dh, mem, node, p->encoding, "UTF-8"); if (!(p->rec_buf = data1_nodetomarc(p->dh, marctab, node, selected, &p->rec_len))) p->diagnostic = 238;