X-Git-Url: http://git.indexdata.com/?p=idzebra-moved-to-github.git;a=blobdiff_plain;f=index%2Fmod_dom.c;h=fd5986669becc8c40b9135c4816f064388cfce66;hp=1cbe3dc96c161bbc385cfaf569dc47d52da1debd;hb=a66b7d79383ae700f3358731eecfe2aafed0e90d;hpb=784bb11a4ee525eb88c78c3fc9e08a8b5c2e9939 diff --git a/index/mod_dom.c b/index/mod_dom.c index 1cbe3dc..fd59866 100644 --- a/index/mod_dom.c +++ b/index/mod_dom.c @@ -1,5 +1,5 @@ /* This file is part of the Zebra server. - Copyright (C) 1995-2008 Index Data + Copyright (C) 1994-2010 Index Data 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 @@ -50,11 +50,26 @@ static const char *zebra_dom_ns = ZEBRA_DOM_NS; #define ZEBRA_PI_NAME "zebra-2.0" static const char *zebra_pi_name = ZEBRA_PI_NAME; +enum convert_type { + convert_xslt_type, + convert_meta_type +}; - -struct convert_s { +struct convert_xslt { const char *stylesheet; xsltStylesheetPtr stylesheet_xsp; +}; + +struct convert_meta { + int dummy; +}; + +struct convert_s { + enum convert_type which; + union { + struct convert_xslt xslt; + struct convert_meta meta; + } u; struct convert_s *next; }; @@ -205,10 +220,13 @@ static int attr_content(struct _xmlAttr *attr, const char *name, static void destroy_xsp(struct convert_s *c) { - while(c) + while (c) { - if (c->stylesheet_xsp) - xsltFreeStylesheet(c->stylesheet_xsp); + if (c->which == convert_xslt_type) + { + if (c->u.xslt.stylesheet_xsp) + xsltFreeStylesheet(c->u.xslt.stylesheet_xsp); + } c = c->next; } } @@ -269,25 +287,25 @@ static ZEBRA_RES parse_convert(struct filter_info *tinfo, xmlNodePtr ptr, if (!XML_STRCMP(ptr->name, "xslt")) { struct _xmlAttr *attr; - struct convert_s *p - = nmem_malloc(tinfo->nmem_config, sizeof(*p)); + struct convert_s *p = nmem_malloc(tinfo->nmem_config, sizeof(*p)); p->next = 0; - p->stylesheet = 0; - p->stylesheet_xsp = 0; + p->which = convert_xslt_type; + p->u.xslt.stylesheet = 0; + p->u.xslt.stylesheet_xsp = 0; for (attr = ptr->properties; attr; attr = attr->next) - if (attr_content(attr, "stylesheet", &p->stylesheet)) + if (attr_content(attr, "stylesheet", &p->u.xslt.stylesheet)) ; else { dom_log(YLOG_WARN, tinfo, ptr, "bad attribute @%s", attr->name); } - if (p->stylesheet) + if (p->u.xslt.stylesheet) { char tmp_xslt_full_name[1024]; - if (!yaz_filepath_resolve(p->stylesheet, + if (!yaz_filepath_resolve(p->u.xslt.stylesheet, tinfo->profile_path, NULL, tmp_xslt_full_name)) @@ -295,30 +313,44 @@ static ZEBRA_RES parse_convert(struct filter_info *tinfo, xmlNodePtr ptr, dom_log(YLOG_WARN, tinfo, 0, "stylesheet %s not found in " "path %s", - p->stylesheet, + p->u.xslt.stylesheet, tinfo->profile_path); return ZEBRA_FAIL; } - p->stylesheet_xsp + p->u.xslt.stylesheet_xsp = xsltParseStylesheetFile((const xmlChar*) tmp_xslt_full_name); - if (!p->stylesheet_xsp) + if (!p->u.xslt.stylesheet_xsp) { dom_log(YLOG_WARN, tinfo, 0, "could not parse xslt stylesheet %s", tmp_xslt_full_name); return ZEBRA_FAIL; } - } - else - { - dom_log(YLOG_WARN, tinfo, ptr, - "missing attribute 'stylesheet' "); - return ZEBRA_FAIL; - } - *l = p; - l = &p->next; + } + else + { + dom_log(YLOG_WARN, tinfo, ptr, + "missing attribute 'stylesheet'"); + return ZEBRA_FAIL; + } + *l = p; + l = &p->next; + } + else if (!XML_STRCMP(ptr->name, "process-meta")) + { + struct _xmlAttr *attr; + struct convert_s *p = nmem_malloc(tinfo->nmem_config, sizeof(*p)); + + p->next = 0; + p->which = convert_meta_type; + + for (attr = ptr->properties; attr; attr = attr->next) + dom_log(YLOG_WARN, tinfo, ptr, + "bad attribute @%s", attr->name); + *l = p; + l = &p->next; } else { @@ -330,8 +362,64 @@ static ZEBRA_RES parse_convert(struct filter_info *tinfo, xmlNodePtr ptr, return ZEBRA_OK; } +static int process_meta(struct filter_info *tinfo, xmlDocPtr doc, xmlNodePtr node, + struct recRetrieveCtrl *retctr) +{ + + if (node->type == XML_ELEMENT_NODE && node->ns && node->ns->href && + 0 == XML_STRCMP(node->ns->href, zebra_dom_ns)) + { + if (0 == XML_STRCMP(node->name, "meta")) + { + const char *element_set_name = 0; + + struct _xmlAttr *attr; + for (attr = node->properties; attr; attr = attr->next) + { + if (attr_content(attr, "name", &element_set_name)) + ; + else + { + dom_log(YLOG_WARN, tinfo, node, + "bad attribute @%s, expected @name", attr->name); + } + } + if (element_set_name) + { + WRBUF result = wrbuf_alloc(); + WRBUF addinfo = wrbuf_alloc(); + const Odr_oid *input_format = yaz_oid_recsyn_xml; + const Odr_oid *output_format = 0; + int ret; + + ret = retctr->special_fetch(retctr->handle, + element_set_name, + input_format, &output_format, + result, addinfo); + if (ret == 0) + { + xmlDocPtr sub_doc = + xmlParseMemory(wrbuf_buf(result), wrbuf_len(result)); + if (sub_doc) + { + xmlNodePtr t = xmlDocGetRootElement(sub_doc); + xmlReplaceNode(node, xmlCopyNode(t, 1)); + xmlFreeDoc(sub_doc); + } + } + wrbuf_destroy(result); + wrbuf_destroy(addinfo); + } + } + } + for (node = node->children; node; node = node->next) + process_meta(tinfo, doc, node, retctr); + return 0; +} + static ZEBRA_RES perform_convert(struct filter_info *tinfo, struct recExtractCtrl *extctr, + struct recRetrieveCtrl *retctr, struct convert_s *convert, const char **params, xmlDocPtr *doc, @@ -339,34 +427,48 @@ static ZEBRA_RES perform_convert(struct filter_info *tinfo, { for (; convert; convert = convert->next) { - xmlChar *buf_out = 0; - int len_out = 0; - xmlDocPtr res_doc = xsltApplyStylesheet(convert->stylesheet_xsp, - *doc, params); - if (last_xsp) - *last_xsp = convert->stylesheet_xsp; - - if (!res_doc) - break; - - /* now saving into buffer and re-reading into DOM to avoid annoing - XSLT problem with thrown-out indentation text nodes */ - xsltSaveResultToString(&buf_out, &len_out, res_doc, - convert->stylesheet_xsp); - xmlFreeDoc(res_doc); - - xmlFreeDoc(*doc); - - *doc = xmlParseMemory((const char *) buf_out, len_out); + if (convert->which == convert_xslt_type) + { + xmlChar *buf_out = 0; + int len_out = 0; + xmlDocPtr res_doc = xsltApplyStylesheet(convert->u.xslt.stylesheet_xsp, + *doc, params); + if (last_xsp) + *last_xsp = convert->u.xslt.stylesheet_xsp; + + if (!res_doc) + break; + + /* now saving into buffer and re-reading into DOM to avoid annoing + XSLT problem with thrown-out indentation text nodes */ + xsltSaveResultToString(&buf_out, &len_out, res_doc, + convert->u.xslt.stylesheet_xsp); + xmlFreeDoc(res_doc); + + xmlFreeDoc(*doc); + + *doc = xmlParseMemory((const char *) buf_out, len_out); + + /* writing debug info out */ + if (extctr && extctr->flagShowRecords) + yaz_log(YLOG_LOG, "%s: XSLT %s\n %.*s", + tinfo->fname ? tinfo->fname : "(none)", + convert->u.xslt.stylesheet, + len_out, buf_out); + + xmlFree(buf_out); + } + else if (convert->which == convert_meta_type) + { + if (retctr) /* only execute meta on retrieval */ + { + process_meta(tinfo, *doc, xmlDocGetRootElement(*doc), retctr); - /* writing debug info out */ - if (extctr && extctr->flagShowRecords) - yaz_log(YLOG_LOG, "%s: XSLT %s\n %.*s", - tinfo->fname ? tinfo->fname : "(none)", - convert->stylesheet, - len_out, buf_out); - - xmlFree(buf_out); + /* last stylesheet absent */ + if (last_xsp) + *last_xsp = 0; + } + } } return ZEBRA_OK; } @@ -699,19 +801,6 @@ static int ioclose_ex(void *context) } -/* DOM filter style indexing */ -static int attr_content_xml(struct _xmlAttr *attr, const char *name, - const char **dst_content) -{ - if (0 == XML_STRCMP(attr->name, name) && attr->children - && attr->children->type == XML_TEXT_NODE) - { - *dst_content = (const char *) (attr->children->content); - return 1; - } - return 0; -} - /* DOM filter style indexing */ static void index_value_of(struct filter_info *tinfo, @@ -724,10 +813,18 @@ static void index_value_of(struct filter_info *tinfo, { xmlChar *text = xmlNodeGetContent(node); size_t text_len = strlen((const char *)text); - + /* if there is no text, we do not need to proceed */ if (text_len) { + /* keep seqno base so that all text will have + identical seqno's for multiple fields , e.g + .. */ + + zint seqno_base = recword->seqno; + zint seqno_max = recword->seqno; + + const char *look = index_p; const char *bval; const char *eval; @@ -775,6 +872,7 @@ static void index_value_of(struct filter_info *tinfo, /* actually indexing the text given */ + recword->seqno = seqno_base; recword->index_name = (const char *)index; if (*type) recword->index_type = (const char *) type; @@ -789,12 +887,16 @@ static void index_value_of(struct filter_info *tinfo, (extctr->tokenAdd)(recword); + if (seqno_max < recword->seqno) + seqno_max = recword->seqno; + /* eat whitespaces */ if (*look && ' ' == *look) { look++; } } + recword->seqno = seqno_max; } xmlFree(text); } @@ -819,7 +921,13 @@ static void set_record_info(struct filter_info *tinfo, if (id_p && *id_p) - sscanf((const char *)id_p, "%255s", extctr->match_criteria); + { + size_t l = strlen(id_p); + if (l >= sizeof(extctr->match_criteria)) + l = sizeof(extctr->match_criteria)-1; + memcpy(extctr->match_criteria, id_p, l); + extctr->match_criteria[l] = '\0'; + } if (rank_p && *rank_p) extctr->staticrank = atozint((const char *)rank_p); @@ -859,14 +967,14 @@ static void process_xml_element_zebra_node(struct filter_info *tinfo, if (node->type == XML_ELEMENT_NODE && node->ns && node->ns->href && 0 == XML_STRCMP(node->ns->href, zebra_dom_ns)) { - if (0 == XML_STRCMP(node->name, "index")) - { + if (0 == XML_STRCMP(node->name, "index")) + { const char *index_p = 0; struct _xmlAttr *attr; for (attr = node->properties; attr; attr = attr->next) { - if (attr_content_xml(attr, "name", &index_p)) + if (attr_content(attr, "name", &index_p)) { index_value_of(tinfo, extctr, recword, node, index_p); } @@ -887,11 +995,11 @@ static void process_xml_element_zebra_node(struct filter_info *tinfo, struct _xmlAttr *attr; for (attr = node->properties; attr; attr = attr->next) { - if (attr_content_xml(attr, "id", &id_p)) + if (attr_content(attr, "id", &id_p)) ; - else if (attr_content_xml(attr, "rank", &rank_p)) + else if (attr_content(attr, "rank", &rank_p)) ; - else if (attr_content_xml(attr, "type", &type_p)) + else if (attr_content(attr, "type", &type_p)) ; else { @@ -919,9 +1027,6 @@ static int attr_content_pi(const char **c_ptr, const char *name, const char *look = *c_ptr; int ret = 0; - *value = '\0'; - while (*look && ' ' == *look) - look++; if (strlen(look) > name_len) { if (look[name_len] == '=' && !memcmp(look, name, name_len)) @@ -938,8 +1043,6 @@ static int attr_content_pi(const char **c_ptr, const char *name, ret = 1; } } - while (*look && ' ' == *look) - look++; *c_ptr = look; return ret; } @@ -967,18 +1070,27 @@ static void process_xml_pi_node(struct filter_info *tinfo, *rank = '\0'; *type = '\0'; look += 6; - while (*look) + for (;;) + { + /* eat whitespace */ + while (' ' == *look) + look++; + if (*look == '\0') + break; if (attr_content_pi(&look, "id", id, sizeof(id))) ; else if (attr_content_pi(&look, "rank", rank, sizeof(rank))) ; else if (attr_content_pi(&look, "type", type, sizeof(type))) + ; + else { dom_log(YLOG_WARN, tinfo, node, "content '%s', can not parse '%s'", pi_p, look); break; } + } set_record_info(tinfo, extctr, node, id, rank, type); } /* parsing index instruction */ @@ -1052,19 +1164,15 @@ static void extract_dom_doc_node(struct filter_info *tinfo, } - - static int convert_extract_doc(struct filter_info *tinfo, struct filter_input *input, struct recExtractCtrl *p, xmlDocPtr doc) - { xmlChar *buf_out; int len_out; const char *params[10]; xsltStylesheetPtr last_xsp = 0; - xmlDocPtr store_doc = 0; /* per default do not ingest record */ tinfo->record_info_invoked = 0; @@ -1081,45 +1189,47 @@ static int convert_extract_doc(struct filter_info *tinfo, { xmlChar *buf_out; int len_out; + xmlDocDumpMemory(doc, &buf_out, &len_out); #if 0 FILE *outf = fopen("extract.xml", "w"); - xmlDocDumpMemory(doc, &buf_out, &len_out); fwrite(buf_out, 1, len_out, outf); -#endif - yaz_log(YLOG_LOG, "Extract Doc: %.*s", len_out, buf_out); -#if 0 fclose(outf); #endif + yaz_log(YLOG_LOG, "Extract Doc: %.*s", len_out, buf_out); } - /* input conversion */ - perform_convert(tinfo, p, input->convert, params, &doc, 0); - - - if (tinfo->store) + if (p->setStoreData) { - /* store conversion */ - store_doc = xmlCopyDoc(doc, 1); - perform_convert(tinfo, p, tinfo->store->convert, - params, &store_doc, &last_xsp); - } - - /* saving either store doc or original doc in case no store doc exists */ - if (last_xsp) - xsltSaveResultToString(&buf_out, &len_out, - store_doc ? store_doc : doc, last_xsp); - else - xmlDocDumpMemory(store_doc ? store_doc : doc, &buf_out, &len_out); + xmlDocPtr store_doc = 0; - if (p->setStoreData) - (*p->setStoreData)(p, buf_out, len_out); - xmlFree(buf_out); + /* input conversion */ + perform_convert(tinfo, p, 0, input->convert, params, &doc, 0); + + if (tinfo->store) + { + /* store conversion */ + store_doc = xmlCopyDoc(doc, 1); + perform_convert(tinfo, p, 0, tinfo->store->convert, + params, &store_doc, &last_xsp); + } + + /* saving either store doc or original doc in case no store doc exists */ + if (last_xsp) + xsltSaveResultToString(&buf_out, &len_out, + store_doc ? store_doc : doc, last_xsp); + else + xmlDocDumpMemory(store_doc ? store_doc : doc, &buf_out, &len_out); + + if (p->setStoreData) + (*p->setStoreData)(p, buf_out, len_out); + xmlFree(buf_out); + if (store_doc) + xmlFreeDoc(store_doc); + } - if (store_doc) - xmlFreeDoc(store_doc); /* extract conversion */ - perform_convert(tinfo, p, tinfo->extract->convert, params, &doc, 0); + perform_convert(tinfo, p, 0, tinfo->extract->convert, params, &doc, 0); /* finally, do the indexing */ @@ -1172,7 +1282,7 @@ static int extract_xml_split(struct filter_info *tinfo, ptr = xmlTextReaderExpand(input->u.xmlreader.reader); if (ptr) - { + { /* we have a new document */ xmlNodePtr ptr2 = xmlCopyNode(ptr, 1); @@ -1318,7 +1428,8 @@ static int filter_extract(void *clientData, struct recExtractCtrl *p) static int ioread_ret(void *context, char *buffer, int len) { struct recRetrieveCtrl *p = context; - return p->stream->readf(p->stream, buffer, len); + int r = p->stream->readf(p->stream, buffer, len); + return r; } static int ioclose_ret(void *context) @@ -1326,7 +1437,7 @@ static int ioclose_ret(void *context) return 0; } -static int filter_retrieve (void *clientData, struct recRetrieveCtrl *p) +static int filter_retrieve(void *clientData, struct recRetrieveCtrl *p) { /* const char *esn = zebra_dom_ns; */ const char *esn = 0; @@ -1356,7 +1467,7 @@ static int filter_retrieve (void *clientData, struct recRetrieveCtrl *p) { p->diagnostic = YAZ_BIB1_SPECIFIED_ELEMENT_SET_NAME_NOT_VALID_FOR_SPECIFIED_; - p->addinfo = odr_strdup(p->odr, esn); + p->addinfo = odr_strdup_null(p->odr, esn); return 0; } @@ -1392,7 +1503,7 @@ static int filter_retrieve (void *clientData, struct recRetrieveCtrl *p) } /* retrieve conversion */ - perform_convert(tinfo, 0, retrieve->convert, params, &doc, &last_xsp); + perform_convert(tinfo, 0, p, retrieve->convert, params, &doc, &last_xsp); if (!doc) { p->diagnostic = YAZ_BIB1_SYSTEM_ERROR_IN_PRESENTING_RECORDS; @@ -1463,6 +1574,7 @@ idzebra_filter /* * Local variables: * c-basic-offset: 4 + * c-file-style: "Stroustrup" * indent-tabs-mode: nil * End: * vim: shiftwidth=4 tabstop=8 expandtab