X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=src%2Flogic.c;h=a1de74816751790be1375517159873eb5c3475f1;hb=e0543b6a779ea1d96775fd4a901cf0ccf0c30ce0;hp=e8c2e8886afb190a77373c0905cc1c3e662bb24a;hpb=7db06f8379d23e8f1bce2c87f3ce082bb56ac2db;p=pazpar2-moved-to-github.git diff --git a/src/logic.c b/src/logic.c index e8c2e88..a1de748 100644 --- a/src/logic.c +++ b/src/logic.c @@ -1,23 +1,23 @@ -/* $Id: logic.c,v 1.64 2007-09-05 08:40:12 adam Exp $ +/* $Id: logic.c,v 1.67 2007-10-02 07:50:12 adam Exp $ Copyright (c) 2006-2007, Index Data. -This file is part of Pazpar2. + This file is part of Pazpar2. -Pazpar2 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. + Pazpar2 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. -Pazpar2 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. + Pazpar2 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 Pazpar2; see the file LICENSE. If not, write to the -Free Software Foundation, 59 Temple Place - Suite 330, Boston, MA -02111-1307, USA. - */ + You should have received a copy of the GNU General Public License + along with Pazpar2; see the file LICENSE. If not, write to the + Free Software Foundation, 59 Temple Place - Suite 330, Boston, MA + 02111-1307, USA. +*/ /** \file logic.c \brief high-level logic; mostly user sessions and settings @@ -95,7 +95,7 @@ struct parameters global_parameters = 0, 0, 180, - 30 + 15 }; // Recursively traverse query structure to extract terms. @@ -107,20 +107,20 @@ void pull_terms(NMEM nmem, struct ccl_rpn_node *n, char **termlist, int *num) switch (n->kind) { - case CCL_RPN_AND: - case CCL_RPN_OR: - case CCL_RPN_NOT: - case CCL_RPN_PROX: - pull_terms(nmem, n->u.p[0], termlist, num); - pull_terms(nmem, n->u.p[1], termlist, num); - break; - case CCL_RPN_TERM: - nmem_strsplit(nmem, " ", n->u.t.term, &words, &numwords); - for (i = 0; i < numwords; i++) - termlist[(*num)++] = words[i]; - break; - default: // NOOP - break; + case CCL_RPN_AND: + case CCL_RPN_OR: + case CCL_RPN_NOT: + case CCL_RPN_PROX: + pull_terms(nmem, n->u.p[0], termlist, num); + pull_terms(nmem, n->u.p[1], termlist, num); + break; + case CCL_RPN_TERM: + nmem_strsplit(nmem, " ", n->u.t.term, &words, &numwords); + for (i = 0; i < numwords; i++) + termlist[(*num)++] = words[i]; + break; + default: // NOOP + break; } } @@ -276,7 +276,7 @@ xmlDoc *record_to_xml(struct session_database *sdb, Z_External *rec) // Add static values from session database settings if applicable static void insert_settings_parameters(struct session_database *sdb, - struct session *se, char **parms) + struct session *se, char **parms) { struct conf_service *service = global_parameters.server->service; int i; @@ -289,7 +289,7 @@ static void insert_settings_parameters(struct session_database *sdb, int setting; if (md->setting == Metadata_setting_parameter && - (setting = settings_offset(md->name)) > 0) + (setting = settings_offset(md->name)) > 0) { const char *val = session_setting_oneval(sdb, setting); if (val && nparms < MAX_XSLT_ARGS) @@ -322,14 +322,14 @@ static void insert_settings_values(struct session_database *sdb, xmlDoc *doc) int offset; if (md->setting == Metadata_setting_postproc && - (offset = settings_offset(md->name)) > 0) + (offset = settings_offset(md->name)) > 0) { const char *val = session_setting_oneval(sdb, offset); if (val) { xmlNode *r = xmlDocGetRootElement(doc); xmlNode *n = xmlNewTextChild(r, 0, (xmlChar *) "metadata", - (xmlChar *) val); + (xmlChar *) val); xmlSetProp(n, (xmlChar *) "type", (xmlChar *) md->name); } } @@ -337,7 +337,7 @@ static void insert_settings_values(struct session_database *sdb, xmlDoc *doc) } xmlDoc *normalize_record(struct session_database *sdb, struct session *se, - Z_External *rec) + Z_External *rec) { struct database_retrievalmap *m; xmlDoc *rdoc = record_to_xml(sdb, rec); @@ -708,15 +708,6 @@ static void session_init_databases_fun(void *context, struct database *db) new->database = db; new->yaz_marc = 0; -#ifdef HAVE_ICU - if (global_parameters.server && global_parameters.server->icu_chn) - new->pct = pp2_charset_create(global_parameters.server->icu_chn); - else - new->pct = pp2_charset_create(0); -#else // HAVE_ICU - new->pct = pp2_charset_create(0); -#endif // HAVE_ICU - new->map = 0; new->settings = nmem_malloc(se->session_nmem, sizeof(struct settings *) * num); @@ -740,8 +731,6 @@ static void session_database_destroy(struct session_database *sdb) xsltFreeStylesheet(m->stylesheet); if (sdb->yaz_marc) yaz_marc_destroy(sdb->yaz_marc); - if (sdb->pct) - pp2_charset_destroy(sdb->pct); } // Initialize session_database list -- this represents this session's view @@ -791,10 +780,10 @@ void session_apply_setting(struct session *se, char *dbname, char *setting, } // Jakub: This breaks the filter setting. /*if (offset == PZ_ID) - { - yaz_log(YLOG_WARN, "No need to set pz:id setting. Ignoring"); - return; - }*/ + { + yaz_log(YLOG_WARN, "No need to set pz:id setting. Ignoring"); + return; + }*/ new->precedence = 0; new->target = dbname; new->name = setting; @@ -806,23 +795,23 @@ void session_apply_setting(struct session *se, char *dbname, char *setting, // (happens when a search starts and client connections are prepared) switch (offset) { - case PZ_NATIVESYNTAX: - if (sdb->yaz_marc) - { - yaz_marc_destroy(sdb->yaz_marc); - sdb->yaz_marc = 0; - } - break; - case PZ_XSLT: - if (sdb->map) - { - struct database_retrievalmap *m; - // We don't worry about the map structure -- it's in nmem - for (m = sdb->map; m; m = m->next) - xsltFreeStylesheet(m->stylesheet); - sdb->map = 0; - } - break; + case PZ_NATIVESYNTAX: + if (sdb->yaz_marc) + { + yaz_marc_destroy(sdb->yaz_marc); + sdb->yaz_marc = 0; + } + break; + case PZ_XSLT: + if (sdb->map) + { + struct database_retrievalmap *m; + // We don't worry about the map structure -- it's in nmem + for (m = sdb->map; m; m = m->next) + xsltFreeStylesheet(m->stylesheet); + sdb->map = 0; + } + break; } } @@ -862,7 +851,6 @@ struct session *new_session(NMEM nmem) session->watchlist[i].data = 0; session->watchlist[i].fun = 0; } - return session; } @@ -933,7 +921,7 @@ struct record_cluster **show(struct session *s, struct reclist_sortparms *sp, NMEM nmem_show) { struct record_cluster **recs = nmem_malloc(nmem_show, *num - * sizeof(struct record_cluster *)); + * sizeof(struct record_cluster *)); struct reclist_sortparms *spp; int i; #if USE_TIMING @@ -1001,14 +989,14 @@ void statistics(struct session *se, struct statistics *stat) stat->num_no_connection++; switch (client_get_state(cl)) { - case Client_Connecting: stat->num_connecting++; break; - case Client_Initializing: stat->num_initializing++; break; - case Client_Searching: stat->num_searching++; break; - case Client_Presenting: stat->num_presenting++; break; - case Client_Idle: stat->num_idle++; break; - case Client_Failed: stat->num_failed++; break; - case Client_Error: stat->num_error++; break; - default: break; + case Client_Connecting: stat->num_connecting++; break; + case Client_Initializing: stat->num_initializing++; break; + case Client_Searching: stat->num_searching++; break; + case Client_Presenting: stat->num_presenting++; break; + case Client_Idle: stat->num_idle++; break; + case Client_Failed: stat->num_failed++; break; + case Client_Error: stat->num_error++; break; + default: break; } count++; } @@ -1084,7 +1072,8 @@ static struct record_metadata *record_metadata_init( char * p = value; p = normalize7bit_generic(p, " ,/.:(["); - rec_md->data.text = nmem_strdup(nmem, p); + rec_md->data.text.disp = nmem_strdup(nmem, p); + rec_md->data.text.sort = 0; } else if (type == Metadata_type_year) { @@ -1103,7 +1092,7 @@ struct record *ingest_record(struct client *cl, Z_External *rec, int record_no) { xmlDoc *xdoc = normalize_record(client_get_database(cl), - client_get_session(cl), rec); + client_get_session(cl), rec); xmlNode *root, *n; struct record *record; struct record_cluster *cluster; @@ -1112,6 +1101,9 @@ struct record *ingest_record(struct client *cl, Z_External *rec, xmlChar *type = 0; xmlChar *value = 0; struct conf_service *service = global_parameters.server->service; + const char *norm_str = 0; + pp2_relevance_token_t prt = 0; + WRBUF norm_wr = 0; if (!xdoc) return 0; @@ -1123,15 +1115,34 @@ struct record *ingest_record(struct client *cl, Z_External *rec, xmlFreeDoc(xdoc); return 0; } - + record = record_create(se->nmem, service->num_metadata, service->num_sortkeys, cl, record_no); - mergekey_norm = (xmlChar *) nmem_strdup(se->nmem, (char*) mergekey); - xmlFree(mergekey); - normalize7bit_mergekey((char *) mergekey_norm, 0); + prt = pp2_relevance_tokenize( + global_parameters.server->mergekey_pct, (const char *) mergekey); + + + norm_wr = wrbuf_alloc(); + + while ((norm_str = pp2_relevance_token_next(prt))) + { + if (*norm_str) + { + if (wrbuf_len(norm_wr)) + wrbuf_puts(norm_wr, " "); + wrbuf_puts(norm_wr, norm_str); + } + } + + mergekey_norm = (xmlChar *)nmem_strdup(se->nmem, wrbuf_cstr(norm_wr)); + wrbuf_destroy(norm_wr); + + pp2_relevance_token_destroy(prt); + xmlFree(mergekey); + cluster = reclist_insert(se->reclist, global_parameters.server->service, record, (char *) mergekey_norm, @@ -1146,99 +1157,119 @@ struct record *ingest_record(struct client *cl, Z_External *rec, return 0; } relevance_newrec(se->relevance, cluster); - - - // now parsing XML record and adding data to cluster or record metadata - for (n = root->children; n; n = n->next) - { - if (type) - xmlFree(type); - if (value) - xmlFree(value); - type = value = 0; - - if (n->type != XML_ELEMENT_NODE) - continue; - if (!strcmp((const char *) n->name, "metadata")) - { - struct conf_metadata *ser_md = 0; - struct conf_sortkey *ser_sk = 0; - struct record_metadata **wheretoput = 0; - struct record_metadata *rec_md = 0; - int md_field_id = -1; - int sk_field_id = -1; - - type = xmlGetProp(n, (xmlChar *) "type"); - value = xmlNodeListGetString(xdoc, n->children, 1); - - if (!type || !value || !*value) - continue; - - md_field_id - = conf_service_metadata_field_id(service, (const char *) type); - if (md_field_id < 0) - { - yaz_log(YLOG_WARN, - "Ignoring unknown metadata element: %s", type); - continue; - } - - ser_md = &service->metadata[md_field_id]; - - if (ser_md->sortkey_offset >= 0){ - sk_field_id = ser_md->sortkey_offset; - ser_sk = &service->sortkeys[sk_field_id]; - } - - // non-merged metadata - rec_md = record_metadata_init(se->nmem, (char *) value, - ser_md->type); - if (!rec_md) - { - yaz_log(YLOG_WARN, "bad metadata data '%s' for element '%s'", - value, type); - continue; - } - rec_md->next = record->metadata[md_field_id]; - record->metadata[md_field_id] = rec_md; - - // merged metadata - rec_md = record_metadata_init(se->nmem, (char *) value, - ser_md->type); - wheretoput = &cluster->metadata[md_field_id]; - - // and polulate with data: - // assign cluster or record based on merge action - if (ser_md->merge == Metadata_merge_unique) - { - struct record_metadata *mnode; - for (mnode = *wheretoput; mnode; mnode = mnode->next) - if (!strcmp((const char *) mnode->data.text, - rec_md->data.text)) - break; - if (!mnode) - { - rec_md->next = *wheretoput; - *wheretoput = rec_md; - } - } - else if (ser_md->merge == Metadata_merge_longest) - { - if (!*wheretoput - || strlen(rec_md->data.text) - > strlen((*wheretoput)->data.text)) - { - *wheretoput = rec_md; - if (ser_sk) - { - char *s = nmem_strdup(se->nmem, rec_md->data.text); - if (!cluster->sortkeys[sk_field_id]) - cluster->sortkeys[sk_field_id] = - nmem_malloc(se->nmem, - sizeof(union data_types)); - normalize7bit_mergekey(s, - (ser_sk->type == Metadata_sortkey_skiparticle)); - cluster->sortkeys[sk_field_id]->text = s; + + + // now parsing XML record and adding data to cluster or record metadata + for (n = root->children; n; n = n->next) + { + if (type) + xmlFree(type); + if (value) + xmlFree(value); + type = value = 0; + + if (n->type != XML_ELEMENT_NODE) + continue; + if (!strcmp((const char *) n->name, "metadata")) + { + struct conf_metadata *ser_md = 0; + struct conf_sortkey *ser_sk = 0; + struct record_metadata **wheretoput = 0; + struct record_metadata *rec_md = 0; + int md_field_id = -1; + int sk_field_id = -1; + + type = xmlGetProp(n, (xmlChar *) "type"); + value = xmlNodeListGetString(xdoc, n->children, 1); + + if (!type || !value || !*value) + continue; + + md_field_id + = conf_service_metadata_field_id(service, (const char *) type); + if (md_field_id < 0) + { + yaz_log(YLOG_WARN, + "Ignoring unknown metadata element: %s", type); + continue; + } + + ser_md = &service->metadata[md_field_id]; + + if (ser_md->sortkey_offset >= 0){ + sk_field_id = ser_md->sortkey_offset; + ser_sk = &service->sortkeys[sk_field_id]; + } + + // non-merged metadata + rec_md = record_metadata_init(se->nmem, (char *) value, + ser_md->type); + if (!rec_md) + { + yaz_log(YLOG_WARN, "bad metadata data '%s' for element '%s'", + value, type); + continue; + } + rec_md->next = record->metadata[md_field_id]; + record->metadata[md_field_id] = rec_md; + + // merged metadata + rec_md = record_metadata_init(se->nmem, (char *) value, + ser_md->type); + wheretoput = &cluster->metadata[md_field_id]; + + // and polulate with data: + // assign cluster or record based on merge action + if (ser_md->merge == Metadata_merge_unique) + { + struct record_metadata *mnode; + for (mnode = *wheretoput; mnode; mnode = mnode->next) + if (!strcmp((const char *) mnode->data.text.disp, + rec_md->data.text.disp)) + break; + if (!mnode) + { + rec_md->next = *wheretoput; + *wheretoput = rec_md; + } + } + else if (ser_md->merge == Metadata_merge_longest) + { + if (!*wheretoput + || strlen(rec_md->data.text.disp) + > strlen((*wheretoput)->data.text.disp)) + { + *wheretoput = rec_md; + if (ser_sk) + { + const char *sort_str = 0; + int skip_article = + ser_sk->type == Metadata_sortkey_skiparticle; + + if (!cluster->sortkeys[sk_field_id]) + cluster->sortkeys[sk_field_id] = + nmem_malloc(se->nmem, + sizeof(union data_types)); + + prt = pp2_relevance_tokenize( + global_parameters.server->sort_pct, + rec_md->data.text.disp); + + pp2_relevance_token_next(prt); + + sort_str = pp2_get_sort(prt, skip_article); + + cluster->sortkeys[sk_field_id]->text.disp = + rec_md->data.text.disp; + cluster->sortkeys[sk_field_id]->text.sort = + nmem_strdup(se->nmem, sort_str); +#if 0 + yaz_log(YLOG_LOG, "text disp=%s", + cluster->sortkeys[sk_field_id]->text.disp); + yaz_log(YLOG_LOG, "text sort=%s", + cluster->sortkeys[sk_field_id]->text.sort); +#endif + pp2_relevance_token_destroy(prt); } } }