-/* $Id: logic.c,v 1.20 2007-04-23 21:05:23 adam Exp $
+/* $Id: logic.c,v 1.25 2007-04-26 21:33:32 marc Exp $
Copyright (c) 2006-2007, Index Data.
This file is part of Pazpar2.
#include "database.h"
#include "client.h"
#include "settings.h"
+#include "normalize7bit.h"
#define MAX_CHUNK 15
}
}
-char *normalize_mergekey(char *buf, int skiparticle)
-{
- char *p = buf, *pout = buf;
-
- if (skiparticle)
- {
- char firstword[64];
- char articles[] = "the den der die des an a "; // must end in space
-
- while (*p && !isalnum(*p))
- p++;
- pout = firstword;
- while (*p && *p != ' ' && pout - firstword < 62)
- *(pout++) = tolower(*(p++));
- *(pout++) = ' ';
- *(pout++) = '\0';
- if (!strstr(articles, firstword))
- p = buf;
- pout = buf;
- }
-
- while (*p)
- {
- while (*p && !isalnum(*p))
- p++;
- while (isalnum(*p))
- *(pout++) = tolower(*(p++));
- if (*p)
- *(pout++) = ' ';
- while (*p && !isalnum(*p))
- p++;
- }
- if (buf != pout)
- do {
- *(pout--) = '\0';
- }
- while (pout > buf && *pout == ' ');
-
- return buf;
-}
-
-// Extract what appears to be years from buf, storing highest and
-// lowest values.
-static int extract_years(const char *buf, int *first, int *last)
-{
- *first = -1;
- *last = -1;
- while (*buf)
- {
- const char *e;
- int len;
-
- while (*buf && !isdigit(*buf))
- buf++;
- len = 0;
- for (e = buf; *e && isdigit(*e); e++)
- len++;
- if (len == 4)
- {
- int value = atoi(buf);
- if (*first < 0 || value < *first)
- *first = value;
- if (*last < 0 || value > *last)
- *last = value;
- }
- buf = e;
- }
- return *first;
-}
static void add_facet(struct session *s, const char *type, const char *value)
}
s->termlists[i].name = nmem_strdup(s->nmem, type);
- s->termlists[i].termlist = termlist_create(s->nmem, s->expected_maxrecs, 15);
+ s->termlists[i].termlist
+ = termlist_create(s->nmem, s->expected_maxrecs, 15);
s->num_termlists = i + 1;
}
termlist_insert(s->termlists[i].termlist, value);
return 0;
}
}
-#endif
-
-#if 0
+#else
// do it another way to detect transformation errors right now
// but does not seem to work either!
{
yaz_log(YLOG_WARN, "No settings for %s", sdb->database->url);
return -1;
}
- if ((s = session_setting_oneval(sdb, PZ_NATIVESYNTAX)) && !strncmp(s, "iso2709", 7))
+ if ((s = session_setting_oneval(sdb, PZ_NATIVESYNTAX))
+ && !strncmp(s, "iso2709", 7))
{
char *encoding = "marc-8s", *e;
yaz_iconv_t cm;
// This analyzes settings and recomputes any supporting data structures
// if necessary.
-static int prepare_session_database(struct session *se, struct session_database *sdb)
+static int prepare_session_database(struct session *se,
+ struct session_database *sdb)
{
if (!sdb->settings)
{
- yaz_log(YLOG_WARN, "No settings associated with %s", sdb->database->url);
+ yaz_log(YLOG_WARN,
+ "No settings associated with %s", sdb->database->url);
return -1;
}
if (sdb->settings[PZ_NATIVESYNTAX] && !sdb->yaz_marc)
}
-void session_set_watch(struct session *s, int what, session_watchfun fun, void *data)
+void session_set_watch(struct session *s, int what,
+ session_watchfun fun, void *data)
{
s->watchlist[what].fun = fun;
s->watchlist[what].data = data;
}
// Associates a set of clients with a session;
-// Note: Session-databases represent databases with per-session setting overrides
+// Note: Session-databases represent databases with per-session
+// setting overrides
int select_targets(struct session *se, struct database_criterion *crit)
{
while (se->clients)
new->values = 0;
for (subi = 0; subi < subnum; subi++)
{
- struct database_criterion_value *newv = nmem_malloc(m, sizeof(*newv));
+ struct database_criterion_value *newv
+ = nmem_malloc(m, sizeof(*newv));
newv->value = subvalues[subi];
newv->next = new->values;
new->values = newv;
{
if (prepare_session_database(se, client_get_database(cl)) < 0)
return "CONFIG_ERROR";
- if (client_parse_query(cl, query) < 0) // Query must parse for all targets
+ // Query must parse for all targets
+ if (client_parse_query(cl, query) < 0)
return "QUERY";
}
new->database = db;
new->yaz_marc = 0;
new->map = 0;
- new->settings = nmem_malloc(se->session_nmem, sizeof(struct settings *) * num);
+ new->settings
+ = nmem_malloc(se->session_nmem, sizeof(struct settings *) * num);
memset(new->settings, 0, sizeof(struct settings*) * num);
+
if (db->settings)
{
for (i = 0; i < num; i++)
// Probably session_init_databases_fun should be refactored instead of
// called here.
-static struct session_database *load_session_database(struct session *se, char *id)
+static struct session_database *load_session_database(struct session *se,
+ char *id)
{
struct database *db = find_database(id, 0);
}
// Find an existing session database. If not found, load it
-static struct session_database *find_session_database(struct session *se, char *id)
+static struct session_database *find_session_database(struct session *se,
+ char *id)
{
struct session_database *sdb;
}
// Apply a session override to a database
-void session_apply_setting(struct session *se, char *dbname, char *setting, char *value)
+void session_apply_setting(struct session *se, char *dbname, char *setting,
+ char *value)
{
struct session_database *sdb = find_session_database(se, dbname);
struct setting *new = nmem_malloc(se->session_nmem, sizeof(*new));
return 0;
}
-struct record_cluster **show(struct session *s, struct reclist_sortparms *sp, int start,
- int *num, int *total, int *sumhits, NMEM nmem_show)
+struct record_cluster **show(struct session *s, struct reclist_sortparms *sp,
+ int start, int *num, int *total, int *sumhits,
+ NMEM nmem_show)
{
struct record_cluster **recs = nmem_malloc(nmem_show, *num
* sizeof(struct record_cluster *));
{
xmlDoc *xdoc = normalize_record(client_get_database(cl), rec);
xmlNode *root, *n;
- struct record *res;
+ struct record *record;
struct record_cluster *cluster;
struct session *se = client_get_session(cl);
xmlChar *mergekey, *mergekey_norm;
return 0;
}
- res = nmem_malloc(se->nmem, sizeof(struct record));
- res->next = 0;
- res->client = cl;
- res->metadata = nmem_malloc(se->nmem,
+#if 0
+ record = nmem_malloc(se->nmem, sizeof(struct record));
+ record->next = 0;
+ record->client = cl;
+ record->metadata = nmem_malloc(se->nmem,
sizeof(struct record_metadata*) * service->num_metadata);
- memset(res->metadata, 0, sizeof(struct record_metadata*) * service->num_metadata);
+ memset(record->metadata, 0,
+ sizeof(struct record_metadata*) * service->num_metadata);
+
+#else
+ record = record_create(se->nmem,
+ service->num_metadata, service->num_sortkeys);
+ record_assign_client(record, cl);
+#endif
mergekey_norm = (xmlChar *) nmem_strdup(se->nmem, (char*) mergekey);
xmlFree(mergekey);
- normalize_mergekey((char *) mergekey_norm, 0);
+ normalize7bit_mergekey((char *) mergekey_norm, 0);
cluster = reclist_insert(se->reclist,
global_parameters.server->service,
- res, (char *) mergekey_norm,
+ record, (char *) mergekey_norm,
&se->total_merged);
if (global_parameters.dump_records)
yaz_log(YLOG_LOG, "Cluster id %d from %s (#%d)", cluster->recid,
continue;
if (!strcmp((const char *) n->name, "metadata"))
{
- struct conf_metadata *md = 0;
- struct conf_sortkey *sk = 0;
- struct record_metadata **wheretoput, *newm;
- int imeta;
+ 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;
int first, last;
type = xmlGetProp(n, (xmlChar *) "type");
if (!type || !value)
continue;
+#if 0
// First, find out what field we're looking at
- for (imeta = 0; imeta < service->num_metadata; imeta++)
- if (!strcmp((const char *) type, service->metadata[imeta].name))
+ for (md_field_id = 0; md_field_id < service->num_metadata;
+ md_field_id++)
+ if (!strcmp((const char *) type,
+ service->metadata[md_field_id].name))
{
- md = &service->metadata[imeta];
- if (md->sortkey_offset >= 0)
- sk = &service->sortkeys[md->sortkey_offset];
+ 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];
+ }
break;
}
- if (!md)
+
+ if (!ser_md)
+ {
+ yaz_log(YLOG_WARN,
+ "Ignoring unknown metadata element: %s", type);
+ continue;
+ }
+
+#else
+ 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);
+ 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];
+ }
+#endif
+
// Find out where we are putting it
- if (md->merge == Metadata_merge_no)
- wheretoput = &res->metadata[imeta];
+ if (ser_md->merge == Metadata_merge_no)
+ wheretoput = &record->metadata[md_field_id];
else
- wheretoput = &cluster->metadata[imeta];
+ wheretoput = &cluster->metadata[md_field_id];
// Put it there
- newm = nmem_malloc(se->nmem, sizeof(struct record_metadata));
- newm->next = 0;
- if (md->type == Metadata_type_generic)
+
+#if 1 //something wrong with constructor, debug tomorrow
+ rec_md = nmem_malloc(se->nmem, sizeof(struct record_metadata));
+ rec_md->next = 0;
+#else
+ rec_md = record_metadata_create(se->nmem);
+#endif
+
+ if (ser_md->type == Metadata_type_generic)
{
char *p, *pe;
for (p = (char *) value; *p && isspace(*p); p++)
for (pe = p + strlen(p) - 1;
pe > p && strchr(" ,/.:([", *pe); pe--)
*pe = '\0';
- newm->data.text = nmem_strdup(se->nmem, p);
+ //char * normalize7bit_generic(char* str, char* rm_chars);
+
+ rec_md->data.text = nmem_strdup(se->nmem, p);
}
- else if (md->type == Metadata_type_year)
+ else if (ser_md->type == Metadata_type_year)
{
if (extract_years((char *) value, &first, &last) < 0)
continue;
}
else
{
- yaz_log(YLOG_WARN, "Unknown type in metadata element %s", type);
+ yaz_log(YLOG_WARN,
+ "Unknown type in metadata element %s", type);
continue;
}
- if (md->type == Metadata_type_year && md->merge != Metadata_merge_range)
+ if (ser_md->type == Metadata_type_year
+ && ser_md->merge != Metadata_merge_range)
{
yaz_log(YLOG_WARN, "Only range merging supported for years");
continue;
}
- if (md->merge == Metadata_merge_unique)
+ 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, newm->data.text))
+ if (!strcmp((const char *) mnode->data.text,
+ rec_md->data.text))
break;
if (!mnode)
{
- newm->next = *wheretoput;
- *wheretoput = newm;
+ rec_md->next = *wheretoput;
+ *wheretoput = rec_md;
}
}
- else if (md->merge == Metadata_merge_longest)
+ else if (ser_md->merge == Metadata_merge_longest)
{
- if (!*wheretoput ||
- strlen(newm->data.text) > strlen((*wheretoput)->data.text))
+ if (!*wheretoput
+ || strlen(rec_md->data.text)
+ > strlen((*wheretoput)->data.text))
{
- *wheretoput = newm;
- if (sk)
+ *wheretoput = rec_md;
+ if (ser_sk)
{
- char *s = nmem_strdup(se->nmem, newm->data.text);
- if (!cluster->sortkeys[md->sortkey_offset])
- cluster->sortkeys[md->sortkey_offset] =
- nmem_malloc(se->nmem, sizeof(union data_types));
- normalize_mergekey(s,
- (sk->type == Metadata_sortkey_skiparticle));
- cluster->sortkeys[md->sortkey_offset]->text = s;
+ 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;
}
}
}
- else if (md->merge == Metadata_merge_all || md->merge == Metadata_merge_no)
+ else if (ser_md->merge == Metadata_merge_all
+ || ser_md->merge == Metadata_merge_no)
{
- newm->next = *wheretoput;
- *wheretoput = newm;
+ rec_md->next = *wheretoput;
+ *wheretoput = rec_md;
}
- else if (md->merge == Metadata_merge_range)
+ else if (ser_md->merge == Metadata_merge_range)
{
- assert(md->type == Metadata_type_year);
+ assert(ser_md->type == Metadata_type_year);
if (!*wheretoput)
{
- *wheretoput = newm;
+ *wheretoput = rec_md;
(*wheretoput)->data.number.min = first;
(*wheretoput)->data.number.max = last;
- if (sk)
- cluster->sortkeys[md->sortkey_offset] = &newm->data;
+ if (ser_sk)
+ cluster->sortkeys[sk_field_id]
+ = &rec_md->data;
}
else
{
(*wheretoput)->data.number.max = last;
}
#ifdef GAGA
- if (sk)
+ if (ser_sk)
{
- union data_types *sdata = cluster->sortkeys[md->sortkey_offset];
- yaz_log(YLOG_LOG, "SK range: %d-%d", sdata->number.min, sdata->number.max);
+ union data_types *sdata
+ = cluster->sortkeys[sk_field_id];
+ yaz_log(YLOG_LOG, "SK range: %d-%d",
+ sdata->number.min, sdata->number.max);
}
#endif
}
else
- yaz_log(YLOG_WARN, "Don't know how to merge on element name %s", md->name);
+ yaz_log(YLOG_WARN,
+ "Don't know how to merge on element name %s",
+ ser_md->name);
- if (md->rank)
+ if (ser_md->rank)
relevance_countwords(se->relevance, cluster,
- (char *) value, md->rank);
- if (md->termlist)
+ (char *) value, ser_md->rank);
+ if (ser_md->termlist)
{
- if (md->type == Metadata_type_year)
+ if (ser_md->type == Metadata_type_year)
{
char year[64];
sprintf(year, "%d", last);
type = value = 0;
}
else
- yaz_log(YLOG_WARN, "Unexpected element %s in internal record", n->name);
+ yaz_log(YLOG_WARN,
+ "Unexpected element %s in internal record", n->name);
}
if (type)
xmlFree(type);
relevance_donerecord(se->relevance, cluster);
se->total_records++;
- return res;
+ return record;
}
projects / pazpar2-moved-to-github.git / blobdiff