per-field weight is part of "rank" attribute content

For example, rank="2 au 3" gives a rank factor of 3 for terms in
au=terms and a rank factor of 2 otherwise.

diff --git a/doc/common b/doc/common
index c6e4db9..92b4d40 160000 (submodule)
--- a/doc/common
+++ b/doc/common
@@ -1 +1 @@
-Subproject commit c6e4db90ce6779ee42fafd38e776176d38c2844b
+Subproject commit 92b4d4032574351cda7ab8762889dea6699bff94

diff --git a/src/pazpar2_config.c b/src/pazpar2_config.c
index d992a33..b2064f9 100644 (file)
--- a/src/pazpar2_config.c
+++ b/src/pazpar2_config.c
@@ -72,12 +72,11 @@ static void conf_metadata_assign(NMEM nmem,
                                  enum conf_setting_type setting,
                                  int brief,
                                  int termlist,
-                                 int rank,
+                                 const char *rank,
                                  int sortkey_offset,
                                  enum conf_metadata_mergekey mt,
                                  const char *facetrule,
-                                 const char *limitmap,
-                                 const char *frank)
+                                 const char *limitmap)
 {
     assert(nmem && metadata && name);
     
@@ -94,12 +93,11 @@ static void conf_metadata_assign(NMEM nmem,
     metadata->setting = setting;
     metadata->brief = brief;   
     metadata->termlist = termlist;
-    metadata->rank = rank;    
+    metadata->rank = nmem_strdup_null(nmem, rank);
     metadata->sortkey_offset = sortkey_offset;
     metadata->mergekey = mt;
     metadata->facetrule = nmem_strdup_null(nmem, facetrule);
     metadata->limitmap = nmem_strdup_null(nmem, limitmap);
-    metadata->frank = nmem_strdup_null(nmem, frank);
 }
 
 
@@ -175,12 +173,11 @@ static struct conf_metadata* conf_service_add_metadata(
     enum conf_setting_type setting,
     int brief,
     int termlist,
-    int rank,
+    const char *rank,
     int sortkey_offset,
     enum conf_metadata_mergekey mt,
     const char *facetrule,
-    const char *limitmap,
-    const char *frank
+    const char *limitmap
     )
 {
     struct conf_metadata * md = 0;
@@ -192,7 +189,7 @@ static struct conf_metadata* conf_service_add_metadata(
     md = service->metadata + field_id;
     conf_metadata_assign(service->nmem, md, name, type, merge, setting,
                          brief, termlist, rank, sortkey_offset,
-                         mt, facetrule, limitmap, frank);
+                         mt, facetrule, limitmap);
     return md;
 }
 
@@ -284,7 +281,6 @@ static int parse_metadata(struct conf_service *service, xmlNode *n,
     enum conf_metadata_mergekey mergekey_type = Metadata_mergekey_no;
     int brief = 0;
     int termlist = 0;
-    int rank = 0;
     int sortkey_offset = 0;
     xmlChar *xml_name = 0;
     xmlChar *xml_brief = 0;
@@ -297,7 +293,6 @@ static int parse_metadata(struct conf_service *service, xmlNode *n,
     xmlChar *xml_mergekey = 0;
     xmlChar *xml_limitmap = 0;
     xmlChar *xml_icu_chain = 0;
-    xmlChar *xml_frank = 0;
 
     struct _xmlAttr *attr;
     for (attr = n->properties; attr; attr = attr->next)
@@ -335,9 +330,6 @@ static int parse_metadata(struct conf_service *service, xmlNode *n,
         else if (!xmlStrcmp(attr->name, BAD_CAST "limitmap") &&
                  attr->children && attr->children->type == XML_TEXT_NODE)
             xml_limitmap = attr->children->content;
-        else if (!xmlStrcmp(attr->name, BAD_CAST "frank") &&
-                 attr->children && attr->children->type == XML_TEXT_NODE)
-            xml_frank = attr->children->content;
         else
         {
             yaz_log(YLOG_FATAL, "Unknown metadata attribute '%s'", attr->name);
@@ -373,9 +365,6 @@ static int parse_metadata(struct conf_service *service, xmlNode *n,
         }
     }
     
-    if (xml_rank)
-        rank = atoi((const char *) xml_rank);
-
     if (xml_type)
     {
         if (!strcmp((const char *) xml_type, "generic"))
@@ -477,11 +466,11 @@ static int parse_metadata(struct conf_service *service, xmlNode *n,
     conf_service_add_metadata(service, *md_node,
                               (const char *) xml_name,
                               type, merge, setting,
-                              brief, termlist, rank, sortkey_offset,
+                              brief, termlist,
+                              (const char *) xml_rank, sortkey_offset,
                               mergekey_type,
                               (const char *) xml_icu_chain,
-                              (const char *) xml_limitmap,
-                              (const char *) xml_frank);
+                              (const char *) xml_limitmap);
     (*md_node)++;
     return 0;
 }

diff --git a/src/pazpar2_config.h b/src/pazpar2_config.h
index 18dfde6..193adcf 100644 (file)
--- a/src/pazpar2_config.h
+++ b/src/pazpar2_config.h
@@ -73,9 +73,7 @@ struct conf_metadata
     char *name;  // The field name. Output by normalization stylesheet
     int brief;   // Is this element to be returned in the brief format?
     int termlist;// Is this field to be treated as a termlist for browsing?
-    int rank;    // Rank factor. 0 means don't use this field for ranking, 
-                 // 1 is default
-                 // values >1  give additional significance to a field
+    const char *rank;
     int sortkey_offset; // -1 if it's not a sortkey, otherwise index
                         // into service/record_cluster->sortkey array
     enum conf_metadata_type type;
@@ -85,7 +83,6 @@ struct conf_metadata
     char *facetrule;
 
     char *limitmap;  // Should be expanded into service-wide default e.g. pz:limitmap:<name>=value setting
-    char *frank;
 };
 
 

diff --git a/src/relevance.c b/src/relevance.c
index 7f1943a..bc9cb1a 100644 (file)
--- a/src/relevance.c
+++ b/src/relevance.c
@@ -45,18 +45,23 @@ struct word_entry {
 };
 
 static int word_entry_match(struct word_entry *entries, const char *norm_str,
-                            const char *frank, int *local_mult)
+                            const char *rank, int *mult)
 {
     for (; entries; entries = entries->next)
     {
         if (!strcmp(norm_str, entries->norm_str))
         {
             const char *cp = 0;
-            if (frank && (cp = strchr(frank, ' ')))
+            int no_read = 0;
+            sscanf(rank, "%d%n", mult, &no_read);
+            rank += no_read;
+            while (*rank == ' ')
+                rank++;
+            if (no_read > 0 && (cp = strchr(rank, ' ')))
             {
-                if ((cp - frank) == strlen(entries->ccl_field) &&
-                    memcmp(entries->ccl_field, frank, cp - frank) == 0)
-                    *local_mult = atoi(cp + 1);
+                if ((cp - rank) == strlen(entries->ccl_field) &&
+                    memcmp(entries->ccl_field, rank, cp - rank) == 0)
+                    *mult = atoi(cp + 1);
             }
             return entries->termno;
         }
@@ -65,8 +70,8 @@ static int word_entry_match(struct word_entry *entries, const char *norm_str,
 }
 
 void relevance_countwords(struct relevance *r, struct record_cluster *cluster,
-                          const char *words, int multiplier, const char *name,
-                          const char *frank)
+                          const char *words, const char *rank,
+                          const char *name)
 {
     int *mult = cluster->term_frequency_vec_tmp;
     const char *norm_str;
@@ -75,10 +80,11 @@ void relevance_countwords(struct relevance *r, struct record_cluster *cluster,
     for (i = 1; i < r->vec_len; i++)
         mult[i] = 0;
 
+    assert(rank);
     while ((norm_str = pp2_charset_token_next(r->prt)))
     {
-        int local_mult = multiplier;
-        int res = word_entry_match(r->entries, norm_str, frank, &local_mult);
+        int local_mult = 0;
+        int res = word_entry_match(r->entries, norm_str, rank, &local_mult);
         if (res)
         {
             assert(res < r->vec_len);

diff --git a/src/relevance.h b/src/relevance.h
index b22a7a0..35949c0 100644 (file)
--- a/src/relevance.h
+++ b/src/relevance.h
@@ -33,8 +33,8 @@ struct relevance *relevance_create_ccl(pp2_charset_fact_t pft,
 void relevance_destroy(struct relevance **rp);
 void relevance_newrec(struct relevance *r, struct record_cluster *cluster);
 void relevance_countwords(struct relevance *r, struct record_cluster *cluster,
-                          const char *words, int multiplier, const char *name,
-                          const char *frank);
+                          const char *words, const char *multiplier,
+                          const char *name);
 void relevance_donerecord(struct relevance *r, struct record_cluster *cluster);
 
 void relevance_prepare_read(struct relevance *rel, struct reclist *rec);

diff --git a/src/session.c b/src/session.c
index 4049360..18ba514 100644 (file)
--- a/src/session.c
+++ b/src/session.c
@@ -1812,8 +1812,8 @@ static int ingest_to_cluster(struct client *cl,
             struct record_metadata *rec_md = 0;
             int md_field_id = -1;
             int sk_field_id = -1;
-            int rank = 0;
-            xmlChar *rank_str = 0;
+            const char *rank;
+            xmlChar *xml_rank;
             
             type = xmlGetProp(n, (xmlChar *) "type");
             value = xmlNodeListGetString(xdoc, n->children, 1);
@@ -1828,15 +1828,6 @@ static int ingest_to_cluster(struct client *cl,
             
             ser_md = &service->metadata[md_field_id];
 
-            rank_str = xmlGetProp(n, (xmlChar *) "rank");
-            if (rank_str)
-            {
-                rank = atoi((const char *) rank_str);
-                xmlFree(rank_str);
-            }
-            else
-                rank = ser_md->rank;
-            
             if (ser_md->sortkey_offset >= 0)
             {
                 sk_field_id = ser_md->sortkey_offset;
@@ -1849,6 +1840,9 @@ static int ingest_to_cluster(struct client *cl,
             if (!rec_md)
                 continue;
 
+            xml_rank = xmlGetProp(n, (xmlChar *) "rank");
+            rank = xml_rank ? (const char *) xml_rank : ser_md->rank;
+
             wheretoput = &cluster->metadata[md_field_id];
 
             // and polulate with data:
@@ -1937,8 +1931,7 @@ static int ingest_to_cluster(struct client *cl,
             if (rank)
             {
                 relevance_countwords(se->relevance, cluster, 
-                                     (char *) value, rank, ser_md->name,
-                                     ser_md->frank);
+                                     (char *) value, rank, ser_md->name);
             }
 
             // construct facets ... unless the client already has reported them
@@ -1961,6 +1954,8 @@ static int ingest_to_cluster(struct client *cl,
             }
 
             // cleaning up
+            if (xml_rank)
+                xmlFree(xml_rank);
             xmlFree(type);
             xmlFree(value);
             type = value = 0;

diff --git a/test/test_http.cfg b/test/test_http.cfg
index 8904f3b..53a0ff0 100644 (file)
--- a/test/test_http.cfg
+++ b/test/test_http.cfg
@@ -14,7 +14,7 @@
       <metadata name="date" brief="yes" sortkey="numeric" type="year" merge="range"
                termlist="yes"/>
       <metadata name="author" brief="yes" termlist="yes" merge="longest"
-            frank="au 3" rank="2" mergekey="optional" />
+                rank="2 au 3" mergekey="optional" /> <!-- rank="2 au 3" -->
       <metadata name="subject" brief="yes" merge="unique" termlist="yes" rank="3"/>
       <metadata name="id"/>
       <metadata name="lccn" merge="unique"/>