Added the support for multiple org names in the type ahead

ORCID · Jul 13, 2024 · 5861752 · 5861752
1 parent d194a26
commit 5861752
Show file tree

Hide file tree

Showing 6 changed files with 70 additions and 7 deletions.
diff --git a/orcid-core/src/main/java/org/orcid/core/manager/impl/OrgDisambiguatedManagerImpl.java b/orcid-core/src/main/java/org/orcid/core/manager/impl/OrgDisambiguatedManagerImpl.java
@@ -209,7 +209,7 @@ synchronized public void processOrgsWithIncorrectPopularity() {
 
     @Override
     public List<OrgDisambiguated> searchOrgsFromSolr(String searchTerm, int firstResult, int maxResult, boolean fundersOnly) {
-        List<OrgDisambiguatedSolrDocument> docs = orcidSolrOrgsClient.getOrgs(searchTerm, firstResult, maxResult, fundersOnly);
+        List<OrgDisambiguatedSolrDocument> docs = orcidSolrOrgsClient.getOrgs(searchTerm, firstResult, maxResult, fundersOnly, true);
         List<OrgDisambiguated> ret = new ArrayList<OrgDisambiguated>();
         for (OrgDisambiguatedSolrDocument doc : docs) {
             OrgDisambiguated org = convertSolrDocument(doc);

diff --git a/orcid-core/src/main/java/org/orcid/core/solr/OrcidSolrOrgsClient.java b/orcid-core/src/main/java/org/orcid/core/solr/OrcidSolrOrgsClient.java
@@ -4,7 +4,9 @@
 
 import java.io.IOException;
 import java.text.MessageFormat;
+import java.util.ArrayList;
 import java.util.List;
+import java.util.Map;
 
 import javax.annotation.Resource;
 
@@ -33,6 +35,8 @@ public class OrcidSolrOrgsClient {
     private static final String SOLR_SELF_SERVICE_ORGS_QUERY = "(org-disambiguated-id-from-source:%s)^50.0 (org-disambiguated-name%s)^50.0 (org-disambiguated-name-string:%s)^25.0";
 
     private static final String SOLR_ORG_BY_ROR_ID_QUERY = "org-disambiguated-id-from-source:%s";
+
+    private static final String ORG_NAMES_HIGHLIGHT_DELIMITATOR ="::";
 
     public OrgDisambiguatedSolrDocument findById(Long id) {
         SolrQuery query = new SolrQuery();
@@ -50,7 +54,7 @@ public OrgDisambiguatedSolrDocument findById(Long id) {
         return null;
     }
 
-   public List<OrgDisambiguatedSolrDocument> getOrgs(String searchTerm, int firstResult, int maxResult, boolean fundersOnly) {
+   public List<OrgDisambiguatedSolrDocument> getOrgs(String searchTerm, int firstResult, int maxResult, boolean fundersOnly, boolean withNamesHighlight) {
         StringBuilder queryString = new StringBuilder(SOLR_ORGS_QUERY.replace("%s", searchTerm));
         if (fundersOnly) {
             queryString.append(" AND is-funding-org:true");
@@ -60,17 +64,52 @@ public List<OrgDisambiguatedSolrDocument> getOrgs(String searchTerm, int firstRe
         query.setQuery(queryString.toString());
         query.addOrUpdateSort("score", ORDER.desc);
         query.addOrUpdateSort("org-disambiguated-popularity", ORDER.desc);
+
+        // Set the preserveMulti parameter
+        query.setParam("preserveMulti", "true");
+
         if(fundersOnly) {
             query.addFilterQuery(String.format("(%s:(%s OR %s))", SolrConstants.ORG_DISAMBIGUATED_ID_SOURCE_TYPE, "ROR", "FUNDREF"));
         } else {
             query.addFilterQuery(String.format("(%s:(%s))", SolrConstants.ORG_DISAMBIGUATED_ID_SOURCE_TYPE, "ROR"));
         }
 
+        if(withNamesHighlight) {
+            query.setHighlight(withNamesHighlight);
+            query.addHighlightField(SolrConstants.ORG_NAMES);
+            query.setHighlightSnippets(maxResult);
+            query.setHighlightSimplePost(ORG_NAMES_HIGHLIGHT_DELIMITATOR);
+            query.setHighlightSimplePre(ORG_NAMES_HIGHLIGHT_DELIMITATOR);
+        }
+
         LOGGER.debug("SOLR Query: " + query.toQueryString());
 
         try {
             QueryResponse queryResponse = solrReadOnlyOrgsClient.query(query);
-            return queryResponse.getBeans(OrgDisambiguatedSolrDocument.class);
+            List<OrgDisambiguatedSolrDocument> orgs = queryResponse.getBeans(OrgDisambiguatedSolrDocument.class);
+            // Get the highlight results
+            if(withNamesHighlight) {
+                List<OrgDisambiguatedSolrDocument> orgsNamesHighlighted = new ArrayList<OrgDisambiguatedSolrDocument>();
+                Map<String, Map<String, List<String>>> highlightMap = queryResponse.getHighlighting();
+                for(OrgDisambiguatedSolrDocument org : orgs) {
+                 // Print highlighted snippets
+                    if (highlightMap.containsKey(org.getOrgDisambiguatedId())) {
+                        Map<String, List<String>> fieldHighlightMap = highlightMap.get(org.getOrgDisambiguatedId());
+                        if (fieldHighlightMap.containsKey(SolrConstants.ORG_NAMES)) {
+                            List<String> highlights = fieldHighlightMap.get(SolrConstants.ORG_NAMES);
+                            OrgDisambiguatedSolrDocument highlightOrg;
+                            for (String highlight : highlights) {
+                                //strip the highlight delimitator ORG_NAMES_HIGHLIGHT_DELIMITATOR
+                                highlightOrg = new OrgDisambiguatedSolrDocument(org);
+                                highlightOrg.setOrgDisambiguatedName(highlight.replaceAll(ORG_NAMES_HIGHLIGHT_DELIMITATOR, ""));
+                                orgsNamesHighlighted.add(highlightOrg);
+                            }
+                        }
+                    }
+                }   
+                return orgsNamesHighlighted;
+            }
+            return orgs;
         } catch (SolrServerException | IOException se) {
             String errorMessage = MessageFormat.format("Error when attempting to search for orgs, with search term {0}", new Object[] { searchTerm });
             throw new NonTransientDataAccessResourceException(errorMessage, se);

diff --git a/orcid-core/src/main/java/org/orcid/pojo/summary/RecordSummaryPojo.java b/orcid-core/src/main/java/org/orcid/pojo/summary/RecordSummaryPojo.java
@@ -1,6 +1,5 @@
 package org.orcid.pojo.summary;
 
-
 import java.util.List;
 
 import javax.xml.bind.annotation.XmlRootElement;

diff --git a/orcid-core/src/test/java/org/orcid/core/solr/OrcidSolrOrgsClientTest.java b/orcid-core/src/test/java/org/orcid/core/solr/OrcidSolrOrgsClientTest.java
@@ -70,7 +70,7 @@ public void pesistAndFindByIdTest() throws IOException, SolrServerException {
     @Test
     public void findFundrefOrgsOnlyTest() throws SolrServerException, IOException {
         ArgumentCaptor<SolrQuery> captor = ArgumentCaptor.forClass(SolrQuery.class);
-        orcidSolrOrgsClient.getOrgs("xxx", 0, 0, true);
+        orcidSolrOrgsClient.getOrgs("xxx", 0, 0, true, false);
         Mockito.verify(solrReadOnlyOrgsClientMock).query(captor.capture());
 
         SolrQuery query = captor.getValue();
@@ -81,7 +81,7 @@ public void findFundrefOrgsOnlyTest() throws SolrServerException, IOException {
     @Test
     public void findNonFundrefOrgsOnlyTest() throws SolrServerException, IOException {
         ArgumentCaptor<SolrQuery> captor = ArgumentCaptor.forClass(SolrQuery.class);
-        orcidSolrOrgsClient.getOrgs("xxx", 0, 0, false);
+        orcidSolrOrgsClient.getOrgs("xxx", 0, 0, false, false);
         Mockito.verify(solrReadOnlyOrgsClientMock).query(captor.capture());
 
         SolrQuery query = captor.getValue();

diff --git a/orcid-utils/src/main/java/org/orcid/utils/solr/entities/OrgDisambiguatedSolrDocument.java b/orcid-utils/src/main/java/org/orcid/utils/solr/entities/OrgDisambiguatedSolrDocument.java
@@ -1,7 +1,9 @@
 package org.orcid.utils.solr.entities;
 
 import java.io.Serializable;
+import java.util.ArrayList;
 import java.util.List;
+import java.util.Set;
 
 import org.apache.solr.client.solrj.beans.Field;
 
@@ -287,5 +289,25 @@ public boolean equals(Object obj) {
             return false;
         return true;
     }
+
+   public  OrgDisambiguatedSolrDocument() {};
+
+   public  OrgDisambiguatedSolrDocument(OrgDisambiguatedSolrDocument originalDoc) {     
+        this.orgDisambiguatedId = String.valueOf(originalDoc.getOrgDisambiguatedId());
+        this.orgDisambiguatedName = originalDoc.getOrgDisambiguatedName();
+        this.orgDisambiguatedCity = originalDoc.getOrgDisambiguatedCity(); 
+        this.orgDisambiguatedRegion = originalDoc.getOrgDisambiguatedRegion();
+        this.orgDisambiguatedCountry = originalDoc.getOrgDisambiguatedCountry();
+        this.orgDisambiguatedIdFromSource = originalDoc.getOrgDisambiguatedIdFromSource();
+        this.orgDisambiguatedIdSourceType = originalDoc.getOrgDisambiguatedIdSourceType();
+        this.orgDisambiguatedType = originalDoc.getOrgDisambiguatedType();
+        this.orgDisambiguatedPopularity =originalDoc.getOrgDisambiguatedPopularity();
+        this.orgNames = originalDoc.getOrgNames();
+        this.isFundingOrg =originalDoc.isFundingOrg();
+        this.orgLocationsJson = originalDoc.getOrgLocationsJson();
+        this.orgNamesJson = originalDoc.getOrgNamesJson();
+        this.isOrgChosenByMember = originalDoc.isOrgChosenByMember();
+        this.orgDisambiguatedStatus = originalDoc.getOrgDisambiguatedStatus();
+    }
 
 }
diff --git a/properties/development.properties b/properties/development.properties
@@ -130,7 +130,7 @@ org.orcid.core.works.bulk.write.max=100
 org.orcid.core.issn.source=APP-POCQZAUC70YOFL9R
 
 # Orgs grouping
-org.orcid.core.orgs.query=(org-disambiguated-name:\"%s\") ^100.0  (org-disambiguated-name:%s*) ^10.0
+org.orcid.core.orgs.query=(org-names:\"%s\")^100.0  (org-names:\"%s\"~15)^80.0 (org-names:*%s*)^10.0
 org.orcid.core.orgsToGroup.query=select a.* from org_disambiguated a full outer join org_disambiguated_external_identifier b on a.id = b.org_disambiguated_id where b.identifier_type = 'FUNDREF' and a.source_type='ROR'
 
 ################
@@ -257,3 +257,6 @@ org.orcid.scheduler.autospam.enabled=true
 # Slack
 org.orcid.core.autospam.slackChannel=collab-spam-reports
 org.orcid.core.autospam.webhookUrl=
+
+#org.orcid.persistence.liquibase.enabled=false
+org.orcid.persistence.solr.read.only.url=http://localhost:8983/solr