litvinovg/vitro
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

NIHVIVO-2458 SolrPagedSearchController: remove code supporting alpha parameter, since not used; refinement links.

Browse source
This commit is contained in:
ryounes 2011-04-29 18:09:38 +00:00
parent 5467d62023
commit 57d9d061b5
2 changed files with 123 additions and 177 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
webapp/src/edu/cornell/mannlib/vitro/webapp/dao/jena/ApplicationDaoJena.java
View file

@ -78,7 +78,7 @@ public class ApplicationDaoJena extends JenaBaseDao implements ApplicationDao {
// namespace with a final slash, so this makes matching easier.
// It also accords with the way the default namespace is defined.
if (!namespace.endsWith("/")) {
namespace = namespace + "/";
namespace += "/";
}
externallyLinkedNamespaces.add(namespace);
}

420
webapp/src/edu/cornell/mannlib/vitro/webapp/search/controller/SolrPagedSearchController.java
View file

@ -4,6 +4,7 @@ package edu.cornell.mannlib.vitro.webapp.search.controller;
import java.io.IOException;
import java.util.ArrayList;
import java.util.Collection;
import java.util.Collections;
import java.util.Comparator;
import java.util.HashMap;
@ -160,9 +161,6 @@ public class SolrPagedSearchController extends FreemarkerHttpServlet {
log.debug("IndividualDao is " + iDao.toString() + " Public classes in the classgroup are " + grpDao.getPublicGroupsWithVClasses().toString());
log.debug("VClassDao is "+ vclassDao.toString() );
// RY Not sure where/how this is used ***
//String alphaFilter = vreq.getParameter("alpha");
int startIndex = 0;
try{
startIndex = Integer.parseInt(vreq.getParameter("startIndex"));
@ -182,10 +180,7 @@ public class SolrPagedSearchController extends FreemarkerHttpServlet {
int maxHitSize = DEFAULT_MAX_SEARCH_SIZE ;
if( startIndex >= DEFAULT_MAX_SEARCH_SIZE - hitsPerPage )
maxHitSize = startIndex + DEFAULT_MAX_SEARCH_SIZE ;
// if( alphaFilter != null ){
// maxHitSize = maxHitSize * 2;
// hitsPerPage = maxHitSize;
// }
log.debug("maxHitSize is " + maxHitSize);
String qtxt = vreq.getParameter(VitroQuery.QUERY_PARAMETER_NAME);
@ -339,40 +334,36 @@ public class SolrPagedSearchController extends FreemarkerHttpServlet {
if (type != null && type.getName() != null)
body.put("typeName", type.getName());
}
//
// /* Add classgroup and type refinement links to body */
// if( wasHtmlRequested ){
// // Search request includes no classgroup and no type, so add classgroup search refinement links.
// if ( !classGroupFilterRequested && !typeFiltereRequested ) {
// List<VClassGroup> classgroups = getClassGroups(grpDao, topDocs, searcherForRequest);
// List<VClassGroupSearchLink> classGroupLinks = new ArrayList<VClassGroupSearchLink>(classgroups.size());
// for (VClassGroup vcg : classgroups) {
// if (vcg.getPublicName() != null) {
// classGroupLinks.add(new VClassGroupSearchLink(qtxt, vcg));
// }
// }
// body.put("classGroupLinks", classGroupLinks);
//
// // Search request is for a classgroup, so add rdf:type search refinement links
// // but try to filter out classes that are subclasses
// } else if ( classGroupFilterRequested && !typeFiltereRequested ) {
// List<VClass> vClasses = getVClasses(vclassDao,topDocs,searcherForRequest);
// List<VClassSearchLink> vClassLinks = new ArrayList<VClassSearchLink>(vClasses.size());
// for (VClass vc : vClasses) {
// vClassLinks.add(new VClassSearchLink(qtxt, vc));
// }
// body.put("classLinks", vClassLinks);
// pagingLinkParams.put("classgroup", classGroupParam);
//
// // This case is never displayed
//// } else if (!StringUtils.isEmpty(alphaFilter)) {
//// body.put("alphas", getAlphas(topDocs, searcherForRequest));
//// alphaSortIndividuals(beans);
// } else {
// pagingLinkParams.put("type", typeParam);
// }
// }
//
/* Add classgroup and type refinement links to body */
if( wasHtmlRequested ){
// Search request includes no classgroup and no type, so add classgroup search refinement links.
if ( !classGroupFilterRequested && !typeFiltereRequested ) {
List<VClassGroup> classgroups = getClassGroups(grpDao, docs);
List<VClassGroupSearchLink> classGroupLinks = new ArrayList<VClassGroupSearchLink>(classgroups.size());
for (VClassGroup vcg : classgroups) {
if (vcg.getPublicName() != null) {
classGroupLinks.add(new VClassGroupSearchLink(qtxt, vcg));
}
}
body.put("classGroupLinks", classGroupLinks);
// Search request is for a classgroup, so add rdf:type search refinement links
// but try to filter out classes that are subclasses
} else if ( classGroupFilterRequested && !typeFiltereRequested ) {
List<VClass> vClasses = getVClasses(vclassDao, docs);
List<VClassSearchLink> vClassLinks = new ArrayList<VClassSearchLink>(vClasses.size());
for (VClass vc : vClasses) {
vClassLinks.add(new VClassSearchLink(qtxt, vc));
}
body.put("classLinks", vClassLinks);
pagingLinkParams.put("classgroup", classGroupParam);
} else {
pagingLinkParams.put("type", typeParam);
}
}
// Convert search result individuals to template model objects
// RY If this diverges significantly from what's used on the index page,
// create a different template model.
@ -407,53 +398,25 @@ public class SolrPagedSearchController extends FreemarkerHttpServlet {
}
}
private void alphaSortIndividuals(List<Individual> beans) {
Collections.sort(beans, new Comparator< Individual >(){
public int compare(Individual o1, Individual o2) {
if( o1 == null || o1.getName() == null )
return 1;
else
return o1.getName().compareTo(o2.getName());
}});
}
private List<String> getAlphas(TopDocs topDocs, IndexSearcher searcher) {
Set<String> alphas = new HashSet<String>();
for(int i=0;i<topDocs.scoreDocs.length; i++){
Document doc;
try {
doc = searcher.doc(topDocs.scoreDocs[i].doc);
String name =doc.get(Entity2LuceneDoc.term.NAME);
if( name != null && name.length() > 0)
alphas.add( name.substring(0, 1));
} catch (CorruptIndexException e) {
log.debug("Could not get alphas for document",e);
} catch (IOException e) {
log.debug("Could not get alphas for document",e);
}
}
return new ArrayList<String>(alphas);
}
/**
* Get the class groups represented for the individuals in the topDocs.
*/
private List<VClassGroup> getClassGroups(VClassGroupDao grpDao, TopDocs topDocs,
IndexSearcher searcherForRequest) {
private List<VClassGroup> getClassGroups(VClassGroupDao grpDao, SolrDocumentList docs) {
LinkedHashMap<String,VClassGroup> grpMap = grpDao.getClassGroupMap();
int n = grpMap.size();
HashSet<String> classGroupsInHits = new HashSet<String>(n);
int grpsFound = 0;
for(int i=0; i<topDocs.scoreDocs.length && n > grpsFound ;i++){
long hitCount = docs.getNumFound();
for(int i=0; i<hitCount && n > grpsFound ;i++){
try{
Document doc = searcherForRequest.doc(topDocs.scoreDocs[i].doc);
Field[] grps = doc.getFields(Entity2LuceneDoc.term.CLASSGROUP_URI);
if(grps != null || grps.length > 0){
for(int j=0;j<grps.length;j++){
String groupUri = grps[j].stringValue();
SolrDocument doc = docs.get(i);
Collection<Object> grps = doc.getFieldValues(Entity2LuceneDoc.term.CLASSGROUP_URI);
if (grps != null) {
for (Object o : grps) {
String groupUri = o.toString();
if( groupUri != null && ! classGroupsInHits.contains(groupUri)){
classGroupsInHits.add(groupUri);
grpsFound++;
@ -462,7 +425,7 @@ public class SolrPagedSearchController extends FreemarkerHttpServlet {
}
}
}
}catch(Exception e){
} catch(Exception e) {
log.error("problem getting VClassGroups from search hits "
+ e.getMessage());
}
@ -485,6 +448,148 @@ public class SolrPagedSearchController extends FreemarkerHttpServlet {
return classgroups;
}
private List<VClass> getVClasses(VClassDao vclassDao, SolrDocumentList docs){
HashSet<String> typesInHits = getVClassUrisForHits(docs);
List<VClass> classes = new ArrayList<VClass>(typesInHits.size());
Iterator<String> it = typesInHits.iterator();
while(it.hasNext()){
String typeUri = it.next();
try{
if( VitroVocabulary.OWL_THING.equals(typeUri))
continue;
VClass type = vclassDao.getVClassByURI(typeUri);
if( ! type.isAnonymous() &&
type.getName() != null && !"".equals(type.getName()) &&
type.getGroupURI() != null ) //don't display classes that aren't in classgroups
classes.add(type);
}catch(Exception ex){
if( log.isDebugEnabled() )
log.debug("could not add type " + typeUri, ex);
}
}
Collections.sort(classes, new Comparator<VClass>(){
public int compare(VClass o1, VClass o2) {
return o1.compareTo(o2);
}});
return classes;
}
private HashSet<String> getVClassUrisForHits(SolrDocumentList docs){
HashSet<String> typesInHits = new HashSet<String>();
for (SolrDocument doc : docs) {
try {
Collection<Object> types = doc.getFieldValues(Entity2LuceneDoc.term.RDFTYPE);
if (types != null) {
for (Object o : types) {
String typeUri = o.toString();
typesInHits.add(typeUri);
}
}
} catch (Exception e) {
log.error("problems getting rdf:type for search hits",e);
}
}
return typesInHits;
}
private Analyzer getAnalyzer(ServletContext servletContext) throws SearchException {
Object obj = servletContext.getAttribute(LuceneSetup.ANALYZER);
if( obj == null || !(obj instanceof Analyzer) )
throw new SearchException("Could not get analyzer");
else
return (Analyzer)obj;
}
private Query getQuery(VitroRequest request,
Analyzer analyzer, String querystr ) throws SearchException, ParseException {
Query query = null;
try{
//String querystr = request.getParameter(VitroQuery.QUERY_PARAMETER_NAME);
if( querystr == null){
log.error("There was no Parameter '"+VitroQuery.QUERY_PARAMETER_NAME
+"' in the request.");
return null;
}else if( querystr.length() > MAX_QUERY_LENGTH ){
log.debug("The search was too long. The maximum " +
"query length is " + MAX_QUERY_LENGTH );
return null;
}
log.debug("Parsing query using QueryParser ");
QueryParser parser = getQueryParser(analyzer);
query = parser.parse(querystr);
//check if this is classgroup filtered
Object param = request.getParameter("classgroup");
if( param != null && !"".equals(param)){
log.debug("Firing classgroup query ");
log.debug("request.getParameter(classgroup) is "+ param.toString());
BooleanQuery boolQuery = new BooleanQuery();
boolQuery.add( query, BooleanClause.Occur.MUST);
boolQuery.add( new TermQuery(
new Term(Entity2LuceneDoc.term.CLASSGROUP_URI,
(String)param)),
BooleanClause.Occur.MUST);
query = boolQuery;
}
//check if this is rdf:type filtered
param = request.getParameter("type");
if( param != null && !"".equals(param)){
log.debug("Firing type query ");
log.debug("request.getParameter(type) is "+ param.toString());
BooleanQuery boolQuery = new BooleanQuery();
boolQuery.add( query, BooleanClause.Occur.MUST);
boolQuery.add( new TermQuery(
new Term(Entity2LuceneDoc.term.RDFTYPE,
(String)param)),
BooleanClause.Occur.MUST);
query = boolQuery;
}
log.debug("Query: " + query);
} catch (ParseException e) {
throw new ParseException(e.getMessage());
} catch (Exception ex){
throw new SearchException(ex.getMessage());
}
return query;
}
@SuppressWarnings("static-access")
private QueryParser getQueryParser(Analyzer analyzer){
//defaultSearchField indicates which field search against when there is no term
//indicated in the query string.
//The analyzer is needed so that we use the same analyzer on the search queries as
//was used on the text that was indexed.
//QueryParser qp = new QueryParser("NAME",analyzer);
//this sets the query parser to AND all of the query terms it finds.
//set up the map of stemmed field names -> unstemmed field names
// HashMap<String,String> map = new HashMap<String, String>();
// map.put(Entity2LuceneDoc.term.ALLTEXT,Entity2LuceneDoc.term.ALLTEXTUNSTEMMED);
// qp.setStemmedToUnstemmed(map);
MultiFieldQueryParser qp = new MultiFieldQueryParser(Version.LUCENE_29, new String[]{
"name", "nameunstemmed", "type", "moniker", "ALLTEXT", "ALLTEXTUNSTEMMED", "nameraw" , "classLocalName", "classLocalNameLowerCase" }, analyzer);
// QueryParser qp = new QueryParser(Version.LUCENE_29, "name", analyzer);
//AND_OPERATOR returns documents even if the terms in the query lie in different fields.
//The only requirement is that they exist in a single document.
//qp.setDefaultOperator(QueryParser.AND_OPERATOR);
return qp;
}
private class VClassGroupSearchLink extends LinkTemplateModel {
VClassGroupSearchLink(String querytext, VClassGroup classgroup) {
@ -554,165 +659,6 @@ public class SolrPagedSearchController extends FreemarkerHttpServlet {
}
}
private List<VClass> getVClasses(VClassDao vclassDao, TopDocs topDocs,
IndexSearcher searherForRequest){
HashSet<String> typesInHits = getVClassUrisForHits(topDocs,searherForRequest);
List<VClass> classes = new ArrayList<VClass>(typesInHits.size());
Iterator<String> it = typesInHits.iterator();
while(it.hasNext()){
String typeUri = it.next();
try{
if( VitroVocabulary.OWL_THING.equals(typeUri))
continue;
VClass type = vclassDao.getVClassByURI(typeUri);
if( ! type.isAnonymous() &&
type.getName() != null && !"".equals(type.getName()) &&
type.getGroupURI() != null ) //don't display classes that aren't in classgroups
classes.add(type);
}catch(Exception ex){
if( log.isDebugEnabled() )
log.debug("could not add type " + typeUri, ex);
}
}
Collections.sort(classes, new Comparator<VClass>(){
public int compare(VClass o1, VClass o2) {
return o1.compareTo(o2);
}});
return classes;
}
private HashSet<String> getVClassUrisForHits(TopDocs topDocs,
IndexSearcher searcherForRequest){
HashSet<String> typesInHits = new HashSet<String>();
for(int i=0; i<topDocs.scoreDocs.length; i++){
try{
Document doc=searcherForRequest.doc(topDocs.scoreDocs[i].doc);
Field[] types = doc.getFields(Entity2LuceneDoc.term.RDFTYPE);
if(types != null ){
for(int j=0;j<types.length;j++){
String typeUri = types[j].stringValue();
typesInHits.add(typeUri);
}
}
}catch(Exception e){
log.error("problems getting rdf:type for search hits",e);
}
}
return typesInHits;
}
private Analyzer getAnalyzer(ServletContext servletContext) throws SearchException {
Object obj = servletContext.getAttribute(LuceneSetup.ANALYZER);
if( obj == null || !(obj instanceof Analyzer) )
throw new SearchException("Could not get analyzer");
else
return (Analyzer)obj;
}
private Query getQuery(VitroRequest request,
Analyzer analyzer, String querystr ) throws SearchException, ParseException {
Query query = null;
try{
//String querystr = request.getParameter(VitroQuery.QUERY_PARAMETER_NAME);
if( querystr == null){
log.error("There was no Parameter '"+VitroQuery.QUERY_PARAMETER_NAME
+"' in the request.");
return null;
}else if( querystr.length() > MAX_QUERY_LENGTH ){
log.debug("The search was too long. The maximum " +
"query length is " + MAX_QUERY_LENGTH );
return null;
}
log.debug("Parsing query using QueryParser ");
QueryParser parser = getQueryParser(analyzer);
query = parser.parse(querystr);
String alpha = request.getParameter("alpha");
if( alpha != null && !"".equals(alpha) && alpha.length() == 1){
log.debug("Firing alpha query ");
log.debug("request.getParameter(alpha) is " + alpha);
BooleanQuery boolQuery = new BooleanQuery();
boolQuery.add( query, BooleanClause.Occur.MUST );
boolQuery.add(
new WildcardQuery(new Term(Entity2LuceneDoc.term.NAME, alpha+'*')),
BooleanClause.Occur.MUST);
query = boolQuery;
}
//check if this is classgroup filtered
Object param = request.getParameter("classgroup");
if( param != null && !"".equals(param)){
log.debug("Firing classgroup query ");
log.debug("request.getParameter(classgroup) is "+ param.toString());
BooleanQuery boolQuery = new BooleanQuery();
boolQuery.add( query, BooleanClause.Occur.MUST);
boolQuery.add( new TermQuery(
new Term(Entity2LuceneDoc.term.CLASSGROUP_URI,
(String)param)),
BooleanClause.Occur.MUST);
query = boolQuery;
}
//check if this is rdf:type filtered
param = request.getParameter("type");
if( param != null && !"".equals(param)){
log.debug("Firing type query ");
log.debug("request.getParameter(type) is "+ param.toString());
BooleanQuery boolQuery = new BooleanQuery();
boolQuery.add( query, BooleanClause.Occur.MUST);
boolQuery.add( new TermQuery(
new Term(Entity2LuceneDoc.term.RDFTYPE,
(String)param)),
BooleanClause.Occur.MUST);
query = boolQuery;
}
log.debug("Query: " + query);
} catch (ParseException e) {
throw new ParseException(e.getMessage());
} catch (Exception ex){
throw new SearchException(ex.getMessage());
}
return query;
}
@SuppressWarnings("static-access")
private QueryParser getQueryParser(Analyzer analyzer){
//defaultSearchField indicates which field search against when there is no term
//indicated in the query string.
//The analyzer is needed so that we use the same analyzer on the search queries as
//was used on the text that was indexed.
//QueryParser qp = new QueryParser("NAME",analyzer);
//this sets the query parser to AND all of the query terms it finds.
//set up the map of stemmed field names -> unstemmed field names
// HashMap<String,String> map = new HashMap<String, String>();
// map.put(Entity2LuceneDoc.term.ALLTEXT,Entity2LuceneDoc.term.ALLTEXTUNSTEMMED);
// qp.setStemmedToUnstemmed(map);
MultiFieldQueryParser qp = new MultiFieldQueryParser(Version.LUCENE_29, new String[]{
"name", "nameunstemmed", "type", "moniker", "ALLTEXT", "ALLTEXTUNSTEMMED", "nameraw" , "classLocalName", "classLocalNameLowerCase" }, analyzer);
// QueryParser qp = new QueryParser(Version.LUCENE_29, "name", analyzer);
//AND_OPERATOR returns documents even if the terms in the query lie in different fields.
//The only requirement is that they exist in a single document.
//qp.setDefaultOperator(QueryParser.AND_OPERATOR);
return qp;
}
private ExceptionResponseValues doSearchError(Throwable e, Format f) {
Map<String, Object> body = new HashMap<String, Object>();
body.put("message", "Search failed: " + e.getMessage());