Author: ssmiweve
Date: 2008-12-04 14:50:01 +0100 (Thu, 04 Dec 2008)
New Revision: 7038
Modified:
branches/2.18/generic.sesam/query-evaluation/src/main/java/no/sesat/search/query/token/FastQueryMatchingEvaluatorFactory.java
branches/2.18/generic.sesam/query-evaluation/src/main/java/no/sesat/search/query/token/JepTokenEvaluator.java
branches/2.18/generic.sesam/query-evaluation/src/main/java/no/sesat/search/query/token/RegExpTokenEvaluator.java
branches/2.18/generic.sesam/query-evaluation/src/main/java/no/sesat/search/query/token/SolrEvaluatorFactory.java
branches/2.18/generic.sesam/query-evaluation/src/main/java/no/sesat/search/query/token/SolrTokenEvaluator.java
branches/2.18/generic.sesam/query-evaluation/src/main/java/no/sesat/search/query/token/VeryFastTokenEvaluator.java
branches/2.18/query-api/src/main/java/no/sesat/search/query/token/TokenEvaluationEngine.java
branches/2.18/query-api/src/main/java/no/sesat/search/query/token/TokenEvaluator.java
Log:
make the evaluation requests (solr + fast) lazy.
provide timeout value for solr evaluations.
move getMatchValues(..) up into TokenEvaluator
Modified:
branches/2.18/generic.sesam/query-evaluation/src/main/java/no/sesat/search/query/token/FastQueryMatchingEvaluatorFactory.java
===================================================================
---
branches/2.18/generic.sesam/query-evaluation/src/main/java/no/sesat/search/query/token/FastQueryMatchingEvaluatorFactory.java
2008-12-04 13:39:16 UTC (rev 7037)
+++
branches/2.18/generic.sesam/query-evaluation/src/main/java/no/sesat/search/query/token/FastQueryMatchingEvaluatorFactory.java
2008-12-04 13:50:01 UTC (rev 7038)
@@ -114,7 +114,7 @@
try {
// when the root logger is set to DEBUG do not limit connection
times
if(Logger.getRootLogger().getLevel().isGreaterOrEqual(Level.INFO)){
-
+ // default timeout is one second. TODO make configuration.
fastEvaluatorCreator.get(1000, TimeUnit.MILLISECONDS);
}else{
fastEvaluatorCreator.get();
Modified:
branches/2.18/generic.sesam/query-evaluation/src/main/java/no/sesat/search/query/token/JepTokenEvaluator.java
===================================================================
---
branches/2.18/generic.sesam/query-evaluation/src/main/java/no/sesat/search/query/token/JepTokenEvaluator.java
2008-12-04 13:39:16 UTC (rev 7037)
+++
branches/2.18/generic.sesam/query-evaluation/src/main/java/no/sesat/search/query/token/JepTokenEvaluator.java
2008-12-04 13:50:01 UTC (rev 7038)
@@ -22,6 +22,8 @@
package no.sesat.search.query.token;
+import java.util.Collections;
+import java.util.Set;
import org.nfunk.jep.JEP;
import org.nfunk.jep.type.Complex;
@@ -113,6 +115,10 @@
return queryDependant;
}
+ public Set<String> getMatchValues(final TokenPredicate token, final String
term) {
+ return Collections.emptySet();
+ }
+
// Y overrides ---------------------------------------------------
// Package protected ---------------------------------------------
Modified:
branches/2.18/generic.sesam/query-evaluation/src/main/java/no/sesat/search/query/token/RegExpTokenEvaluator.java
===================================================================
---
branches/2.18/generic.sesam/query-evaluation/src/main/java/no/sesat/search/query/token/RegExpTokenEvaluator.java
2008-12-04 13:39:16 UTC (rev 7037)
+++
branches/2.18/generic.sesam/query-evaluation/src/main/java/no/sesat/search/query/token/RegExpTokenEvaluator.java
2008-12-04 13:50:01 UTC (rev 7038)
@@ -18,6 +18,8 @@
import java.util.ArrayList;
import java.util.Collection;
+import java.util.Collections;
+import java.util.Set;
import java.util.regex.Pattern;
import java.util.regex.Matcher;
import org.apache.log4j.Logger;
@@ -81,4 +83,9 @@
public boolean isQueryDependant(final TokenPredicate predicate) {
return queryDependant;
}
+
+ public Set<String> getMatchValues(final TokenPredicate token, final String
term) {
+ return Collections.emptySet();
+ }
+
}
Modified:
branches/2.18/generic.sesam/query-evaluation/src/main/java/no/sesat/search/query/token/SolrEvaluatorFactory.java
===================================================================
---
branches/2.18/generic.sesam/query-evaluation/src/main/java/no/sesat/search/query/token/SolrEvaluatorFactory.java
2008-12-04 13:39:16 UTC (rev 7037)
+++
branches/2.18/generic.sesam/query-evaluation/src/main/java/no/sesat/search/query/token/SolrEvaluatorFactory.java
2008-12-04 13:50:01 UTC (rev 7038)
@@ -289,7 +289,7 @@
// when the root logger is set to DEBUG do not limit connection
times
if(Logger.getRootLogger().getLevel().isGreaterOrEqual(Level.INFO)){
-
+ // default timeout is one second. TODO make configuration.
solrEvaluatorCreator.get(1000, TimeUnit.MILLISECONDS);
}else{
solrEvaluatorCreator.get();
Modified:
branches/2.18/generic.sesam/query-evaluation/src/main/java/no/sesat/search/query/token/SolrTokenEvaluator.java
===================================================================
---
branches/2.18/generic.sesam/query-evaluation/src/main/java/no/sesat/search/query/token/SolrTokenEvaluator.java
2008-12-04 13:39:16 UTC (rev 7037)
+++
branches/2.18/generic.sesam/query-evaluation/src/main/java/no/sesat/search/query/token/SolrTokenEvaluator.java
2008-12-04 13:50:01 UTC (rev 7038)
@@ -31,6 +31,7 @@
import java.util.regex.Matcher;
import java.util.regex.Pattern;
import no.sesat.search.query.token.AbstractEvaluatorFactory.Context;
+import org.apache.log4j.Level;
import static no.sesat.search.query.parser.AbstractQueryParser.SKIP_REGEX;
import static no.sesat.search.query.parser.AbstractQueryParser.OPERATOR_REGEX;
import org.apache.log4j.Logger;
@@ -69,7 +70,7 @@
private final Context context;
private SolrEvaluatorFactory factory;
- private final Map<String, List<TokenMatch>> analysisResult;
+ private transient Map<String, List<TokenMatch>> analysisResult;
// Static --------------------------------------------------------
@@ -85,13 +86,11 @@
* @param factory
* @throws EvaluationException
*/
- public SolrTokenEvaluator(final Context cxt, final SolrEvaluatorFactory
factory) throws EvaluationException{
+ SolrTokenEvaluator(final Context cxt, final SolrEvaluatorFactory factory)
throws EvaluationException{
context = cxt;
this.factory = factory;
- // Remove whitespace (except space itself) and operator characters.
- analysisResult = query(cleanString(cxt.getQueryString()));
}
// Public --------------------------------------------------------
@@ -107,6 +106,8 @@
final String listname = listnames[i];
+ if(null == analysisResult){ analysisResult =
query(cleanString(context.getQueryString())); }
+
if (analysisResult.containsKey(listname)) {
if (term == null) {
evaluation = true;
@@ -150,6 +151,9 @@
if(null != listnames){
for(int i = 0; i < listnames.length; i++){
final String listname = listnames[i];
+
+ if(null == analysisResult){ analysisResult =
query(cleanString(context.getQueryString())); }
+
if (analysisResult.containsKey(listname)) {
// HACK since DefaultOperatorClause wraps its children in
parenthesis
@@ -190,7 +194,7 @@
* @param query
*/
@SuppressWarnings("unchecked")
- private Map<String, List<TokenMatch>> query(final String query) throws
EvaluationException{
+ private synchronized Map<String, List<TokenMatch>> query(final String
query){
LOG.trace("queryFast( " + query + " )");
Map<String, List<TokenMatch>> result = null;
@@ -214,6 +218,12 @@
.setQuery("list_entry_shingle:\"" + token + "\"")
.setRows(Integer.MAX_VALUE);
+ // when the root logger is set to DEBUG do not limit
connection times
+
if(Logger.getRootLogger().getLevel().isGreaterOrEqual(Level.INFO)){
+ // default timeout is half second. TODO make
configuration.
+ solrQuery.setTimeAllowed(500);
+ }
+
DUMP.info(solrQuery.toString());
// query
@@ -250,7 +260,7 @@
} catch (SolrServerException ex) {
LOG.error(ex.getMessage(), ex);
- throw new EvaluationException(ERR_QUERY_FAILED + url, ex);
+ throw new EvaluationRuntimeException(new
EvaluationException(ERR_QUERY_FAILED + url, ex));
}finally{
if(!updatedCache){
Modified:
branches/2.18/generic.sesam/query-evaluation/src/main/java/no/sesat/search/query/token/VeryFastTokenEvaluator.java
===================================================================
---
branches/2.18/generic.sesam/query-evaluation/src/main/java/no/sesat/search/query/token/VeryFastTokenEvaluator.java
2008-12-04 13:39:16 UTC (rev 7037)
+++
branches/2.18/generic.sesam/query-evaluation/src/main/java/no/sesat/search/query/token/VeryFastTokenEvaluator.java
2008-12-04 13:50:01 UTC (rev 7038)
@@ -105,7 +105,7 @@
private final Context context;
private final Site site;
- private final Map<String, List<TokenMatch>> analysisResult;
+ private transient Map<String, List<TokenMatch>> analysisResult;
// Static --------------------------------------------------------
@@ -128,27 +128,10 @@
init();
- // Remove whitespace (except space itself) and operator characters.
- analysisResult = queryFast(cleanString(context.getQueryString()));
-
}
// Public --------------------------------------------------------
- /**
- * Find out if given token is on or more of the following.
- * <li>GEO
- * <li>FIRSTNAME
- * <li>LASTNAME
- * <li>COMPANY
- * <li>KEYWORDS
- * <li>CATEGORY
- * </ul>
- *
- * @param token can be any of the above
- * @param query
- * @return true if the query contains any of the above
- */
public boolean evaluateToken(final TokenPredicate token, final String
term, final String query) {
boolean evaluation = false;
@@ -159,6 +142,8 @@
final String listname = listnames[i];
+ if(null == analysisResult){ analysisResult =
queryFast(cleanString(context.getQueryString())); }
+
if (analysisResult.containsKey(listname)) {
if (term == null) {
evaluation = true;
@@ -186,14 +171,6 @@
return evaluation;
}
-
- /**
- * get all match values and values for given Fast list .
- *
- * @param token
- * @param term
- * @return a list of Tokens
- */
public Set<String> getMatchValues(final TokenPredicate token, final String
term) {
final Set<String> values = new HashSet<String>();
@@ -202,6 +179,9 @@
if(null != listnames){
for(int i = 0; i < listnames.length; i++){
final String listname = listnames[i];
+
+ if(null == analysisResult){ analysisResult =
queryFast(cleanString(context.getQueryString())); }
+
if (analysisResult.containsKey(listname)) {
// HACK since DefaultOperatorClause wraps its children in
parenthesis
@@ -221,11 +201,6 @@
return Collections.unmodifiableSet(values);
}
- /**
- *
- * @param predicate
- * @return
- */
public boolean isQueryDependant(final TokenPredicate predicate) {
return predicate.name().startsWith(EXACT_PREFIX.toUpperCase());
}
@@ -346,7 +321,7 @@
* @param query
*/
@SuppressWarnings("unchecked")
- private Map<String, List<TokenMatch>> queryFast(final String query) throws
EvaluationException{
+ private synchronized Map<String, List<TokenMatch>> queryFast(final String
query){
LOG.trace("queryFast( " + query + " )");
Map<String, List<TokenMatch>> result = null;
@@ -426,11 +401,11 @@
} catch (IOException e1) {
LOG.error(ERR_QUERY_FAILED + url, e1);
result = (Map<String,
List<TokenMatch>>)nre.getCacheContent();
- throw new EvaluationException(ERR_QUERY_FAILED + url, e1);
+ throw new EvaluationRuntimeException(new
EvaluationException(ERR_QUERY_FAILED + url, e1));
} catch (SAXException e1) {
LOG.error(ERR_PARSE_FAILED + url, e1);
result = (Map<String,
List<TokenMatch>>)nre.getCacheContent();
- throw new EvaluationException(ERR_PARSE_FAILED + url, e1);
+ throw new EvaluationRuntimeException(new
EvaluationException(ERR_PARSE_FAILED + url, e1));
}finally{
if(!updatedCache){
CACHE_QUERY.cancelUpdate(query);
Modified:
branches/2.18/query-api/src/main/java/no/sesat/search/query/token/TokenEvaluationEngine.java
===================================================================
---
branches/2.18/query-api/src/main/java/no/sesat/search/query/token/TokenEvaluationEngine.java
2008-12-04 13:39:16 UTC (rev 7037)
+++
branches/2.18/query-api/src/main/java/no/sesat/search/query/token/TokenEvaluationEngine.java
2008-12-04 13:50:01 UTC (rev 7038)
@@ -18,6 +18,7 @@
package no.sesat.search.query.token;
import java.io.Serializable;
+import java.util.Collections;
import java.util.Set;
import no.sesat.commons.ioc.BaseContext;
import no.sesat.search.query.Clause;
@@ -58,10 +59,12 @@
public boolean evaluateToken(final TokenPredicate token, final String
term, final String query) {
return false;
}
-
public boolean isQueryDependant(final TokenPredicate predicate) {
return false;
}
+ public Set<String> getMatchValues(final TokenPredicate token, final
String term) {
+ return Collections.emptySet();
+ }
};
/**
@@ -71,10 +74,12 @@
public boolean evaluateToken(final TokenPredicate token, final String
term, final String query) {
return true;
}
-
public boolean isQueryDependant(final TokenPredicate predicate) {
return false;
}
+ public Set<String> getMatchValues(final TokenPredicate token, final
String term) {
+ return Collections.emptySet();
+ }
};
/**
@@ -88,6 +93,9 @@
public boolean isQueryDependant(final TokenPredicate predicate) {
return false;
}
+ public Set<String> getMatchValues(final TokenPredicate token, final
String term) {
+ return Collections.emptySet();
+ }
};
/**
Modified:
branches/2.18/query-api/src/main/java/no/sesat/search/query/token/TokenEvaluator.java
===================================================================
---
branches/2.18/query-api/src/main/java/no/sesat/search/query/token/TokenEvaluator.java
2008-12-04 13:39:16 UTC (rev 7037)
+++
branches/2.18/query-api/src/main/java/no/sesat/search/query/token/TokenEvaluator.java
2008-12-04 13:50:01 UTC (rev 7038)
@@ -1,5 +1,5 @@
/*
- * Copyright (2005-2007) Schibsted Søk AS
+ * Copyright (2005-2008) Schibsted Søk AS
* This file is part of SESAT.
*
* SESAT is free software: you can redistribute it and/or modify
@@ -16,13 +16,22 @@
* along with SESAT. If not, see <http://www.gnu.org/licenses/>.
*/
package no.sesat.search.query.token;
+import java.util.Set;
-/**
+/** TokenEvaluator works behind a TokenPredicate doing the actual work to
prove the predicate true or false.
+ * This work is the evaluating process. Each TokenValuator is constructed and
obtained through a corresponding
+ * AbstractEvaluatorFactory.
*
+ * evaluateToken(TokenPredicate, String, String) is the primary method in the
class,
+ * but evaluators may also indicate whether the position of words within the
query influences evaluation
+ * by the isQueryDependant(TokenPredicate) method.
*
- * @version <tt>$Revision$</tt>
+ * And synonyms may also be provided, extending predicates from true|false to
being a richer meta data source.
+ *
+ *
+ * @version <tt>$Id$</tt>
*/
public interface TokenEvaluator {
@@ -45,4 +54,12 @@
* @return
*/
boolean isQueryDependant(TokenPredicate predicate);
+
+ /** Each true evaluation may also provide a synonym (or "match value").
+ *
+ * @param token
+ * @param term
+ * @return a list of Tokens
+ */
+ Set<String> getMatchValues(final TokenPredicate token, final String term);
}
_______________________________________________
Kernel-commits mailing list
[email protected]
http://sesat.no/mailman/listinfo/kernel-commits
