SOLR-1071: spellcheck.extendedResults format change

git-svn-id: https://svn.apache.org/repos/asf/lucene/solr/trunk@812714 13f79535-47bb-0310-9956-ffa450edef68
This commit is contained in:
Yonik Seeley 2009-09-08 22:22:33 +00:00
parent 19e0a7a794
commit b3ffddf116
5 changed files with 101 additions and 94 deletions

View File

@ -47,6 +47,9 @@ a TokenStream (that may be or may not be a Tokenizer). This change is required
to take advantage of the Token reuse improvements in lucene 2.9. For more
information, see SOLR-1377.
If spellcheck.extendedResults=true, the response format for suggestions
has changed, see SOLR-1071.
Versions of Major Components
----------------------------
@ -535,6 +538,11 @@ Bug Fixes
64. SOLR-1400: Properly handle zero-length tokens in TrimFilter. This
was not a bug in any released version. (Peter Wolanin, gsingers)
65. SOLR-1071: spellcheck.extendedResults returns an invalid JSON response
when count > 1. To fix, the extendedResults format was changed.
(Uri Boness, yonik)
Other Changes
----------------------
1. Upgraded to Lucene 2.4.0 (yonik)

View File

@ -201,17 +201,30 @@ public class SpellCheckComponent extends SearchComponent implements SolrCoreAwar
suggestionList.add("numFound", theSuggestions.size());
suggestionList.add("startOffset", inputToken.startOffset());
suggestionList.add("endOffset", inputToken.endOffset());
// Logical structure of normal (non-extended) results:
// "suggestion":["alt1","alt2"]
//
// Logical structure of the extended results:
// "suggestion":[
// {"word":"alt1","freq":7},
// {"word":"alt2","freq":4}
// ]
if (extendedResults && hasFreqInfo) {
suggestionList.add("origFreq", spellingResult.getTokenFrequency(inputToken));
ArrayList<SimpleOrderedMap> sugs = new ArrayList<SimpleOrderedMap>();
suggestionList.add("suggestion", sugs);
for (Map.Entry<String, Integer> suggEntry : theSuggestions.entrySet()) {
SimpleOrderedMap<Object> suggestionItem = new SimpleOrderedMap<Object>();
suggestionItem.add("frequency", suggEntry.getValue());
suggestionItem.add("word", suggEntry.getKey());
suggestionList.add("suggestion", suggestionItem);
SimpleOrderedMap sugEntry = new SimpleOrderedMap();
sugEntry.add("word",suggEntry.getKey());
sugEntry.add("freq",suggEntry.getValue());
sugs.add(sugEntry);
}
} else {
suggestionList.add("suggestion", theSuggestions.keySet());
}
if (collate == true ){//set aside the best suggestion for this token
best.put(inputToken, theSuggestions.keySet().iterator().next());
}

View File

@ -17,7 +17,6 @@ package org.apache.solr.client.solrj.response;
*/
import org.apache.solr.common.util.NamedList;
import org.apache.solr.common.util.SimpleOrderedMap;
import java.util.ArrayList;
import java.util.LinkedHashMap;
@ -74,10 +73,8 @@ public class SpellCheckResponse {
public String getFirstSuggestion(String token) {
Suggestion s = suggestionMap.get(token);
if (s != null) {
return s.getSuggestions().isEmpty() ? null : s.getSuggestions().get(0);
}
return null;
if (s==null || s.getAlternatives().isEmpty()) return null;
return s.getAlternatives().get(0);
}
public String getCollatedResult() {
@ -90,8 +87,8 @@ public class SpellCheckResponse {
private int startOffset;
private int endOffset;
private int originalFrequency;
private List<String> suggestions = new ArrayList<String>();
private List<Integer> suggestionFrequencies = new ArrayList<Integer>();
private List<String> alternatives = new ArrayList<String>();
private List<Integer> alternativeFrequencies;
public Suggestion(String token, NamedList<Object> suggestion) {
this.token = token;
@ -107,14 +104,16 @@ public class SpellCheckResponse {
} else if ("origFreq".equals(n)) {
originalFrequency = (Integer) suggestion.getVal(i);
} else if ("suggestion".equals(n)) {
Object o = suggestion.getVal(i);
if (o instanceof List) {
List<String> list = (List<String>) o;
suggestions.addAll(list);
} else if (o instanceof SimpleOrderedMap) {
SimpleOrderedMap map = (SimpleOrderedMap) o;
suggestions.add((String) map.get("word"));
suggestionFrequencies.add((Integer) map.get("frequency"));
List list = (List)suggestion.getVal(i);
if (list.size() > 0 && list.get(0) instanceof NamedList) {
// extended results detected
alternativeFrequencies = new ArrayList<Integer>();
for (NamedList nl : (List<NamedList>)list) {
alternatives.add((String)nl.get("word"));
alternativeFrequencies.add((Integer)nl.get("freq"));
}
} else {
alternatives.addAll(list);
}
}
}
@ -140,12 +139,27 @@ public class SpellCheckResponse {
return originalFrequency;
}
public List<String> getSuggestions() {
return suggestions;
/** The list of alternatives */
public List<String> getAlternatives() {
return alternatives;
}
public List<Integer> getSuggestionFrequencies() {
return suggestionFrequencies;
/** The frequencies of the alternatives in the corpus, or null if this information was not returned */
public List<Integer> getAlternativeFrequencies() {
return alternativeFrequencies;
}
@Deprecated
/** @see #getAlternatives */
public List<String> getSuggestions() {
return alternatives;
}
@Deprecated
/** @see #getAlternativeFrequencies */
public List<Integer> getSuggestionFrequencies() {
return alternativeFrequencies;
}
}
}

View File

@ -27,6 +27,8 @@ import org.apache.solr.common.SolrInputDocument;
import org.apache.solr.common.params.CommonParams;
import org.apache.solr.common.params.SpellingParams;
import java.util.List;
/**
* Test for SpellCheckComponent's response in Solrj
*
@ -67,7 +69,7 @@ public class TestSpellCheckResponse extends SolrExampleTestBase {
query.set(SpellingParams.SPELLCHECK_BUILD, true);
QueryRequest request = new QueryRequest(query);
SpellCheckResponse response = request.process(server).getSpellCheckResponse();
Assert.assertEquals("Incorrect spelling results", "samsung", response.getFirstSuggestion("samsang"));
Assert.assertEquals("samsung", response.getFirstSuggestion("samsang"));
}
public void testSpellCheckResponse_Extended() throws Exception {
@ -85,7 +87,25 @@ public class TestSpellCheckResponse extends SolrExampleTestBase {
query.set(SpellingParams.SPELLCHECK_EXTENDED_RESULTS, true);
QueryRequest request = new QueryRequest(query);
SpellCheckResponse response = request.process(server).getSpellCheckResponse();
Assert.assertEquals("Incorrect spelling results", "samsung", response.getFirstSuggestion("samsang"));
assertEquals("samsung", response.getFirstSuggestion("samsang"));
SpellCheckResponse.Suggestion sug = response.getSuggestion("samsang");
List<SpellCheckResponse.Suggestion> sugs = response.getSuggestions();
assertEquals(sug.getAlternatives().size(), sug.getAlternativeFrequencies().size());
assertEquals(sugs.get(0).getAlternatives().size(), sugs.get(0).getAlternativeFrequencies().size());
assertEquals("samsung", sug.getAlternatives().get(0));
assertEquals("samsung", sugs.get(0).getAlternatives().get(0));
// basic test if fields were filled in
assertTrue(sug.getEndOffset()>0);
assertTrue(sug.getToken().length() > 0);
assertTrue(sug.getNumFound() > 0);
// assertTrue(sug.getOriginalFrequency() > 0);
// Hmmm... the API for SpellCheckResponse could be nicer:
response.getSuggestions().get(0).getAlternatives().get(0);
}
protected SolrServer getSolrServer() {

View File

@ -18,10 +18,7 @@
package org.apache.solr.handler.component;
import java.io.File;
import java.util.ArrayList;
import java.util.Collection;
import java.util.HashMap;
import java.util.Map;
import java.util.*;
import org.apache.solr.common.params.CommonParams;
import org.apache.solr.common.params.MapSolrParams;
@ -86,26 +83,19 @@ public class SpellCheckComponentTest extends AbstractSolrTestCase {
handler.handleRequest(new LocalSolrQueryRequest(core, params), rsp);
NamedList values = rsp.getValues();
String cmdExec = (String) values.get("command");
assertTrue("command is null and it shouldn't be", cmdExec != null);
assertTrue(cmdExec + " is not equal to " + "build",
cmdExec.equals("build") == true);
assertEquals("build",cmdExec);
NamedList spellCheck = (NamedList) values.get("spellcheck");
assertTrue("spellCheck is null and it shouldn't be", spellCheck != null);
NamedList suggestions = (NamedList) spellCheck.get("suggestions");
assertTrue("suggestions is null and it shouldn't be", suggestions != null);
NamedList blue = (NamedList) suggestions.get("bluo");
assertTrue(blue.get("numFound") + " is not equal to " + "5", blue
.get("numFound").toString().equals("5") == true);
assertEquals(5,blue.get("numFound"));
Collection<String> theSuggestion = (Collection<String>) blue.get("suggestion");
assertTrue("theSuggestion is null and it shouldn't be: " + blue,
theSuggestion != null);
assertTrue("theSuggestion Size: " + theSuggestion.size() + " is not: " + 5,
theSuggestion.size() == 5);
assertEquals(5,theSuggestion.size());
//we know there are at least 5, but now only get 3
params.remove(SpellCheckComponent.SPELLCHECK_COUNT);
params.remove(SpellCheckComponent.SPELLCHECK_EXTENDED_RESULTS);
params.remove(SpellCheckComponent.SPELLCHECK_BUILD);
params.add(SpellCheckComponent.SPELLCHECK_COUNT, String.valueOf(3));
params.add(SpellCheckComponent.SPELLCHECK_COUNT, "3");
params.add(SpellCheckComponent.SPELLCHECK_EXTENDED_RESULTS, String.valueOf(true));
params.add(SpellCheckComponent.SPELLCHECK_BUILD, "false");
rsp = new SolrQueryResponse();
@ -113,36 +103,17 @@ public class SpellCheckComponentTest extends AbstractSolrTestCase {
values = rsp.getValues();
spellCheck = (NamedList) values.get("spellcheck");
assertTrue("spellCheck is null and it shouldn't be", spellCheck != null);
suggestions = (NamedList) spellCheck.get("suggestions");
assertTrue("suggestions is null and it shouldn't be", suggestions != null);
blue = (NamedList) suggestions.get("bluo");
assertTrue(blue.get("numFound") + " is not equal to " + "3", blue
.get("numFound").toString().equals("3") == true);
SimpleOrderedMap theSuggestions;
int idx = blue.indexOf("suggestion", 0);
theSuggestions = (SimpleOrderedMap) blue.get("suggestion", idx);
assertTrue("theSuggestion is null and it shouldn't be: " + blue,
theSuggestions != null);
assertTrue("theSuggestions Size: " + theSuggestions.size() + " is not: " + 2,
theSuggestions.size() == 2);//the word and the frequency
assertEquals(3, blue.get("numFound"));
idx = blue.indexOf("suggestion", idx + 1);
theSuggestions = (SimpleOrderedMap) blue.get("suggestion", idx);
assertTrue("theSuggestion is null and it shouldn't be: " + blue,
theSuggestions != null);
assertTrue("theSuggestions Size: " + theSuggestions.size() + " is not: " + 2,
theSuggestions.size() == 2);//the word and the frequency
List<SimpleOrderedMap> theSuggestions = (List<SimpleOrderedMap>)blue.get("suggestion");
assertEquals(3, theSuggestions.size());
idx = blue.indexOf("suggestion", idx + 1);
theSuggestions = (SimpleOrderedMap) blue.get("suggestion", idx);
assertTrue("theSuggestion is null and it shouldn't be: " + blue,
theSuggestions != null);
assertTrue("theSuggestions Size: " + theSuggestions.size() + " is not: " + 2,
theSuggestions.size() == 2);//the word and the frequency
idx = blue.indexOf("suggestion", idx + 1);
assertTrue(idx + " does not equal: " + -1, idx == -1);
for (SimpleOrderedMap sug : theSuggestions) {
assertNotNull(sug.get("word"));
assertNotNull(sug.get("freq"));
}
}
public void test() throws Exception {
@ -169,19 +140,12 @@ public class SpellCheckComponentTest extends AbstractSolrTestCase {
NamedList suggestions = (NamedList) spellCheck.get("suggestions");
assertTrue("suggestions is null and it shouldn't be", suggestions != null);
NamedList document = (NamedList) suggestions.get("documemt");
assertTrue(document.get("numFound") + " is not equal to " + "1", document
.get("numFound").toString().equals("1") == true);
assertTrue(document.get("startOffset") + " is not equal to " + "0", document
.get("startOffset").toString().equals("0") == true);
assertTrue(document.get("endOffset") + " is not equal to " + "documemt".length(), document
.get("endOffset").toString().equals(String.valueOf("documemt".length())) == true);
assertEquals(1, document.get("numFound"));
assertEquals(0, document.get("startOffset"));
assertEquals(document.get("endOffset"), "documemt".length());
Collection<String> theSuggestion = (Collection<String>) document.get("suggestion");
assertTrue("theSuggestion is null and it shouldn't be: " + document,
theSuggestion != null);
assertTrue("theSuggestion Size: " + theSuggestion.size() + " is not: " + 1,
theSuggestion.size() == 1);
assertTrue(theSuggestion.iterator().next() + " is not equal to " + "document", theSuggestion.iterator().next().equals("document") == true);
assertEquals(1, theSuggestion.size());
assertEquals("document", theSuggestion.iterator().next());
}
@ -203,12 +167,9 @@ public class SpellCheckComponentTest extends AbstractSolrTestCase {
handler.handleRequest(new LocalSolrQueryRequest(core, params), rsp);
NamedList values = rsp.getValues();
NamedList spellCheck = (NamedList) values.get("spellcheck");
assertTrue("spellCheck is null and it shouldn't be", spellCheck != null);
NamedList suggestions = (NamedList) spellCheck.get("suggestions");
assertTrue("suggestions is null and it shouldn't be", suggestions != null);
String collation = (String) suggestions.get("collation");
assertTrue("collation is null and it shouldn't be", collation != null);
assertTrue(collation + " is not equal to " + "document", collation.equals("document") == true);
assertEquals("document", collation);
params.remove(CommonParams.Q);
params.add(CommonParams.Q, "documemt lowerfilt:broen^4");
handler = core.getRequestHandler("spellCheckCompRH");
@ -217,12 +178,9 @@ public class SpellCheckComponentTest extends AbstractSolrTestCase {
handler.handleRequest(new LocalSolrQueryRequest(core, params), rsp);
values = rsp.getValues();
spellCheck = (NamedList) values.get("spellcheck");
assertTrue("spellCheck is null and it shouldn't be", spellCheck != null);
suggestions = (NamedList) spellCheck.get("suggestions");
assertTrue("suggestions is null and it shouldn't be", suggestions != null);
collation = (String) suggestions.get("collation");
assertTrue("collation is null and it shouldn't be", collation != null);
assertTrue(collation + " is not equal to " + "document lowerfilt:brown^4", collation.equals("document lowerfilt:brown^4") == true);
assertEquals("document lowerfilt:brown^4", collation);
params.remove(CommonParams.Q);
params.add(CommonParams.Q, "documemtsss broens");
@ -232,15 +190,9 @@ public class SpellCheckComponentTest extends AbstractSolrTestCase {
handler.handleRequest(new LocalSolrQueryRequest(core, params), rsp);
values = rsp.getValues();
spellCheck = (NamedList) values.get("spellcheck");
assertTrue("spellCheck is null and it shouldn't be", spellCheck != null);
suggestions = (NamedList) spellCheck.get("suggestions");
assertTrue("suggestions is null and it shouldn't be", suggestions != null);
collation = (String) suggestions.get("collation");
assertTrue("collation is null and it shouldn't be", collation != null);
System.out.println("Collation: " + collation);
assertTrue(collation + " is not equal to " + "document brown", collation.equals("document brown") == true);
assertEquals("document brown",collation);
}
public void testCorrectSpelling() throws Exception {