s434730/RandomSec
3
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Eliminate hash maps from recon objects--they are expensive.

Browse Source 
git-svn-id: http://google-refine.googlecode.com/svn/trunk@88 7d457c2a-affb-35e4-300a-418c747d4874
This commit is contained in:
David Huynh 2010-02-10 19:25:21 +00:00
parent 5cd147ea3c
commit d227db0cc6
2 changed files with 50 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

53
src/main/java/com/metaweb/gridworks/model/Recon.java
View File

@ -22,20 +22,60 @@ public class Recon implements Serializable, HasFields, Jsonizable {
New
}
public Map<String, Object> features = new HashMap<String, Object>();
static public int Feature_typeMatch = 0;
static public int Feature_nameMatch = 1;
static public int Feature_nameLevenshtein = 2;
static public int Feature_nameWordDistance = 3;
static public int Feature_max = 4;
static protected Map<String, Integer> s_featureMap;
static {
s_featureMap = new HashMap<String, Integer>();
s_featureMap.put("typeMatch", Feature_typeMatch);
s_featureMap.put("nameMatch", Feature_nameMatch);
s_featureMap.put("nameLevenshtein", Feature_nameLevenshtein);
s_featureMap.put("nameWordDistance", Feature_nameWordDistance);
}
public Object[] features = new Object[Feature_max];
public List<ReconCandidate> candidates = new LinkedList<ReconCandidate>();
public Judgment judgment = Judgment.None;
public ReconCandidate match = null;
public Recon dup() {
Recon r = new Recon();
r.features.putAll(new HashMap<String, Object>(features));
System.arraycopy(features, 0, r.features, 0, features.length);
r.candidates.addAll(candidates);
r.judgment = judgment;
r.match = match;
return r;
}
public Object getFeature(int feature) {
return feature < features.length ? features[feature] : null;
}
public void setFeature(int feature, Object v) {
if (feature >= features.length) {
if (feature >= Feature_max) {
return;
}
// We deserialized this object from an older version of the class
// that had fewer features, so we can just try to extend it
Object[] newFeatures = new Object[Feature_max];
System.arraycopy(features, 0, newFeatures, 0, features.length);
features = newFeatures;
}
features[feature] = v;
}
public Object getField(String name, Properties bindings) {
if ("best".equals(name)) {
return candidates.size() > 0 ? candidates.get(0) : null;
@ -48,11 +88,7 @@ public class Recon implements Serializable, HasFields, Jsonizable {
} else if ("match".equals(name)) {
return match;
} else if ("features".equals(name)) {
return new HasFields() {
public Object getField(String name, Properties bindings) {
return features.get(name);
}
};
return new Features();
}
return null;
}
@ -69,7 +105,8 @@ public class Recon implements Serializable, HasFields, Jsonizable {
public class Features implements HasFields {
public Object getField(String name, Properties bindings) {
return features.get(name);
int index = s_featureMap.get(name);
return index < features.length ? features[index] : null;
}
}

10
src/main/java/com/metaweb/gridworks/model/operations/ReconOperation.java
View File

@ -317,14 +317,14 @@ public class ReconOperation extends EngineDependentOperation {
// best match
if (i == 0) {
recon.features.put("nameMatch", text.equalsIgnoreCase(candidate.topicName));
recon.features.put("nameLevenshtein", StringUtils.getLevenshteinDistance(text, candidate.topicName));
recon.features.put("nameWordDistance", wordDistance(text, candidate.topicName));
recon.setFeature(Recon.Feature_nameMatch, text.equalsIgnoreCase(candidate.topicName));
recon.setFeature(Recon.Feature_nameLevenshtein, StringUtils.getLevenshteinDistance(text, candidate.topicName));
recon.setFeature(Recon.Feature_nameWordDistance, wordDistance(text, candidate.topicName));
recon.features.put("typeMatch", false);
recon.setFeature(Recon.Feature_typeMatch, false);
for (String typeID : candidate.typeIDs) {
if (_typeID.equals(typeID)) {
recon.features.put("typeMatch", true);
recon.setFeature(Recon.Feature_typeMatch, true);
break;
}
}