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

make the kNN clustering report the right counts for the facet values (and order them in the clusters by counts)

Browse Source 
git-svn-id: http://google-refine.googlecode.com/svn/trunk@286 7d457c2a-affb-35e4-300a-418c747d4874
This commit is contained in:
Stefano Mazzocchi 2010-03-12 19:10:22 +00:00
parent d72c07b715
commit 00a81c5fc4
1 changed files with 30 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

33
src/main/java/com/metaweb/gridworks/clustering/knn/kNNClusterer.java
View File

@ -11,6 +11,7 @@ import java.util.Map;
import java.util.Properties; import java.util.Properties;
import java.util.Set; import java.util.Set;
import java.util.TreeSet; import java.util.TreeSet;
import java.util.Map.Entry;
import org.json.JSONException; import org.json.JSONException;
import org.json.JSONObject; import org.json.JSONObject;
@ -44,7 +45,9 @@ public class kNNClusterer extends Clusterer {
static protected Map<String, Distance> _distances = new HashMap<String, Distance>(); static protected Map<String, Distance> _distances = new HashMap<String, Distance>();
ArrayList<Set<Serializable>> _clusters; List<Set<Serializable>> _clusters;
Map<Serializable, Integer> _counts = new HashMap<Serializable, Integer>();
static { static {
_distances.put("levenshtein", new LevenshteinDistance()); _distances.put("levenshtein", new LevenshteinDistance());
@ -82,6 +85,7 @@ public class kNNClusterer extends Clusterer {
Object v = cell.value; Object v = cell.value;
String s = (v instanceof String) ? ((String) v) : v.toString(); String s = (v instanceof String) ? ((String) v) : v.toString();
_treeBuilder.populate(s); _treeBuilder.populate(s);
count(s);
} }
return false; return false;
} }
@ -120,6 +124,7 @@ public class kNNClusterer extends Clusterer {
Object v = cell.value; Object v = cell.value;
String s = (v instanceof String) ? ((String) v) : v.toString().intern(); String s = (v instanceof String) ? ((String) v) : v.toString().intern();
_data.add(s); _data.add(s);
count(s);
} }
return false; return false;
} }
@ -184,6 +189,12 @@ public class kNNClusterer extends Clusterer {
} }
} }
public class ValuesComparator implements Comparator<Entry<Serializable,Integer>> {
public int compare(Entry<Serializable,Integer> o1, Entry<Serializable,Integer> o2) {
return o2.getValue() - o1.getValue();
}
}
public void initializeFromJSON(Project project, JSONObject o) throws Exception { public void initializeFromJSON(Project project, JSONObject o) throws Exception {
super.initializeFromJSON(project, o); super.initializeFromJSON(project, o);
_distance = _distances.get(o.getString("function").toLowerCase()); _distance = _distances.get(o.getString("function").toLowerCase());
@ -204,11 +215,17 @@ public class kNNClusterer extends Clusterer {
writer.array(); writer.array();
for (Set<Serializable> m : _clusters) { for (Set<Serializable> m : _clusters) {
if (m.size() > 1) { if (m.size() > 1) {
writer.array(); Map<Serializable,Integer> internal_counts = new HashMap<Serializable,Integer>();
for (Serializable s : m) { for (Serializable s : m) {
internal_counts.put(s,_counts.get(s));
}
List<Entry<Serializable,Integer>> values = new ArrayList<Entry<Serializable,Integer>>(internal_counts.entrySet());
Collections.sort(values, new ValuesComparator());
writer.array();
for (Entry<Serializable,Integer> e : values) {
writer.object(); writer.object();
writer.key("v"); writer.value(s); writer.key("v"); writer.value(e.getKey());
writer.key("c"); writer.value(1); writer.key("c"); writer.value(e.getValue());
writer.endObject(); writer.endObject();
} }
writer.endArray(); writer.endArray();
@ -216,4 +233,12 @@ public class kNNClusterer extends Clusterer {
} }
writer.endArray(); writer.endArray();
} }
private void count(Serializable s) {
if (_counts.containsKey(s)) {
_counts.put(s, _counts.get(s) + 1);
} else {
_counts.put(s, 1);
}
}
} }