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

Merge pull request #1925 from OpenRefine/issue1887

Browse Source 
Compute reconciliation features when creating new items
This commit is contained in:
Antonin Delpeuch 2019-01-10 04:44:45 +01:00 committed by GitHub
commit 6dd9f41639
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
3 changed files with 50 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

23
extensions/wikidata/src/org/openrefine/wikidata/editing/NewItemLibrary.java
View File

@ -37,6 +37,8 @@ import com.google.refine.model.Recon;
import com.google.refine.model.ReconCandidate;
import com.google.refine.model.ReconStats;
import com.google.refine.model.Row;
import com.google.refine.model.recon.ReconConfig;
import com.google.refine.model.recon.StandardReconConfig;
/**
* This keeps track of the new items that we have created for each internal
@ -106,17 +108,34 @@ public class NewItemLibrary {
continue;
}
Recon recon = cell.recon;
boolean changed = false;
if (Recon.Judgment.New.equals(recon.judgment) && !reset
&& map.containsKey(recon.id)) {
recon.judgment = Recon.Judgment.Matched;
recon.match = new ReconCandidate(map.get(recon.id), cell.value.toString(),
new String[0], 100);
impactedColumns.add(i);
recon.addCandidate(recon.match);
changed = true;
} else if (Recon.Judgment.Matched.equals(recon.judgment) && reset
&& map.containsKey(recon.id)) {
recon.judgment = Recon.Judgment.New;
if(recon.candidates != null) {
recon.candidates.remove(recon.candidates.size()-1);
}
recon.match = null;
impactedColumns.add(i);
changed = true;
}
if (changed) {
impactedColumns.add(i);
// Compute features
Column column = project.columnModel.getColumnByCellIndex(i);
ReconConfig config = column.getReconConfig();
if (config instanceof StandardReconConfig) {
StandardReconConfig stdConfig = (StandardReconConfig)config;
stdConfig.computeFeatures(recon, cell.getValueAsString());
}
}
}
}

13
extensions/wikidata/tests/src/org/openrefine/wikidata/editing/NewItemLibraryTest.java
View File

@ -24,6 +24,9 @@
package org.openrefine.wikidata.editing;
import static org.junit.Assert.assertEquals;
import static org.testng.Assert.assertTrue;
import java.util.Collections;
import org.openrefine.wikidata.testing.JacksonSerializationTest;
import org.openrefine.wikidata.testing.TestingData;
@ -33,6 +36,7 @@ import org.testng.annotations.Test;
import com.google.refine.model.Cell;
import com.google.refine.model.Project;
import com.google.refine.model.Recon;
import com.google.refine.model.recon.StandardReconConfig;
import com.google.refine.tests.RefineTest;
public class NewItemLibraryTest extends RefineTest {
@ -54,6 +58,10 @@ public class NewItemLibraryTest extends RefineTest {
@Test
public void testUpdateReconciledCells() {
Project project = createCSVProject(TestingData.inceptionWithNewCsv);
StandardReconConfig config = new StandardReconConfig("http://my.endpoint",
"http://my.schema", "http://my.schema", "Q5", "human", true, Collections.emptyList());
project.columnModel.columns.get(0).setReconConfig(config);
project.rows.get(0).cells.set(0, TestingData.makeNewItemCell(3289L, "University of Ljubljana"));
project.rows.get(1).cells.set(0, TestingData.makeMatchedCell("Q865528", "University of Warwick"));
project.rows.get(2).cells.set(0, TestingData.makeNewItemCell(1234L, "new uni"));
@ -61,9 +69,12 @@ public class NewItemLibraryTest extends RefineTest {
isMatchedTo("Q865528", project.rows.get(1).cells.get(0));
isNewTo(1234L, project.rows.get(2).cells.get(0));
library.updateReconciledCells(project, false);
isMatchedTo("Q384", project.rows.get(0).cells.get(0));
Cell firstCell = project.rows.get(0).cells.get(0);
isMatchedTo("Q384", firstCell);
assertTrue((Boolean) firstCell.recon.getFeature(Recon.Feature_nameMatch));
isMatchedTo("Q865528", project.rows.get(1).cells.get(0));
isMatchedTo("Q345", project.rows.get(2).cells.get(0));
assertTrue(project.rows.get(2).cells.get(0).recon.getFeature(Recon.Feature_nameLevenshtein).equals(0));
library.updateReconciledCells(project, true);
isNewTo(3289L, project.rows.get(0).cells.get(0));
isMatchedTo("Q865528", project.rows.get(1).cells.get(0));

21
main/src/com/google/refine/model/recon/StandardReconConfig.java
View File

@ -551,8 +551,20 @@ public class StandardReconConfig extends ReconConfig {
recon.addCandidate(candidate);
count++;
}
if (count > 0) {
computeFeatures(recon, text);
return recon;
}
/**
* Recomputes the features associated with this reconciliation
* object (only if we have at least one candidate).
*
* @param text
* the cell value to compare the reconciliation data to
*/
public void computeFeatures(Recon recon, String text) {
if (!recon.candidates.isEmpty()) {
ReconCandidate candidate = recon.candidates.get(0);
recon.setFeature(Recon.Feature_nameMatch, text.equalsIgnoreCase(candidate.name));
@ -569,9 +581,10 @@ public class StandardReconConfig extends ReconConfig {
}
}
}
} else {
recon.features = new Object[Recon.Feature_max];
}
return recon;
}
}
static protected double wordDistance(String s1, String s2) {
Set<String> words1 = breakWords(s1);