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

Report duplicate items in DistinctValuesScrutinizer

Browse Source
This commit is contained in:
Antonin Delpeuch 2018-01-12 19:32:45 +00:00
parent 212b1b3dd3
commit 699ec16082
2 changed files with 9 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
extensions/wikidata/module/langs/translation-en.json
View File

@ -95,7 +95,7 @@
}, },
"identical-values-for-distinct-valued-property": { "identical-values-for-distinct-valued-property": {
"title": "Identical values for {property_entity}", "title": "Identical values for {property_entity}",
"body": "This property should have distinct values." "body": "This property should have distinct values, but the same value was found on {item1_entity} and {item2_entity} for instance."
}, },
"no-edit-generated": { "no-edit-generated": {
"title": "No edit was generated.", "title": "No edit was generated.",

14
extensions/wikidata/src/org/openrefine/wikidata/qa/scrutinizers/DistinctValuesScrutinizer.java
View File

@ -20,7 +20,7 @@ import org.wikidata.wdtk.datamodel.interfaces.Value;
*/ */
public class DistinctValuesScrutinizer extends StatementScrutinizer { public class DistinctValuesScrutinizer extends StatementScrutinizer {
private Map<PropertyIdValue, Set<Value>> _seenValues; private Map<PropertyIdValue, Map<Value, EntityIdValue>> _seenValues;
public DistinctValuesScrutinizer() { public DistinctValuesScrutinizer() {
_seenValues = new HashMap<>(); _seenValues = new HashMap<>();
@ -31,22 +31,24 @@ public class DistinctValuesScrutinizer extends StatementScrutinizer {
PropertyIdValue pid = statement.getClaim().getMainSnak().getPropertyId(); PropertyIdValue pid = statement.getClaim().getMainSnak().getPropertyId();
if (_fetcher.hasDistinctValues(pid)) { if (_fetcher.hasDistinctValues(pid)) {
Value mainSnakValue = statement.getClaim().getMainSnak().getValue(); Value mainSnakValue = statement.getClaim().getMainSnak().getValue();
Set<Value> seen = _seenValues.get(pid); Map<Value, EntityIdValue> seen = _seenValues.get(pid);
if (seen == null) { if (seen == null) {
seen = new HashSet<Value>(); seen = new HashMap<Value, EntityIdValue>();
_seenValues.put(pid, seen); _seenValues.put(pid, seen);
} }
if (seen.contains(mainSnakValue)) { if (seen.containsKey(mainSnakValue)) {
EntityIdValue otherId = seen.get(mainSnakValue);
QAWarning issue = new QAWarning( QAWarning issue = new QAWarning(
"identical-values-for-distinct-valued-property", "identical-values-for-distinct-valued-property",
pid.getId(), pid.getId(),
QAWarning.Severity.IMPORTANT, QAWarning.Severity.IMPORTANT,
1); 1);
issue.setProperty("property_entity", pid); issue.setProperty("property_entity", pid);
// TODO also report the items on which the property is duplicated issue.setProperty("item1_entity", entityId);
issue.setProperty("item2_entity", otherId);
addIssue(issue); addIssue(issue);
} else { } else {
seen.add(mainSnakValue); seen.put(mainSnakValue, entityId);
} }
} }
} }