Marmelator/redaction-service
1
0
Fork 0
Code Issues Pull Requests 14 Actions Packages Projects Releases 622 Wiki Activity

RSS-266: Allow duplicated entries with different types for scm

Browse Source
This commit is contained in:
deiflaender 2022-11-25 14:09:02 +01:00
parent 47a13f5a99
commit 0d353f99cd
4 changed files with 11 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
redaction-service-v1/redaction-service-server-v1/src/main/java/com/iqser/red/service/redaction/v1/server/redaction/model/Entity.java
View File

@ -18,6 +18,8 @@ import java.util.Set;
public class Entity implements ReasonHolder { public class Entity implements ReasonHolder {
private String word; private String word;
@EqualsAndHashCode.Include
private String type; private String type;
private boolean redaction; private boolean redaction;
private boolean falsePositive; private boolean falsePositive;

8
redaction-service-v1/redaction-service-server-v1/src/main/java/com/iqser/red/service/redaction/v1/server/redaction/service/RedactionLogCreatorService.java
View File

@ -7,6 +7,7 @@ import java.util.Map;
import java.util.Set; import java.util.Set;
import org.apache.commons.collections4.CollectionUtils; import org.apache.commons.collections4.CollectionUtils;
import org.apache.commons.lang3.tuple.Pair;
import org.springframework.stereotype.Service; import org.springframework.stereotype.Service;
import com.iqser.red.service.redaction.v1.model.Point; import com.iqser.red.service.redaction.v1.model.Point;
@ -92,7 +93,7 @@ public class RedactionLogCreatorService {
List<RedactionLogEntry> redactionLogEntities = new ArrayList<>(); List<RedactionLogEntry> redactionLogEntities = new ArrayList<>();
// Duplicates can exist due table extraction columns over multiple rows. // Duplicates can exist due table extraction columns over multiple rows.
Set<String> processedIds = new HashSet<>(); Set<Pair<String, String>> processedIds = new HashSet<>();
entityLoop: entityLoop:
for (Entity entity : entities.get(page)) { for (Entity entity : entities.get(page)) {
@ -100,12 +101,11 @@ public class RedactionLogCreatorService {
for (EntityPositionSequence entityPositionSequence : entity.getPositionSequences()) { for (EntityPositionSequence entityPositionSequence : entity.getPositionSequences()) {
RedactionLogEntry redactionLogEntry = createRedactionLogEntry(entity, dossierTemplateId); RedactionLogEntry redactionLogEntry = createRedactionLogEntry(entity, dossierTemplateId);
if (processedIds.contains(entityPositionSequence.getId())) { if (processedIds.contains(Pair.of(entityPositionSequence.getId(), entity.getType()))) {
// TODO refactor this outer loop jump as soon as we have the time.
continue entityLoop; continue entityLoop;
} else { } else {
processedIds.add(entityPositionSequence.getId()); processedIds.add(Pair.of(entityPositionSequence.getId(), entity.getType()));
} }
redactionLogEntry.setId(entityPositionSequence.getId()); redactionLogEntry.setId(entityPositionSequence.getId());

8
redaction-service-v1/redaction-service-server-v1/src/main/java/com/iqser/red/service/redaction/v1/server/redaction/service/RedactionLogMergeService.java
View File

@ -152,10 +152,10 @@ public class RedactionLogMergeService {
return true; return true;
} }
} }
if (processedIds.contains(entry.getId())) { // if (processedIds.contains(entry.getId())) {
log.info("Duplicate annotation found with id {}", entry.getId()); // log.info("Duplicate annotation found with id {}", entry.getId());
return true; // return true;
} // }
processedIds.add(entry.getId()); processedIds.add(entry.getId());
return false; return false;
}); });

2
redaction-service-v1/redaction-service-server-v1/src/main/java/com/iqser/red/service/redaction/v1/server/redaction/utils/EntitySearchUtils.java
View File

@ -286,7 +286,7 @@ public class EntitySearchUtils {
existing.setRedaction(true); existing.setRedaction(true);
} }
} }
} else if (dictionary.getDictionaryRank(existing.getType()) <= dictionary.getDictionaryRank(found.getType())) { } else if (dictionary.getDictionaryRank(existing.getType()) < dictionary.getDictionaryRank(found.getType())) {
entities.remove(found); entities.remove(found);
entities.add(found); entities.add(found);
} }