Merge branch 'feature/RED-10072' into 'master'

RED-10072: AI description field and toggle for entities

Closes RED-10072

See merge request redactmanager/redaction-service!539
This commit is contained in:
Maverick Studer 2024-11-07 14:43:51 +01:00
commit e415234bf8
8 changed files with 70 additions and 19 deletions

View File

@ -4,7 +4,7 @@ plugins {
} }
description = "redaction-service-api-v1" description = "redaction-service-api-v1"
val persistenceServiceVersion = "2.611.0" val persistenceServiceVersion = "2.612.0-RED10072.1"
dependencies { dependencies {
implementation("org.springframework:spring-web:6.0.12") implementation("org.springframework:spring-web:6.0.12")

View File

@ -16,8 +16,8 @@ val layoutParserVersion = "0.181.0"
val jacksonVersion = "2.15.2" val jacksonVersion = "2.15.2"
val droolsVersion = "9.44.0.Final" val droolsVersion = "9.44.0.Final"
val pdfBoxVersion = "3.0.0" val pdfBoxVersion = "3.0.0"
val persistenceServiceVersion = "2.611.0" val persistenceServiceVersion = "2.612.0-RED10072.1"
val llmServiceVersion = "1.11.0" val llmServiceVersion = "1.20.0-RED10072.2"
val springBootStarterVersion = "3.1.5" val springBootStarterVersion = "3.1.5"
val springCloudVersion = "4.0.4" val springCloudVersion = "4.0.4"
val testContainersVersion = "1.19.7" val testContainersVersion = "1.19.7"

View File

@ -82,4 +82,14 @@ public class NerEntities {
LLM_NER LLM_NER
} }
public static com.iqser.red.service.persistence.service.v1.api.shared.model.analysislog.entitylog.Engine mapToPrimaryEngine(NerEntities.Engine nerEntityEngine) {
return switch (nerEntityEngine) {
case NER, CLOUD_NER -> com.iqser.red.service.persistence.service.v1.api.shared.model.analysislog.entitylog.Engine.NER;
case LLM_NER -> com.iqser.red.service.persistence.service.v1.api.shared.model.analysislog.entitylog.Engine.LLM_NER;
};
}
} }

View File

@ -20,6 +20,7 @@ import com.iqser.red.service.persistence.service.v1.api.shared.model.AnalyzeRequ
import com.iqser.red.service.persistence.service.v1.api.shared.model.RuleFileType; import com.iqser.red.service.persistence.service.v1.api.shared.model.RuleFileType;
import com.iqser.red.service.persistence.service.v1.api.shared.model.analysislog.entitylog.EntityLog; import com.iqser.red.service.persistence.service.v1.api.shared.model.analysislog.entitylog.EntityLog;
import com.iqser.red.service.persistence.service.v1.api.shared.model.analysislog.entitylog.imported.ImportedRedactions; import com.iqser.red.service.persistence.service.v1.api.shared.model.analysislog.entitylog.imported.ImportedRedactions;
import com.iqser.red.service.persistence.service.v1.api.shared.model.dossiertemplate.type.Type;
import com.iqser.red.service.redaction.v1.server.RedactionServiceSettings; import com.iqser.red.service.redaction.v1.server.RedactionServiceSettings;
import com.iqser.red.service.redaction.v1.server.client.model.NerEntitiesModel; import com.iqser.red.service.redaction.v1.server.client.model.NerEntitiesModel;
import com.iqser.red.service.redaction.v1.server.model.KieWrapper; import com.iqser.red.service.redaction.v1.server.model.KieWrapper;
@ -39,6 +40,7 @@ import com.iqser.red.service.redaction.v1.server.service.document.SectionFinderS
import com.iqser.red.service.redaction.v1.server.service.drools.KieContainerCreationService; import com.iqser.red.service.redaction.v1.server.service.drools.KieContainerCreationService;
import com.iqser.red.service.redaction.v1.server.storage.ObservedStorageService; import com.iqser.red.service.redaction.v1.server.storage.ObservedStorageService;
import com.iqser.red.service.redaction.v1.server.storage.RedactionStorageService; import com.iqser.red.service.redaction.v1.server.storage.RedactionStorageService;
import com.knecon.fforesight.llm.service.LlmNerEntities;
import io.micrometer.core.annotation.Timed; import io.micrometer.core.annotation.Timed;
import io.micrometer.observation.annotation.Observed; import io.micrometer.observation.annotation.Observed;
@ -295,8 +297,10 @@ public class AnalysisPreparationService {
private NerEntities getLlmNerEntities(AnalyzeRequest analyzeRequest) { private NerEntities getLlmNerEntities(AnalyzeRequest analyzeRequest) {
return new NerEntities(redactionStorageService.getLlmNerEntities(analyzeRequest.getDossierId(), analyzeRequest.getFileId()).getEntities() LlmNerEntities llmNerEntities = redactionStorageService.getLlmNerEntities(analyzeRequest.getDossierId(), analyzeRequest.getFileId());
return new NerEntities(llmNerEntities.getEntities()
.stream() .stream()
//.filter(e -> allTypes.contains(e.getType()))
.map(e -> new NerEntities.NerEntity(e.getValue(), .map(e -> new NerEntities.NerEntity(e.getValue(),
new TextRange(e.getStartOffset(), e.getEndOffset()), new TextRange(e.getStartOffset(), e.getEndOffset()),
e.getType(), e.getType(),

View File

@ -20,6 +20,7 @@ import com.iqser.gin4.commons.metrics.meters.FunctionTimerValues;
import com.iqser.red.service.persistence.service.v1.api.shared.model.AnalyzeRequest; import com.iqser.red.service.persistence.service.v1.api.shared.model.AnalyzeRequest;
import com.iqser.red.service.persistence.service.v1.api.shared.model.AnalyzeResult; import com.iqser.red.service.persistence.service.v1.api.shared.model.AnalyzeResult;
import com.iqser.red.service.persistence.service.v1.api.shared.model.FileAttribute; import com.iqser.red.service.persistence.service.v1.api.shared.model.FileAttribute;
import com.iqser.red.service.persistence.service.v1.api.shared.model.MessageType;
import com.iqser.red.service.persistence.service.v1.api.shared.model.RuleFileType; import com.iqser.red.service.persistence.service.v1.api.shared.model.RuleFileType;
import com.iqser.red.service.persistence.service.v1.api.shared.model.analysislog.componentlog.ComponentLog; import com.iqser.red.service.persistence.service.v1.api.shared.model.analysislog.componentlog.ComponentLog;
import com.iqser.red.service.persistence.service.v1.api.shared.model.analysislog.entitylog.EntityLog; import com.iqser.red.service.persistence.service.v1.api.shared.model.analysislog.entitylog.EntityLog;

View File

@ -1,6 +1,7 @@
package com.iqser.red.service.redaction.v1.server.service; package com.iqser.red.service.redaction.v1.server.service;
import java.awt.Color; import java.awt.Color;
import java.util.ArrayList;
import java.util.Comparator; import java.util.Comparator;
import java.util.HashSet; import java.util.HashSet;
import java.util.LinkedList; import java.util.LinkedList;
@ -497,4 +498,12 @@ public class DictionaryService {
.collect(Collectors.toList()); .collect(Collectors.toList());
} }
public List<Type> getAllTypes(String dossierTemplateId, String dossierId) {
List<Type> allTypes = dictionaryClient.getAllTypesForDossierTemplate(dossierTemplateId, null, false);
allTypes.addAll(dictionaryClient.getAllTypesForDossier(dossierId, null, false));
return allTypes;
}
} }

View File

@ -9,6 +9,7 @@ import static com.iqser.red.service.redaction.v1.server.service.document.EntityC
import static com.iqser.red.service.redaction.v1.server.service.document.EntityCreationUtility.truncateEndIfLineBreakIsBetween; import static com.iqser.red.service.redaction.v1.server.service.document.EntityCreationUtility.truncateEndIfLineBreakIsBetween;
import static com.iqser.red.service.redaction.v1.server.utils.SeparatorUtils.boundaryIsSurroundedBySeparators; import static com.iqser.red.service.redaction.v1.server.utils.SeparatorUtils.boundaryIsSurroundedBySeparators;
import java.util.Arrays;
import java.util.Collection; import java.util.Collection;
import java.util.Collections; import java.util.Collections;
import java.util.Comparator; import java.util.Comparator;
@ -1176,7 +1177,7 @@ public class EntityCreationService {
*/ */
public TextEntity byNerEntity(NerEntities.NerEntity nerEntity, EntityType entityType, SemanticNode semanticNode) { public TextEntity byNerEntity(NerEntities.NerEntity nerEntity, EntityType entityType, SemanticNode semanticNode) {
return byTextRangeWithEngine(nerEntity.textRange(), nerEntity.type(), entityType, semanticNode, Set.of(Engine.NER)).orElseThrow(() -> new NotFoundException( return byTextRangeWithEngine(nerEntity.textRange(), nerEntity.type(), entityType, semanticNode, getNerEngines(nerEntity)).orElseThrow(() -> new NotFoundException(
"No entity present!")); "No entity present!"));
} }
@ -1192,7 +1193,11 @@ public class EntityCreationService {
*/ */
public TextEntity byNerEntity(NerEntities.NerEntity nerEntity, String type, EntityType entityType, SemanticNode semanticNode) { public TextEntity byNerEntity(NerEntities.NerEntity nerEntity, String type, EntityType entityType, SemanticNode semanticNode) {
return byTextRangeWithEngine(nerEntity.textRange(), type, entityType, semanticNode, Set.of(Engine.NER)).orElseThrow(() -> new NotFoundException("No entity present!")); return byTextRangeWithEngine(nerEntity.textRange(),
type,
entityType,
semanticNode,
getNerEngines(nerEntity)).orElseThrow(() -> new NotFoundException("No entity present!"));
} }
@ -1206,7 +1211,7 @@ public class EntityCreationService {
*/ */
public Optional<TextEntity> optionalByNerEntity(NerEntities.NerEntity nerEntity, EntityType entityType, SemanticNode semanticNode) { public Optional<TextEntity> optionalByNerEntity(NerEntities.NerEntity nerEntity, EntityType entityType, SemanticNode semanticNode) {
return byTextRangeWithEngine(nerEntity.textRange(), nerEntity.type(), entityType, semanticNode, Set.of(Engine.NER)); return byTextRangeWithEngine(nerEntity.textRange(), nerEntity.type(), entityType, semanticNode, getNerEngines(nerEntity));
} }
@ -1221,7 +1226,7 @@ public class EntityCreationService {
*/ */
public Optional<TextEntity> optionalByNerEntity(NerEntities.NerEntity nerEntity, String type, EntityType entityType, SemanticNode semanticNode) { public Optional<TextEntity> optionalByNerEntity(NerEntities.NerEntity nerEntity, String type, EntityType entityType, SemanticNode semanticNode) {
return byTextRangeWithEngine(nerEntity.textRange(), type, entityType, semanticNode, Set.of(Engine.NER)); return byTextRangeWithEngine(nerEntity.textRange(), type, entityType, semanticNode, getNerEngines(nerEntity));
} }
@ -1240,7 +1245,7 @@ public class EntityCreationService {
if (nerEntity.confidence() != null && nerEntity.confidence() < minConfidence) { if (nerEntity.confidence() != null && nerEntity.confidence() < minConfidence) {
return Optional.empty(); return Optional.empty();
} }
return byTextRangeWithEngine(nerEntity.textRange(), nerEntity.type(), entityType, semanticNode, Set.of(Engine.NER)); return byTextRangeWithEngine(nerEntity.textRange(), nerEntity.type(), entityType, semanticNode, getNerEngines(nerEntity));
} }
@ -1264,7 +1269,7 @@ public class EntityCreationService {
if (nerEntity.confidence() != null && nerEntity.confidence() < minConfidence) { if (nerEntity.confidence() != null && nerEntity.confidence() < minConfidence) {
return Optional.empty(); return Optional.empty();
} }
return byTextRangeWithEngine(nerEntity.textRange(), type, entityType, semanticNode, Set.of(Engine.NER)); return byTextRangeWithEngine(nerEntity.textRange(), type, entityType, semanticNode, getNerEngines(nerEntity));
} }
@ -1280,7 +1285,7 @@ public class EntityCreationService {
public Stream<TextEntity> combineNerEntitiesToCbiAddressDefaults(NerEntities nerEntities, String type, EntityType entityType, SemanticNode semanticNode) { public Stream<TextEntity> combineNerEntitiesToCbiAddressDefaults(NerEntities nerEntities, String type, EntityType entityType, SemanticNode semanticNode) {
return NerEntitiesAdapter.combineNerEntitiesToCbiAddressDefaults(nerEntities) return NerEntitiesAdapter.combineNerEntitiesToCbiAddressDefaults(nerEntities)
.map(boundary -> byTextRangeWithEngine(boundary, type, entityType, semanticNode, Set.of(Engine.NER))) .map(boundary -> byTextRangeWithEngine(boundary, type, entityType, semanticNode, getNerEngines(nerEntities)))
.filter(Optional::isPresent) .filter(Optional::isPresent)
.map(Optional::get); .map(Optional::get);
} }
@ -1307,7 +1312,7 @@ public class EntityCreationService {
int minPartsToCombine) { int minPartsToCombine) {
return NerEntitiesAdapter.combineNerEntitiesOfAllGivenTypes(nerEntities, essentialTypes, typesToCombine, minPartsToCombine) return NerEntitiesAdapter.combineNerEntitiesOfAllGivenTypes(nerEntities, essentialTypes, typesToCombine, minPartsToCombine)
.map(boundary -> byTextRangeWithEngine(boundary, type, entityType, semanticNode, Set.of(Engine.NER))) .map(boundary -> byTextRangeWithEngine(boundary, type, entityType, semanticNode, getNerEngines(nerEntities)))
.filter(Optional::isPresent) .filter(Optional::isPresent)
.map(Optional::get); .map(Optional::get);
} }
@ -1336,7 +1341,7 @@ public class EntityCreationService {
int minPartsToCombine) { int minPartsToCombine) {
return NerEntitiesAdapter.combineNerEntitiesOfAllGivenTypes(nerEntities, essentialTypes, typesToCombine, maxDistanceBetweenParts, minPartsToCombine) return NerEntitiesAdapter.combineNerEntitiesOfAllGivenTypes(nerEntities, essentialTypes, typesToCombine, maxDistanceBetweenParts, minPartsToCombine)
.map(boundary -> byTextRangeWithEngine(boundary, type, entityType, semanticNode, Set.of(Engine.NER))) .map(boundary -> byTextRangeWithEngine(boundary, type, entityType, semanticNode, getNerEngines(nerEntities)))
.filter(Optional::isPresent) .filter(Optional::isPresent)
.map(Optional::get); .map(Optional::get);
} }
@ -1372,7 +1377,7 @@ public class EntityCreationService {
maxDistanceBetweenParts, maxDistanceBetweenParts,
minPartsToCombine, minPartsToCombine,
minEssentialTypesCombined) minEssentialTypesCombined)
.map(boundary -> byTextRangeWithEngine(boundary, type, entityType, semanticNode, Set.of(Engine.NER))) .map(boundary -> byTextRangeWithEngine(boundary, type, entityType, semanticNode, getNerEngines(nerEntities)))
.filter(Optional::isPresent) .filter(Optional::isPresent)
.map(Optional::get); .map(Optional::get);
} }
@ -1411,7 +1416,7 @@ public class EntityCreationService {
minPartsToCombine, minPartsToCombine,
minEssentialTypesCombined, minEssentialTypesCombined,
confidence) confidence)
.map(boundary -> byTextRangeWithEngine(boundary, type, entityType, semanticNode, Set.of(Engine.NER))) .map(boundary -> byTextRangeWithEngine(boundary, type, entityType, semanticNode, getNerEngines(nerEntities)))
.filter(Optional::isPresent) .filter(Optional::isPresent)
.map(Optional::get); .map(Optional::get);
} }
@ -1440,7 +1445,7 @@ public class EntityCreationService {
DocumentTree documentTree = node.getDocumentTree(); DocumentTree documentTree = node.getDocumentTree();
if (node.getEntities().contains(entity)) { if (node.getEntities().contains(entity)) {
// If entity already exists and it has a different text range, we add the text range to the list of duplicated text ranges // If entity already exists, and it has a different text range, we add the text range to the list of duplicated text ranges
node.getEntities() node.getEntities()
.stream()// .stream()//
.filter(e -> e.equals(entity))// .filter(e -> e.equals(entity))//
@ -1455,6 +1460,23 @@ public class EntityCreationService {
} }
private static Set<Engine> getNerEngines(NerEntities nerEntities) {
return getNerEngines(nerEntities.getNerEntityList()
.toArray(new NerEntities.NerEntity[0]));
}
private static Set<Engine> getNerEngines(NerEntities.NerEntity... nerEntities) {
return Arrays.stream(nerEntities)
.map(NerEntities.NerEntity::engine)
.map(NerEntities::mapToPrimaryEngine)
.collect(Collectors.toSet());
}
private void addDuplicateEntityToGraph(TextEntity entityToDuplicate, TextRange newTextRange, SemanticNode node) { private void addDuplicateEntityToGraph(TextEntity entityToDuplicate, TextRange newTextRange, SemanticNode node) {
entityToDuplicate.addTextRange(newTextRange); entityToDuplicate.addTextRange(newTextRange);

View File

@ -6,6 +6,7 @@ import static org.springframework.boot.autoconfigure.task.TaskExecutionAutoConfi
import java.io.File; import java.io.File;
import java.io.FileInputStream; import java.io.FileInputStream;
import java.io.InputStream; import java.io.InputStream;
import java.util.ArrayList;
import java.util.Collection; import java.util.Collection;
import java.util.List; import java.util.List;
import java.util.Set; import java.util.Set;
@ -31,7 +32,6 @@ import com.iqser.red.service.redaction.v1.server.RedactionServiceSettings;
import com.iqser.red.service.redaction.v1.server.client.model.NerEntitiesModel; import com.iqser.red.service.redaction.v1.server.client.model.NerEntitiesModel;
import com.iqser.red.service.redaction.v1.server.model.document.DocumentData; import com.iqser.red.service.redaction.v1.server.model.document.DocumentData;
import com.iqser.red.service.redaction.v1.server.utils.exception.NotFoundException; import com.iqser.red.service.redaction.v1.server.utils.exception.NotFoundException;
import com.iqser.red.storage.commons.exception.StorageException;
import com.iqser.red.storage.commons.exception.StorageObjectDoesNotExist; import com.iqser.red.storage.commons.exception.StorageObjectDoesNotExist;
import com.iqser.red.storage.commons.service.StorageService; import com.iqser.red.storage.commons.service.StorageService;
import com.knecon.fforesight.llm.service.LlmNerEntities; import com.knecon.fforesight.llm.service.LlmNerEntities;
@ -137,13 +137,13 @@ public class RedactionStorageService {
} }
@SneakyThrows @SneakyThrows
public void saveComponentLog(String dossierId, String fileId, ComponentLog componentLog) { public void saveComponentLog(String dossierId, String fileId, ComponentLog componentLog) {
componentLogMongoService.saveComponentLog(dossierId, fileId, componentLog); componentLogMongoService.saveComponentLog(dossierId, fileId, componentLog);
} }
@SneakyThrows @SneakyThrows
public void updateEntityLogEntries(String dossierId, String fileId, List<EntityLogEntry> entityLogEntries) { public void updateEntityLogEntries(String dossierId, String fileId, List<EntityLogEntry> entityLogEntries) {
@ -373,7 +373,12 @@ public class RedactionStorageService {
public LlmNerEntities getLlmNerEntities(String dossierId, String fileId) { public LlmNerEntities getLlmNerEntities(String dossierId, String fileId) {
return storageService.readJSONObject(TenantContext.getTenantId(), StorageIdUtils.getStorageId(dossierId, fileId, FileType.LLM_NER_ENTITIES), LlmNerEntities.class); String objectId = StorageIdUtils.getStorageId(dossierId, fileId, FileType.LLM_NER_ENTITIES);
if (storageService.objectExists(TenantContext.getTenantId(), objectId)) {
return storageService.readJSONObject(TenantContext.getTenantId(), objectId, LlmNerEntities.class);
} else {
return LlmNerEntities.builder().entities(new ArrayList<>()).build();
}
} }