Pull request #421: RED-4548: PORT: Create SIMPLIFIED_TEXT.json from TEXT.json (for NER-service)

Merge in RED/redaction-service from RED-4548-rs1PORT to release/3.107.x

* commit '8eef6ab3faa92da5f4d510131805614abda14c9a':
  RED-4548: PORT: Create SIMPLIFIED_TEXT.json from TEXT.json (for NER-service)
This commit is contained in:
Ali Oezyetimoglu 2022-07-14 11:53:53 +02:00 committed by Philipp Schramm
commit a76eedb4cf
4 changed files with 57 additions and 2 deletions

View File

@ -12,7 +12,7 @@
<artifactId>redaction-service-api-v1</artifactId> <artifactId>redaction-service-api-v1</artifactId>
<properties> <properties>
<persistence-service.version>1.143.0</persistence-service.version> <persistence-service.version>1.156.20</persistence-service.version>
</properties> </properties>
<dependencies> <dependencies>

View File

@ -0,0 +1,18 @@
package com.iqser.red.service.redaction.v1.server.classification.model;
import lombok.AllArgsConstructor;
import lombok.Builder;
import lombok.Data;
import lombok.NoArgsConstructor;
@Data
@Builder
@NoArgsConstructor
@AllArgsConstructor
public class SimplifiedSectionText {
private int sectionNumber;
private String text;
}

View File

@ -0,0 +1,20 @@
package com.iqser.red.service.redaction.v1.server.classification.model;
import java.util.ArrayList;
import java.util.List;
import lombok.AllArgsConstructor;
import lombok.Builder;
import lombok.Data;
import lombok.NoArgsConstructor;
@Data
@Builder
@NoArgsConstructor
@AllArgsConstructor
public class SimplifiedText {
private int numberOfPages;
private List<SimplifiedSectionText> simplifiedSectionTexts = new ArrayList<>();
}

View File

@ -2,10 +2,12 @@ package com.iqser.red.service.redaction.v1.server.redaction.service;
import static com.iqser.red.service.redaction.v1.server.redaction.service.ImportedRedactionService.IMPORTED_REDACTION_TYPE; import static com.iqser.red.service.redaction.v1.server.redaction.service.ImportedRedactionService.IMPORTED_REDACTION_TYPE;
import java.util.ArrayList;
import java.util.HashMap; import java.util.HashMap;
import java.util.HashSet; import java.util.HashSet;
import java.util.List; import java.util.List;
import java.util.Map; import java.util.Map;
import java.util.Objects;
import java.util.Set; import java.util.Set;
import java.util.stream.Collectors; import java.util.stream.Collectors;
import java.util.stream.Stream; import java.util.stream.Stream;
@ -30,6 +32,8 @@ import com.iqser.red.service.redaction.v1.model.SectionGrid;
import com.iqser.red.service.redaction.v1.model.StructureAnalyzeRequest; import com.iqser.red.service.redaction.v1.model.StructureAnalyzeRequest;
import com.iqser.red.service.redaction.v1.server.classification.model.Document; import com.iqser.red.service.redaction.v1.server.classification.model.Document;
import com.iqser.red.service.redaction.v1.server.classification.model.SectionText; import com.iqser.red.service.redaction.v1.server.classification.model.SectionText;
import com.iqser.red.service.redaction.v1.server.classification.model.SimplifiedSectionText;
import com.iqser.red.service.redaction.v1.server.classification.model.SimplifiedText;
import com.iqser.red.service.redaction.v1.server.classification.model.Text; import com.iqser.red.service.redaction.v1.server.classification.model.Text;
import com.iqser.red.service.redaction.v1.server.client.LegalBasisClient; import com.iqser.red.service.redaction.v1.server.client.LegalBasisClient;
import com.iqser.red.service.redaction.v1.server.client.model.NerEntities; import com.iqser.red.service.redaction.v1.server.client.model.NerEntities;
@ -104,8 +108,9 @@ public class AnalyzeService {
.map(SectionArea::getPage) .map(SectionArea::getPage)
.collect(Collectors.toSet()), sectionText.getSectionAreas()))); .collect(Collectors.toSet()), sectionText.getSectionAreas())));
log.info("Store text and section grid for file {} in dossier {}", analyzeRequest.getFileId(), analyzeRequest.getDossierId()); log.info("Store text, simplified text and section grid for file {} in dossier {}", analyzeRequest.getFileId(), analyzeRequest.getDossierId());
redactionStorageService.storeObject(analyzeRequest.getDossierId(), analyzeRequest.getFileId(), FileType.TEXT, text); redactionStorageService.storeObject(analyzeRequest.getDossierId(), analyzeRequest.getFileId(), FileType.TEXT, text);
redactionStorageService.storeObject(analyzeRequest.getDossierId(), analyzeRequest.getFileId(), FileType.SIMPLIFIED_TEXT, convert(text));
redactionStorageService.storeObject(analyzeRequest.getDossierId(), analyzeRequest.getFileId(), FileType.SECTION_GRID, classifiedDoc.getSectionGrid()); redactionStorageService.storeObject(analyzeRequest.getDossierId(), analyzeRequest.getFileId(), FileType.SECTION_GRID, classifiedDoc.getSectionGrid());
return AnalyzeResult.builder() return AnalyzeResult.builder()
@ -310,4 +315,16 @@ public class AnalyzeService {
} }
} }
private SimplifiedText convert(Text text) {
List<SimplifiedSectionText> sectionTexts = new ArrayList<>();
int numberOfPages = 0;
if (text != null) {
numberOfPages = text.getNumberOfPages();
text.getSectionTexts().stream().filter(Objects::nonNull).forEach(st -> sectionTexts.add(new SimplifiedSectionText(st.getSectionNumber(), st.getText())));
}
return SimplifiedText.builder().numberOfPages(numberOfPages).simplifiedSectionTexts(sectionTexts).build();
}
} }