Marmelator/layout-parser
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 289 Wiki Activity

RED-3813: Recategorize same image as experimental feature

Browse Source 
now writing hash into structure
This commit is contained in:
yhampe 2024-06-05 14:20:33 +02:00
parent 9be672c728
commit 9ecf9ca19f
3 changed files with 29 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

36
layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/LayoutParsingPipeline.java
View File

@ -105,21 +105,28 @@ public class LayoutParsingPipeline {
log.info("Starting layout parsing for {}", layoutParsingRequest.identifier()); log.info("Starting layout parsing for {}", layoutParsingRequest.identifier());
File originFile = layoutParsingStorageService.getOriginFile(layoutParsingRequest.originFileStorageId()); File originFile = layoutParsingStorageService.getOriginFile(layoutParsingRequest.originFileStorageId());
File viewerDocumentFile = layoutParsingStorageService.getViewerDocFile(layoutParsingRequest.viewerDocumentStorageId()).orElse(originFile); File viewerDocumentFile = layoutParsingStorageService.getViewerDocFile(layoutParsingRequest.viewerDocumentStorageId())
.orElse(originFile);
VisualLayoutParsingResponse visualLayoutParsingResponse = new VisualLayoutParsingResponse(); VisualLayoutParsingResponse visualLayoutParsingResponse = new VisualLayoutParsingResponse();
if (layoutParsingRequest.visualLayoutParsingFileId().isPresent()) { if (layoutParsingRequest.visualLayoutParsingFileId()
visualLayoutParsingResponse = layoutParsingStorageService.getVisualLayoutParsingFile(layoutParsingRequest.visualLayoutParsingFileId().get()); .isPresent()) {
visualLayoutParsingResponse = layoutParsingStorageService.getVisualLayoutParsingFile(layoutParsingRequest.visualLayoutParsingFileId()
.get());
} }
ImageServiceResponse imageServiceResponse = new ImageServiceResponse(); ImageServiceResponse imageServiceResponse = new ImageServiceResponse();
if (layoutParsingRequest.imagesFileStorageId().isPresent()) { if (layoutParsingRequest.imagesFileStorageId()
imageServiceResponse = layoutParsingStorageService.getImagesFile(layoutParsingRequest.imagesFileStorageId().get()); .isPresent()) {
imageServiceResponse = layoutParsingStorageService.getImagesFile(layoutParsingRequest.imagesFileStorageId()
.get());
} }
TableServiceResponse tableServiceResponse = new TableServiceResponse(); TableServiceResponse tableServiceResponse = new TableServiceResponse();
if (layoutParsingRequest.tablesFileStorageId().isPresent()) { if (layoutParsingRequest.tablesFileStorageId()
tableServiceResponse = layoutParsingStorageService.getTablesFile(layoutParsingRequest.tablesFileStorageId().get()); .isPresent()) {
tableServiceResponse = layoutParsingStorageService.getTablesFile(layoutParsingRequest.tablesFileStorageId()
.get());
} }
ClassificationDocument classificationDocument = parseLayout(settings.getLayoutParsingTypeOverride() == null // ClassificationDocument classificationDocument = parseLayout(settings.getLayoutParsingTypeOverride() == null //
@ -130,14 +137,17 @@ public class LayoutParsingPipeline {
visualLayoutParsingResponse, visualLayoutParsingResponse,
layoutParsingRequest.identifier()); layoutParsingRequest.identifier());
log.info("Building document graph for {}", layoutParsingRequest.identifier());
Document documentGraph = observeBuildDocumentGraph(settings.getLayoutParsingTypeOverride() == null // Document documentGraph = observeBuildDocumentGraph(settings.getLayoutParsingTypeOverride() == null //
? layoutParsingRequest.layoutParsingType() : settings.getLayoutParsingTypeOverride(), classificationDocument); ? layoutParsingRequest.layoutParsingType() : settings.getLayoutParsingTypeOverride(), classificationDocument);
log.info("Creating viewer document for {}", layoutParsingRequest.identifier()); log.info("Creating viewer document for {}", layoutParsingRequest.identifier());
layoutGridService.addLayoutGrid(viewerDocumentFile, documentGraph, viewerDocumentFile, false, layoutParsingRequest.visualLayoutParsingFileId().isPresent()); layoutGridService.addLayoutGrid(viewerDocumentFile,
documentGraph,
viewerDocumentFile,
false,
layoutParsingRequest.visualLayoutParsingFileId()
.isPresent());
log.info("Storing resulting files for {}", layoutParsingRequest.identifier()); log.info("Storing resulting files for {}", layoutParsingRequest.identifier());
@ -281,7 +291,11 @@ public class LayoutParsingPipeline {
pdfImages.computeIfAbsent(pageNumber, x -> new ArrayList<>()) pdfImages.computeIfAbsent(pageNumber, x -> new ArrayList<>())
.addAll(graphics.stream() .addAll(graphics.stream()
.map(g -> new ClassifiedImage(new Rectangle2D.Double(g.x1, g.y1, g.width(), g.height()), ImageType.GRAPHIC, false, stripper.getPageNumber(),"")) .map(g -> new ClassifiedImage(new Rectangle2D.Double(g.x1, g.y1, g.width(), g.height()),
ImageType.GRAPHIC,
false,
stripper.getPageNumber(),
""))
.toList()); .toList());
ClassificationPage classificationPage = switch (layoutParsingType) { ClassificationPage classificationPage = switch (layoutParsingType) {

2
layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/python_api/adapter/VisualLayoutParsingAdapter.java
View File

@ -79,7 +79,7 @@ public class VisualLayoutParsingAdapter {
ClassifiedImage signature = new ClassifiedImage(new Rectangle2D.Float(t.getBox().getX1(), ClassifiedImage signature = new ClassifiedImage(new Rectangle2D.Float(t.getBox().getX1(),
t.getBox().getY1(), t.getBox().getY1(),
t.getBox().getX2() - t.getBox().getX1(), t.getBox().getX2() - t.getBox().getX1(),
t.getBox().getY2() - t.getBox().getY1()), ImageType.SIGNATURE, true, false, false, pageNumber); t.getBox().getY2() - t.getBox().getY1()), ImageType.SIGNATURE, true, false, false, pageNumber,"");
signatures.add(signature); signatures.add(signature);
} }

3
layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/factory/DocumentGraphFactory.java
View File

@ -43,8 +43,10 @@ import lombok.Builder;
import lombok.Getter; import lombok.Getter;
import lombok.experimental.FieldDefaults; import lombok.experimental.FieldDefaults;
import lombok.experimental.UtilityClass; import lombok.experimental.UtilityClass;
import lombok.extern.slf4j.Slf4j;
@UtilityClass @UtilityClass
@Slf4j
public class DocumentGraphFactory { public class DocumentGraphFactory {
public Document buildDocumentGraph(LayoutParsingType layoutParsingType, ClassificationDocument document) { public Document buildDocumentGraph(LayoutParsingType layoutParsingType, ClassificationDocument document) {
@ -138,6 +140,7 @@ public class DocumentGraphFactory {
.position(position) .position(position)
.transparent(image.isHasTransparency()) .transparent(image.isHasTransparency())
.page(page) .page(page)
.representationHash(image.getRepresentation())
.documentTree(context.getDocumentTree()) .documentTree(context.getDocumentTree())
.build(); .build();
page.getMainBody().add(imageNode); page.getMainBody().add(imageNode);