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

RED-7384: Fixes for ClassCastException

Browse Source 
* changed save -> incrementalSave
* always use origin file instead of reusing viewerdoc
* Sometimes the viewer document is corrupted after saving and missing the contentstreams on a random page, for the files we viewed it did not seem to happen with incrementalSave.might only be a timing issue though
This commit is contained in:
Kilian Schuettler 2024-03-08 12:42:40 +01:00
parent a266d98f11
commit 007cbfd1ee
3 changed files with 16 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/LayoutParsingPipeline.java
View File

@ -94,8 +94,10 @@ public class LayoutParsingPipeline {
log.info("Starting layout parsing for {}", layoutParsingRequest.identifier());
File originFile = layoutParsingStorageService.getOriginFile(layoutParsingRequest.originFileStorageId());
File viewerDocumentFile = layoutParsingStorageService.getViewerDocFile(layoutParsingRequest.viewerDocumentStorageId())
.orElse(originFile);
// File viewerDocumentFile = layoutParsingStorageService.getViewerDocFile(layoutParsingRequest.viewerDocumentStorageId())
// .orElse(originFile);
File viewerDocumentFile = originFile;
ImageServiceResponse imageServiceResponse = new ImageServiceResponse();
if (layoutParsingRequest.imagesFileStorageId()

9
layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/model/text/RedTextPosition.java
View File

@ -1,7 +1,6 @@
package com.knecon.fforesight.service.layoutparser.processor.model.text;
import org.apache.pdfbox.text.TextPosition;
import org.springframework.beans.BeanUtils;
import com.fasterxml.jackson.annotation.JsonIgnore;
@ -50,7 +49,13 @@ public class RedTextPosition {
public static RedTextPosition fromTextPosition(TextPosition textPosition) {
var pos = new RedTextPosition();
BeanUtils.copyProperties(textPosition, pos);
pos.setRotation(textPosition.getRotation());
pos.setPageHeight(textPosition.getPageHeight());
pos.setPageWidth(textPosition.getPageWidth());
pos.setUnicode(textPosition.getUnicode());
pos.setDir(textPosition.getDir());
pos.setWidthOfSpace(textPosition.getWidthOfSpace());
pos.setFontSizeInPt(textPosition.getFontSizeInPt());
pos.setFontName(textPosition.getFont().getName());
pos.setFontSizeInPt(textPosition.getFontSizeInPt());

8
layoutparser-service/viewer-doc-processor/src/main/java/com/knecon/fforesight/service/viewerdoc/service/ViewerDocumentService.java
View File

@ -126,8 +126,8 @@ public class ViewerDocumentService {
pdDocument = openPDDocument(tmpFile.toFile());
}
}
observedIncrementalSave(pdDocument, destinationFile);
observedIncrementalSave(pdDocument, destinationFile);
pdDocument.close();
assert tmpFile.toFile().delete();
}
@ -282,10 +282,12 @@ public class ViewerDocumentService {
@SneakyThrows
private void observedIncrementalSave(PDDocument pdDocument, File outputFile) {
/*
Sometimes the viewer document is corrupted after saving and missing the content streams on a random page, for the files we viewed it did not seem to happen with incrementalSave. It might only be a timing issue though
*/
Observation.createNotStarted("ViewerDocumentService", registry).contextualName("incremental-save").observe(() -> {
try (var out = new FileOutputStream(outputFile)) {
pdDocument.save(out);
pdDocument.saveIncremental(out);
} catch (IOException e) {
throw new RuntimeException(e);
}