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

RED-6164: Fixed calculation of image is ocr on scanned pages with cv analysis found tables

Browse Source
This commit is contained in:
deiflaender 2023-02-17 12:13:39 +01:00
parent f8a0a911bc
commit 0cf867b97c
1 changed files with 5 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
redaction-service-v1/redaction-service-server-v1/src/main/java/com/iqser/red/service/redaction/v1/server/segmentation/PdfSegmentationService.java
View File

@ -101,15 +101,16 @@ public class PdfSegmentationService {
page.setPageWidth(cropbox.getWidth());
page.setPageHeight(cropbox.getHeight());
tableExtractionService.extractTables(cleanRulings, page);
buildPageStatistics(page);
increaseDocumentStatistics(page, document);
// If images is ocr needs to be calculated before textBlocks are moved into tables, otherwise findOcr algorithm needs to be adopted.
if (pdfImages != null && pdfImages.containsKey(pageNumber)) {
page.setImages(pdfImages.get(pageNumber));
imageService.findOcr(page);
}
tableExtractionService.extractTables(cleanRulings, page);
buildPageStatistics(page);
increaseDocumentStatistics(page, document);
pages.add(page);
}