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

RED-6619 - backmerge

Browse Source
This commit is contained in:
Thomas Beyer 2023-04-28 16:52:30 +02:00
parent bdcd3c7add
commit e5b8f7f6be
3 changed files with 13 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
redaction-service-v1/redaction-service-server-v1/src/main/java/com/iqser/red/service/redaction/v1/server/tableextraction/model/Cell.java
View File

@ -25,6 +25,8 @@ public class Cell extends Rectangle {
private boolean isHeaderCell; private boolean isHeaderCell;
private static final int MIN_SIZE = 1;
public Cell(Point2D topLeft, Point2D bottomRight) { public Cell(Point2D topLeft, Point2D bottomRight) {
@ -66,4 +68,10 @@ public class Cell extends Rectangle {
return TextNormalizationUtilities.removeHyphenLineBreaks(sb.toString()).replaceAll("\n", " ").replaceAll(" {2}", " "); return TextNormalizationUtilities.removeHyphenLineBreaks(sb.toString()).replaceAll("\n", " ").replaceAll(" {2}", " ");
} }
public boolean hasMinimumSize() {
return this.getHeight() >= MIN_SIZE && this.getWidth() >= MIN_SIZE;
}
} }

4
redaction-service-v1/redaction-service-server-v1/src/main/java/com/iqser/red/service/redaction/v1/server/tableextraction/model/Table.java
View File

@ -261,7 +261,9 @@ public class Table extends AbstractTextContainer {
if (intersectionCell.isPresent()) { if (intersectionCell.isPresent()) {
cell.getTextBlocks().addAll(intersectionCell.get().getTextBlocks()); cell.getTextBlocks().addAll(intersectionCell.get().getTextBlocks());
} }
row.add(cell); if (cell.hasMinimumSize()) {
row.add(cell);
}
} }
prevX = x; prevX = x;
} }

4
redaction-service-v1/redaction-service-server-v1/src/main/java/com/iqser/red/service/redaction/v1/server/tableextraction/service/TableExtractionService.java
View File

@ -88,7 +88,7 @@ public class TableExtractionService {
for (AbstractTextContainer abstractTextContainer : page.getTextBlocks()) { for (AbstractTextContainer abstractTextContainer : page.getTextBlocks()) {
TextBlock textBlock = (TextBlock) abstractTextContainer; TextBlock textBlock = (TextBlock) abstractTextContainer;
for (Cell cell : cells) { for (Cell cell : cells) {
if (cell.intersects(textBlock.getPdfMinX(), if (cell.hasMinimumSize() && cell.intersects(textBlock.getPdfMinX(),
textBlock.getPdfMinY(), textBlock.getPdfMinY(),
textBlock.getPdfMaxX() - textBlock.getPdfMinX(), textBlock.getPdfMaxX() - textBlock.getPdfMinX(),
textBlock.getPdfMaxY() - textBlock.getPdfMinY())) { textBlock.getPdfMaxY() - textBlock.getPdfMinY())) {
@ -109,7 +109,7 @@ public class TableExtractionService {
List<Cell> overlappingCells = new ArrayList<>(); List<Cell> overlappingCells = new ArrayList<>();
for (Cell c : cells) { for (Cell c : cells) {
if (c.intersects(area)) { if (c.hasMinimumSize() && c.intersects(area)) {
overlappingCells.add(c); overlappingCells.add(c);
} }
} }