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

Merge branch 'RED-7141' into 'main'

Browse Source 
RED-7141: Fixed more overlap problems

See merge request fforesight/layout-parser!119
This commit is contained in:
Dominique Eifländer 2024-03-14 16:46:10 +01:00
commit ac850c2626
3 changed files with 14 additions and 11 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

21
layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/blockification/DocstrumBlockificationService.java
View File

@ -54,7 +54,11 @@ public class DocstrumBlockificationService {
var zones = docstrumSegmentationService.segmentPage(textPositions, xyOrder); var zones = docstrumSegmentationService.segmentPage(textPositions, xyOrder);
var pageBlocks = toAbstractPageBlocks(zones, usedHorizonalRulings, usedVerticalRulings, xyOrder); var pageBlocks = toAbstractPageBlocks(zones, usedHorizonalRulings, usedVerticalRulings, xyOrder);
return new ClassificationPage(pageBlocks); var classificationPage = new ClassificationPage(pageBlocks);
mergeIntersectingBlocks(classificationPage.getTextBlocks(), 0, 0);
return classificationPage;
} }
@ -91,8 +95,6 @@ public class DocstrumBlockificationService {
public void combineBlocks(ClassificationPage page) { public void combineBlocks(ClassificationPage page) {
mergeIntersectingBlocks(page.getTextBlocks(), 0, 0);
TextPageBlock previous = new TextPageBlock(); TextPageBlock previous = new TextPageBlock();
ListIterator<AbstractPageBlock> itty = page.getTextBlocks().listIterator(); ListIterator<AbstractPageBlock> itty = page.getTextBlocks().listIterator();
while (itty.hasNext()) { while (itty.hasNext()) {
@ -112,22 +114,22 @@ public class DocstrumBlockificationService {
} }
if (areTheOnlyTwoBlocksOnHeightsWithBothMoreThanTwoLines(previous, current, page)) { if (areTheOnlyTwoBlocksOnHeightsWithBothMoreThanTwoLines(previous, current, page)) {
combineBlocksAndResetIterator(previous, current, itty, true); previous = combineBlocksAndResetIterator(previous, current, itty, true);
continue; continue;
} }
if (previous.almostIntersects(current, 0, 0)) { if (previous.almostIntersects(current, 0, 0)) {
combineBlocksAndResetIterator(previous, current, itty, previous.isToDuplicate()); previous = combineBlocksAndResetIterator(previous, current, itty, previous.isToDuplicate());
continue; continue;
} }
if (isSameTopOrBottomWithPreviousHavingMaxTwoLinesAndCurrentThanOneAndMax4OtherBlocksOnHeight(previous, current, page)) { if (isSameTopOrBottomWithPreviousHavingMaxTwoLinesAndCurrentThanOneAndMax4OtherBlocksOnHeight(previous, current, page)) {
combineBlocksAndResetIterator(previous, current, itty, previous.isToDuplicate()); previous = combineBlocksAndResetIterator(previous, current, itty, previous.isToDuplicate());
continue; continue;
} }
if (isOnlyIntersectingYAndOnelineOrPrevoiusTwoLines(previous, current, page)) { if (isOnlyIntersectingYAndOnelineOrPrevoiusTwoLines(previous, current, page)) {
combineBlocksAndResetIterator(previous, current, itty, previous.isToDuplicate()); previous = combineBlocksAndResetIterator(previous, current, itty, previous.isToDuplicate());
continue; continue;
} }
@ -135,7 +137,7 @@ public class DocstrumBlockificationService {
previous = current; previous = current;
} }
mergeIntersectingBlocks(page.getTextBlocks(), 0, 0); mergeIntersectingBlocks(page.getTextBlocks(), 0, 6.5f);
} }
@ -165,7 +167,7 @@ public class DocstrumBlockificationService {
} }
private void combineBlocksAndResetIterator(TextPageBlock previous, TextPageBlock current, ListIterator<AbstractPageBlock> itty, boolean toDuplicate) { private TextPageBlock combineBlocksAndResetIterator(TextPageBlock previous, TextPageBlock current, ListIterator<AbstractPageBlock> itty, boolean toDuplicate) {
previous.getSequences().addAll(current.getSequences()); previous.getSequences().addAll(current.getSequences());
previous = buildTextBlock(previous.getSequences(), 0); previous = buildTextBlock(previous.getSequences(), 0);
@ -174,6 +176,7 @@ public class DocstrumBlockificationService {
itty.previous(); itty.previous();
itty.set(previous); itty.set(previous);
itty.next(); itty.next();
return previous;
} }

2
layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/factory/SectionNodeFactory.java
View File

@ -91,7 +91,7 @@ public class SectionNodeFactory {
if (abstractPageBlock instanceof TextPageBlock) { if (abstractPageBlock instanceof TextPageBlock) {
switch (layoutParsingType) { switch (layoutParsingType) {
case CLARIFYND, CLARIFYND_PARAGRAPH_DEBUG -> { case REDACT_MANAGER, CLARIFYND, CLARIFYND_PARAGRAPH_DEBUG -> {
alreadyMerged.add(abstractPageBlock); alreadyMerged.add(abstractPageBlock);
remainingBlocks.remove(abstractPageBlock); remainingBlocks.remove(abstractPageBlock);
DocumentGraphFactory.addParagraphOrHeadline(section, (TextPageBlock) abstractPageBlock, context, new ArrayList<>()); DocumentGraphFactory.addParagraphOrHeadline(section, (TextPageBlock) abstractPageBlock, context, new ArrayList<>());

2
layoutparser-service/layoutparser-service-server/src/test/java/com/knecon/fforesight/service/layoutparser/server/graph/ViewerDocumentTest.java
View File

@ -27,7 +27,7 @@ public class ViewerDocumentTest extends BuildDocumentTest {
@SneakyThrows @SneakyThrows
public void testViewerDocument() { public void testViewerDocument() {
String fileName = "files/new/270 rotated text on non rotated pages.pdf"; String fileName = "files/100 Trinexapac-ethyl_RAR_20_Volume_3CP_B-9_ 2018-01-10.pdf";
String tmpFileName = "/tmp/" + Path.of(fileName).getFileName() + "_VIEWER.pdf"; String tmpFileName = "/tmp/" + Path.of(fileName).getFileName() + "_VIEWER.pdf";
var documentFile = new ClassPathResource(fileName).getFile(); var documentFile = new ClassPathResource(fileName).getFile();