Compare commits
| Author | SHA1 | Date | |
|---|---|---|---|
|
|
e535861da8 | ||
|
|
0e8dfed441 | ||
|
|
2fe0eb024a | ||
|
|
37548ddbf8 |
@@ -5,6 +5,10 @@ RUN tar xvzf OCRModuleLinux.tar.gz
|
||||
RUN mkdir /OCRModule
|
||||
RUN mv Lib/* /OCRModule/
|
||||
|
||||
RUN apt-get -y update
|
||||
# Ghostscript somehow improves ocr quality using pdfton, do not remove!
|
||||
RUN apt-get -y install ghostscript
|
||||
|
||||
ARG PLATFORM_JAR
|
||||
|
||||
ENV PLATFORM_JAR ${PLATFORM_JAR}
|
||||
|
||||
+2
-2
@@ -285,7 +285,7 @@ public class OCRService {
|
||||
var gState = element.getGState();
|
||||
|
||||
//See PDF Reference 5.3 Text rendering modes, 3 = Invisible, however this ocr does not use it.
|
||||
if (!filledRectangleIntersection && gState.getTextRenderMode() != 3) {
|
||||
if (!filledRectangleIntersection && gState.getTextRenderMode() != 3 || filledRectangleIntersection && gState.getTextRenderMode() == 0) {
|
||||
writer.writeElement(element);
|
||||
}
|
||||
}
|
||||
@@ -294,10 +294,10 @@ public class OCRService {
|
||||
@SneakyThrows
|
||||
private void processPath(Element element, ElementWriter writer, Set<Rect> filledRectangles) {
|
||||
|
||||
writer.writeElement(element);
|
||||
if (element.getPathData() != null && element.getPathData().getPoints().length > 4) {
|
||||
filledRectangles.add(element.getBBox());
|
||||
}
|
||||
writer.writeElement(element);
|
||||
}
|
||||
|
||||
|
||||
|
||||
+2
@@ -76,6 +76,8 @@ public class OcrServiceIntegrationTest {
|
||||
|
||||
var out = FileUtils.openOutputStream(new File(getTemporaryDirectory() + "/" + fileName + ".pdf"));
|
||||
IOUtils.copy(response, out);
|
||||
|
||||
System.out.println("File:" + getTemporaryDirectory() + "/" + fileName + ".pdf");
|
||||
}
|
||||
|
||||
|
||||
|
||||
Reference in New Issue
Block a user