Closes #373: high quality PDF to image conversion before OCR

docs-core/src/main/java/com/sismics/docs/core/util/format/PdfFormatHandler.java

@@ -6,6 +6,7 @@ import com.sismics.util.mime.MimeType;
 import org.apache.pdfbox.io.MemoryUsageSetting;
 import org.apache.pdfbox.multipdf.PDFMergerUtility;
 import org.apache.pdfbox.pdmodel.PDDocument;
+import org.apache.pdfbox.rendering.ImageType;
 import org.apache.pdfbox.rendering.PDFRenderer;
 import org.apache.pdfbox.text.PDFTextStripper;
 import org.slf4j.Logger;
@@ -60,7 +61,7 @@ public class PdfFormatHandler implements FormatHandler {
                 for (int pageIndex = 0; pageIndex < pdfDocument.getNumberOfPages(); pageIndex++) {
                     log.info("OCR page " + (pageIndex + 1) + "/" + pdfDocument.getNumberOfPages() + " of PDF file containing only images");
                     sb.append(" ");
-                    sb.append(FileUtil.ocrFile(language, renderer.renderImage(pageIndex)));
+                    sb.append(FileUtil.ocrFile(language, renderer.renderImageWithDPI(pageIndex, 300, ImageType.GRAY)));
                 }
                 return sb.toString();
             } catch (Exception e) {

docs-core/src/test/java/com/sismics/util/format/TestPdfFormatHandler.java

@@ -0,0 +1,19 @@
+package com.sismics.util.format;
+
+import com.sismics.docs.core.util.format.PdfFormatHandler;
+import org.junit.Assert;
+import org.junit.Test;
+
+import java.nio.file.Paths;
+
+public class TestPdfFormatHandler {
+    @Test
+    public void testIssue373() throws Exception {
+        PdfFormatHandler formatHandler = new PdfFormatHandler();
+        String content = formatHandler.extractContent("deu", Paths.get(ClassLoader.getSystemResource("file/issue373.pdf").toURI()));
+        Assert.assertTrue(content.contains("Aufrechterhaltung"));
+        Assert.assertTrue(content.contains("Außentemperatur"));
+        Assert.assertTrue(content.contains("Grundumsatzmessungen"));
+        Assert.assertTrue(content.contains("ermitteln"));
+    }
+}