close #249: OOM due to font loading. Fonts are not cached, loading a …

…single instance of the same font in memory, per document.

sejda-fonts/src/main/java/org/sejda/fonts/UnicodeType0Font.java

@@ -34,6 +34,11 @@ public enum UnicodeType0Font implements FontResource {
         this.resource = resource;
     }
 
+    @Override
+    public String getResource() {
+        return resource;
+    }
+
     @Override
     public InputStream getFontStream() {
         return this.getClass().getResourceAsStream(resource);

sejda-model/src/main/java/org/sejda/model/pdf/FontResource.java

@@ -24,6 +24,11 @@
  * @author Andrea Vacondio
  */
 public interface FontResource {
+    /**
+     * @return Unique identifier
+     *
+     */
+    String getResource();
     /**
      * @return the stream where the font is read from
      */

sejda-optional-fonts/src/main/java/org/sejda/fonts/OptionalUnicodeType0Font.java

@@ -30,6 +30,11 @@ public enum OptionalUnicodeType0Font implements FontResource {
         this.resource = resource;
     }
 
+    @Override
+    public String getResource() {
+        return resource;
+    }
+
     @Override
     public InputStream getFontStream() {
         return this.getClass().getResourceAsStream(resource);

sejda-sambox/src/main/java/org/sejda/impl/sambox/EditTask.java

@@ -37,6 +37,7 @@
 import org.sejda.impl.sambox.component.PageGeometricalShapeWriter;
 import org.sejda.impl.sambox.component.PageImageWriter;
 import org.sejda.impl.sambox.component.PageTextWriter;
+import org.sejda.impl.sambox.util.FontUtils;
 import org.sejda.model.RectangularBox;
 import org.sejda.model.exception.TaskException;
 import org.sejda.model.input.PdfSource;
@@ -196,6 +197,8 @@ public void execute(EditParameters parameters) throws TaskException {
                     nameRequest().originalName(source.getName()).fileNumber(currentStep));
             outputWriter.addOutput(file(tmpFile).name(outName));
 
+            FontUtils.clearLoadedFontCache(documentHandler.getUnderlyingPDDocument());
+
             notifyEvent(executionContext().notifiableTaskMetadata()).stepsCompleted(currentStep).outOf(totalSteps);
         }
 

sejda-sambox/src/main/java/org/sejda/impl/sambox/SetHeaderFooterTask.java

@@ -32,6 +32,7 @@
 import org.sejda.impl.sambox.component.DefaultPdfSourceOpener;
 import org.sejda.impl.sambox.component.PDDocumentHandler;
 import org.sejda.impl.sambox.component.SetHeaderFooterWriter;
+import org.sejda.impl.sambox.util.FontUtils;
 import org.sejda.model.exception.TaskException;
 import org.sejda.model.input.PdfSource;
 import org.sejda.model.input.PdfSourceOpener;
@@ -95,6 +96,7 @@ public void execute(SetHeaderFooterParameters parameters) throws TaskException {
                 footerWriter.write(parameters, currentFileCounter, outName);
                 documentHandler.savePDDocument(tmpFile);
                 outputWriter.addOutput(file(tmpFile).name(outName));
+                FontUtils.clearLoadedFontCache(documentHandler.getUnderlyingPDDocument());
             }
 
             notifyEvent(executionContext().notifiableTaskMetadata()).stepsCompleted(currentStep).outOf(totalSteps);

sejda-sambox/src/main/java/org/sejda/impl/sambox/util/FontUtils.java

@@ -24,6 +24,7 @@
 import java.io.InputStream;
 import java.util.Collections;
 import java.util.EnumMap;
+import java.util.HashMap;
 import java.util.Map;
 import java.util.function.Supplier;
 
@@ -100,11 +101,34 @@ public static PDFont fontOrFallback(String text, PDFont font, Supplier<PDFont> f
         return font;
     }
 
+    // caches fonts, PER DOCUMENT
+    // has no auto-magical way to clear the cache when doc processing is done
+    // if you use this in a long lived process, call the cache clear method to avoid leaking memory
+    private static Map<PDDocument, Map<String, PDFont>> loadedFontCache = new HashMap<>();
+
+    public static void clearLoadedFontCache() {
+        loadedFontCache.clear();
+    }
+
+    public static void clearLoadedFontCache(PDDocument document) {
+        loadedFontCache.remove(document);
+    }
+
     private static PDFont loadFont(PDDocument document, FontResource font) {
+        if(!loadedFontCache.containsKey(document)){
+            loadedFontCache.put(document, new HashMap<>());
+        }
+
+        Map<String, PDFont> docCache = loadedFontCache.get(document);
+        if(docCache.containsKey(font.getResource())){
+            return docCache.get(font.getResource());
+        }
+
         InputStream in = font.getFontStream();
         try {
             PDType0Font loaded = PDType0Font.load(document, in);
             LOG.trace("Loaded font {}", loaded.getName());
+            docCache.put(font.getResource(), loaded);
             return loaded;
         } catch (IOException e) {
             LOG.warn("Failed to load font " + font, e);

sejda-sambox/src/test/java/org/sejda/impl/sambox/util/FontUtilsTest.java

@@ -19,6 +19,8 @@
  */
 package org.sejda.impl.sambox.util;
 
+import static org.hamcrest.MatcherAssert.assertThat;
+import static org.hamcrest.Matchers.is;
 import static org.junit.Assert.assertEquals;
 import static org.junit.Assert.assertFalse;
 import static org.junit.Assert.assertNull;
@@ -31,6 +33,7 @@
 import org.junit.Test;
 import org.sejda.model.pdf.StandardType1Font;
 import org.sejda.sambox.pdmodel.PDDocument;
+import org.sejda.sambox.pdmodel.font.PDFont;
 import org.sejda.sambox.pdmodel.font.PDType1Font;
 
 /**
@@ -89,4 +92,13 @@ public void testFontOrFallbackNullSipplier() {
         PDType1Font expected = getStandardType1Font(StandardType1Font.CURIER);
         assertEquals(expected, fontOrFallback("कसौटी", expected, null));
     }
+
+    @Test
+    public void testCaching() {
+        PDDocument doc = new PDDocument();
+        PDFont expected = FontUtils.findFontFor(doc, "ทดสอบ");
+
+        PDFont actual = findFontFor(doc, "ทด");
+        assertTrue("Font is cached, same instance is returned", expected == actual);
+    }
 }