⭐ #3 support recognition of watermarks wrapped in marked content

Author: hwding

src/com/amastigote/unstamper/core/Processor.java

@@ -1,6 +1,6 @@
 /*
   AUTH | hwding
-  DATE | Dec 25 2018
+  DATE | Feb 19 2019
   DESC | textual watermark remover for PDF files
   MAIL | [email protected]
   GITH | github.com/hwding
@@ -17,21 +17,29 @@
 import org.apache.pdfbox.pdfwriter.ContentStreamWriter;
 import org.apache.pdfbox.pdmodel.PDDocument;
 import org.apache.pdfbox.pdmodel.common.PDStream;
+import org.apache.pdfbox.pdmodel.documentinterchange.markedcontent.PDMarkedContent;
 import org.apache.pdfbox.pdmodel.font.PDFont;
 import org.apache.pdfbox.pdmodel.interactive.annotation.PDAnnotation;
+import org.apache.pdfbox.text.PDFMarkedContentExtractor;
+import org.apache.pdfbox.text.TextPosition;
 
 import java.io.File;
 import java.io.IOException;
 import java.io.OutputStream;
 import java.util.*;
+import java.util.stream.Collectors;
 
 public class Processor {
 
     private static final byte[] empBytes = new byte[0];
+    private static final Object empObj = new Object();
+    private static final Collection<Object> empObjCol = Collections.singleton(empObj);
     private static final List<PDAnnotation> empList = Collections.emptyList();
 
     private static final String SHOW_TEXT = "Tj";
     private static final String SHOW_TEXT_ADJUSTED = "TJ";
+    private static final String BEGIN_MARKED_CONTENT_WITH_PROPERTY = "BDC";
+    private static final String END_MARKDED_CONTENT = "EMC";
 
     public static void process(
             @NotNull File file,
@@ -47,11 +55,13 @@ public static void process(
             }
 
             PDDocument pdDocument = PDDocument.load(file);
+            PDFMarkedContentExtractor extractor = new PDFMarkedContentExtractor();
+
             pdDocument.getPages().forEach(pdPage -> {
                 try {
-                    boolean needRewrite = false;
+                    boolean needRewrite;
 
-                    /* START: loading font resources from current page */
+                    /* >> loading font resources from current page */
                     PDFStreamParser pdfStreamParser = new PDFStreamParser(pdPage);
                     pdfStreamParser.parse();
 
@@ -66,14 +76,52 @@ public static void process(
                         } catch (Exception ignored) {
                         }
                     });
-                    /* END */
+                    /* << loading font resources from current page */
+
+                    /* >> pre-check marked contents and remember orders */
+                    extractor.processPage(pdPage);
+
+                    List<PDMarkedContent> markedContents = extractor.getMarkedContents();
+                    List<Boolean> markedContentMatchRecords = markedContents
+                            .stream()
+                            .map(PDMarkedContent::getContents)
+                            .map(c -> c.stream()
+                                    .filter(e -> e instanceof TextPosition)
+                                    .map(e -> ((TextPosition) e).getUnicode())
+                                    .collect(Collectors.joining()))
+                            .map(s -> TextStampRecognizer.recognizePlain(strings, s.getBytes(), useStrict))
+                            .collect(Collectors.toList());
+
+                    needRewrite = markedContentMatchRecords.contains(true);
+                    /* << pre-check marked contents and remember orders */
+
+                    int mcCount = -1;
+                    boolean mcRemovingFlag = false;
 
                     /* handle both string array and string */
                     List<Object> objects = pdfStreamParser.getTokens();
                     Object object, prevObject;
                     for (int i = 0; i < objects.size(); i++) {
                         object = objects.get(i);
 
+                        /* >> mark marked content */
+                        if (object instanceof Operator) {
+                            Operator op = (Operator) object;
+                            if (op.getName().equals(BEGIN_MARKED_CONTENT_WITH_PROPERTY)) {
+                                ++mcCount;
+                                mcRemovingFlag = markedContentMatchRecords.get(mcCount);
+                            } else if (op.getName().equals(END_MARKDED_CONTENT)) {
+                                mcRemovingFlag = false;
+                                objects.set(i, empObj);
+                            }
+                        }
+
+                        if (mcRemovingFlag) {
+                            objects.set(i, empObj);
+                            continue;
+                        }
+                        /* << mark marked content */
+
                         if (object instanceof Operator) {
                             Operator op = (Operator) object;
                             String testStr;
@@ -131,6 +179,9 @@ public static void process(
                         OutputStream out = newContents.createOutputStream(COSName.FLATE_DECODE);
                         ContentStreamWriter writer = new ContentStreamWriter(out);
 
+                        /* remove all marked marked contents */
+                        objects.removeAll(empObjCol);
+
                         writer.writeTokens(objects);
                         out.close();
                         /* << write modified tokens back to the stream */

src/com/amastigote/unstamper/core/TextStampRecognizer.java

@@ -1,6 +1,6 @@
 /*
   AUTH | hwding
-  DATE | Dec 25 2018
+  DATE | Feb 19 2019
   DESC | textual watermark remover for PDF files
   MAIL | [email protected]
   GITH | github.com/hwding
@@ -42,7 +42,7 @@ private static boolean recognizeWithFont(
         return false;
     }
 
-    private static boolean recognizePlain(
+    static boolean recognizePlain(
             @NotNull String[] keywords,
             @NotNull byte[] inputText,
             @NotNull boolean useStrict