pdfParser: updated lib, fixed ClassNotFoundException: CMSError

git-svn-id: https://svn.berlios.de/svnroot/repos/yacy/trunk@4776 6c8d7289-2bf4-0310-a012-ef5d649a1542

.classpath

@@ -36,7 +36,8 @@
 	<classpathentry kind="lib" path="libx/commons-logging.jar"/>
 	<classpathentry kind="lib" path="libx/PDFBox-0.7.3.jar"/>
 	<classpathentry kind="lib" path="libx/FontBox-0.1.0-dev.jar"/>
-	<classpathentry kind="lib" path="libx/bcprov-jdk14-132.jar"/>
+	<classpathentry kind="lib" path="libx/bcprov-jdk14-139.jar"/>
+	<classpathentry kind="lib" path="libx/bcmail-jdk14-139.jar"/>
 	<classpathentry kind="lib" path="libx/commons-codec-1.3.jar"/>
 	<classpathentry kind="lib" path="lib/commons-httpclient-3.1.jar"/>
 	<classpathentry kind="output" path="gen"/>

source/de/anomic/plasma/parser/pdf/build.xml

@@ -18,7 +18,8 @@
 				<pathelement location="${libx}/PDFBox-0.7.3.jar" />
 				<!-- additional fontparsing lib, part of PDFBox -->
 				<pathelement location="${libx}/FontBox-0.1.0-dev.jar" />
-				<pathelement location="${libx}/bcprov-jdk14-132.jar" />
+				<pathelement location="${libx}/bcprov-jdk14-139.jar" />
+				<pathelement location="${libx}/bcmail-jdk14-139.jar" />
 			</classpath>
 		</javac>
 	</target>

source/de/anomic/plasma/parser/pdf/pdfParser.java

@@ -45,6 +45,7 @@ package de.anomic.plasma.parser.pdf;
 
 import java.io.File;
 import java.io.FileOutputStream;
+import java.io.IOException;
 import java.io.InputStream;
 import java.io.OutputStreamWriter;
 import java.io.Writer;
@@ -53,6 +54,7 @@ import java.util.Hashtable;
 import org.pdfbox.pdfparser.PDFParser;
 import org.pdfbox.pdmodel.PDDocument;
 import org.pdfbox.pdmodel.PDDocumentInformation;
+import org.pdfbox.pdmodel.encryption.AccessPermission;
 import org.pdfbox.pdmodel.encryption.StandardDecryptionMaterial;
 import org.pdfbox.util.PDFTextStripper;
 
@@ -62,6 +64,7 @@ import de.anomic.plasma.parser.AbstractParser;
 import de.anomic.plasma.parser.Parser;
 import de.anomic.plasma.parser.ParserException;
 import de.anomic.server.serverCharBuffer;
+import de.anomic.server.serverFileUtils;
 import de.anomic.yacy.yacyURL;
 
 public class pdfParser extends AbstractParser implements Parser {
@@ -78,7 +81,7 @@ public class pdfParser extends AbstractParser implements Parser {
      * @see Parser#getLibxDependences()
      */
     private static final String[] LIBX_DEPENDENCIES = new String[] {
-        "PDFBox-0.7.3.jar", "FontBox-0.1.0-dev.jar", "bcprov-jdk14-132.jar"
+        "PDFBox-0.7.3.jar", "FontBox-0.1.0-dev.jar", "bcprov-jdk14-139.jar"
     };        
     
     public pdfParser() {        
@@ -121,7 +124,8 @@ public class pdfParser extends AbstractParser implements Parser {
             
             if (theDocument.isEncrypted()) {
                 theDocument.openProtection(new StandardDecryptionMaterial(""));
-                if (!theDocument.getCurrentAccessPermission().canExtractContent())
+                final AccessPermission perm = theDocument.getCurrentAccessPermission();
+                if (perm == null || !perm.canExtractContent())
                     throw new ParserException("Document is encrypted",location,ErrorURL.DENIED_DOCUMENT_ENCRYPTED);
             }
             
@@ -193,6 +197,7 @@ public class pdfParser extends AbstractParser implements Parser {
             // delete the file
             if (writerFile != null) try { writerFile.delete(); } catch (Exception ex)  {/* ignore this */}
             
+            e.printStackTrace();
             throw new ParserException("Unexpected error while parsing pdf file. " + e.getMessage(),location); 
         } finally {
             if (theDocument != null) try { theDocument.close(); } catch (Exception e) {/* ignore this */}
@@ -205,5 +210,58 @@ public class pdfParser extends AbstractParser implements Parser {
         // Nothing todo here at the moment
         super.reset();
     }
+    
+    /**
+     * test
+     * @param args
+     */
+    public static void main(String[] args) {
+        if(args.length > 0 && args[0].length() > 0) {
+            // file
+            final File pdfFile = new File(args[0]);
+            if(pdfFile.canRead()) {
+                
+                System.out.println(pdfFile.getAbsolutePath());
+                final long startTime = System.currentTimeMillis();
+                
+                // parse
+                final AbstractParser parser = new pdfParser();
+                plasmaParserDocument document = null;
+                try {
+                    document = parser.parse(null, "application/pdf", null, pdfFile);
+                    
+                } catch (ParserException e) {
+                    System.err.println("Cannot parse file "+ pdfFile.getAbsolutePath());
+                    e.printStackTrace();
+                } catch (InterruptedException e) {
+                    System.err.println("Interrupted while parsing!");
+                    e.printStackTrace();
+                } catch (NoClassDefFoundError e) {
+                    System.err.println("class not found: " + e.getMessage());
+                }
+                
+                // statistics
+                System.out.println("\ttime elapsed: " + (System.currentTimeMillis() - startTime) + " ms");
+                
+                // output
+                if(document == null) {
+                    System.out.println("\t!!!Parsing without result!!!");
+                } else {
+                    System.out.println("\tParsed text with " + document.getTextLength() + " chars of text and " + document.getAnchors().size() + " anchors");
+                    try {
+                        // write file
+                        serverFileUtils.copy(document.getText(), new File("parsedPdf.txt"));
+                    } catch (IOException e) {
+                        System.err.println("error saving parsed document");
+                        e.printStackTrace();
+                    }
+                }
+            } else {
+                System.err.println("Cannot read file "+ pdfFile.getAbsolutePath());
+            }
+        } else {
+            System.out.println("Please give a filename as first argument.");
+        }
+    }
 
 }