1 files changed, 956 insertions, 0 deletions
diff --git a/pdf-as-lib/src/main/java/at/knowcenter/wag/egov/egiz/pdf/AbsoluteTextSignature.java b/pdf-as-lib/src/main/java/at/knowcenter/wag/egov/egiz/pdf/AbsoluteTextSignature.java
new file mode 100644
index 0000000..fd59d34
--- /dev/null
+++ b/pdf-as-lib/src/main/java/at/knowcenter/wag/egov/egiz/pdf/AbsoluteTextSignature.java
@@ -0,0 +1,956 @@
+/**
+ * <copyright> Copyright 2006 by Know-Center, Graz, Austria </copyright>
+ * PDF-AS has been contracted by the E-Government Innovation Center EGIZ, a
+ * joint initiative of the Federal Chancellery Austria and Graz University of
+ * Technology.
+ *
+ * Licensed under the EUPL, Version 1.1 or - as soon they will be approved by
+ * the European Commission - subsequent versions of the EUPL (the "Licence");
+ * You may not use this work except in compliance with the Licence.
+ * You may obtain a copy of the Licence at:
+ * http://www.osor.eu/eupl/
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the Licence is distributed on an "AS IS" basis,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the Licence for the specific language governing permissions and
+ * limitations under the Licence.
+ *
+ * This product combines work with different licenses. See the "NOTICE" text
+ * file for details on the various modules and licenses.
+ * The "NOTICE" text file is part of the distribution. Any derivative works
+ * that you distribute must include a readable copy of the "NOTICE" text file.
+ *
+ * $Id: AbsoluteTextSignature.java,v 1.1 2006/10/31 08:08:33 wprinz Exp $
+ */
+package at.knowcenter.wag.egov.egiz.pdf;
+
+import java.util.ArrayList;
+import java.util.Collections;
+import java.util.Comparator;
+import java.util.Iterator;
+import java.util.List;
+import java.util.Vector;
+
+import org.apache.log4j.Logger;
+
+import at.knowcenter.wag.egov.egiz.PdfAS;
+import at.knowcenter.wag.egov.egiz.cfg.ConfigLogger;
+import at.knowcenter.wag.egov.egiz.exceptions.SignatureException;
+import at.knowcenter.wag.egov.egiz.exceptions.SignatureTypesException;
+import at.knowcenter.wag.egov.egiz.framework.FoundBlock;
+import at.knowcenter.wag.egov.egiz.framework.FoundKey;
+import at.knowcenter.wag.egov.egiz.sig.SignatureObject;
+import at.knowcenter.wag.egov.egiz.sig.SignatureTypeDefinition;
+import at.knowcenter.wag.egov.egiz.sig.SignatureTypes;
+
+/**
+ * Contains methods and helpers that implement the absolute text signature.
+ * 
+ * @author wprinz
+ */
+public class AbsoluteTextSignature
+{
+
+  /**
+   * The logger definition.
+   */
+  private static final Logger logger = ConfigLogger.getLogger(AbsoluteTextSignature.class);
+
+  /**
+   * Returns a List of SignatureTypeDefinitions that can be extracted from text.
+   * 
+   * <p>
+   * These SignatureTypeDefinitions are all text extractable, which means that they define all required fields as visible.
+   * </p>
+   * <p>
+   * This method filters out all SignatureTypeDefinitions that are not text extractable.
+   * </p>
+   * 
+   * @return Returns a List of SignatureTypeDefinitions that can be extracted from text.
+   * @throws SignatureTypesException F.e.
+   */
+  public static List getSignatureTypesForTextAnalysis() throws SignatureTypesException
+  {
+    SignatureTypes sig_types = SignatureTypes.getInstance();
+    List allSignatureTypes = sig_types.getSignatureTypeDefinitions();
+    
+    List textSignatureTypes = new ArrayList(allSignatureTypes.size());
+    Iterator it = allSignatureTypes.iterator();
+    while (it.hasNext())
+    {
+      SignatureTypeDefinition std = (SignatureTypeDefinition) it.next();
+      if (!std.isTextExtractable())
+      {
+        logger.debug("The profile " + std.getType() + " is not text extractable and is thereby not used for text analysis.");
+        continue;
+      }
+      textSignatureTypes.add(std);
+    }
+
+    return textSignatureTypes;
+  }
+  
+  /**
+   * Extracts all signature holders from a given text.
+   * 
+   * <p>
+   * First the latest signature holder is extracted. Then the latest signature
+   * holder in the rest text, which is the second latest one, is extracted. Then
+   * the third latest signature holder is extracted and so forth until no more
+   * signature holders are found.
+   * </p>
+   * 
+   * @param text
+   *          The text.
+   * @return Returns the List of extracted signature holders ordered by their
+   *         date ascendingly (the lowest, earliest date first, the latest,
+   *         newest date last). An empty list is returned if no signature
+   *         holders were found.
+   * @throws SignatureException
+   *           F.e.
+   * @throws SignatureTypesException
+   *           F.e.
+   */
+  public static List extractSignatureHoldersFromText(String text) throws SignatureException, SignatureTypesException
+  {
+    List holders = new ArrayList();
+    String current_text = text;
+    for (;;)
+    {
+      TextualSignatureHolder signature_holder = extractLatestBlock(current_text);
+      if (signature_holder == null)
+      {
+        break;
+      }
+      holders.add(0, signature_holder);
+      current_text = signature_holder.getSignedText();
+    }
+    return holders;
+  }
+
+  /**
+   * Extracts the latest signature block from the given text and creates a
+   * SignatureHolder object that can be verified.
+   * 
+   * @param text
+   *          The text.
+   * @return Returns the SignatureObject extracted from the text, or null, if no
+   *         latest block was found.
+   * @throws SignatureException
+   *           F.e.
+   * @throws SignatureTypesException
+   *           F.e.
+   */
+  public static TextualSignatureHolder extractLatestBlock(String text) throws SignatureException, SignatureTypesException
+  {
+    FoundBlock latest_block = findLatestBlock(text);
+    if (latest_block == null)
+    {
+      return null;
+    }
+    String reconstructed_text = cutOutBlock(text, latest_block);
+    SignatureObject so = createSignatureObjectFromFoundBlock(text, latest_block);
+    TextualSignatureHolder tsh = new TextualSignatureHolder(reconstructed_text, so);
+    return tsh;
+  }
+
+  /**
+   * Finds the latest signature block for a given text.
+   * 
+   * <p>
+   * The latest block is the one with the highest, most recent date. Usually
+   * this block will be extracted (cut out) of the text which will result in the
+   * originally signed text of this signature to be verified using the cut out
+   * data.
+   * </p>
+   * 
+   * @param text
+   *          The text to be analyzed.
+   * @return Returns the latest found block or null, if there was none.
+   * @throws SignatureException
+   *           F.e.
+   * @throws SignatureTypesException
+   *           F.e.
+   */
+  public static FoundBlock findLatestBlock(String text) throws SignatureException, SignatureTypesException
+  {
+    List signatureTypes_ = getSignatureTypesForTextAnalysis();
+    
+    List found_potential_candidates = new ArrayList();
+ 
+    for (int i = 0; i < signatureTypes_.size(); i++)
+    {
+      SignatureTypeDefinition block_type = (SignatureTypeDefinition) signatureTypes_.get(i);
+      List found_candidates_for_type = findPotentialSignaturesForProfile(text, block_type);
+
+      found_potential_candidates.addAll(found_candidates_for_type);
+    }
+
+    if (found_potential_candidates.isEmpty())
+    {
+      logger.debug("no candidates found at all");
+      return null;
+    }
+
+    List found_candidates = new ArrayList();
+    logger.debug("checking block integrity");
+    for (int i = 0; i < found_potential_candidates.size(); i++)
+    {
+      FoundBlock found_block = (FoundBlock) found_potential_candidates.get(i);
+      String date_value = getDateValue(text, found_block);
+      try
+      {
+        EGIZDate date = EGIZDate.parseFromString(date_value);
+
+        logger.debug("found_block = " + date + " - " + found_block);
+
+        checkBlockIntegrity(text, found_block);
+        found_candidates.add(found_block);
+      }
+      catch (Exception e)
+      {
+        logger.debug("Exception while checking the integrity of the found block " + found_block + ". Ignoring this block.", e);
+      }
+    }
+
+    sortFoundBlocksByDate(text, found_candidates);
+    if (logger.isDebugEnabled())
+    {
+      logger.debug("sorted blocks:");
+      for (int i = 0; i < found_candidates.size(); i++)
+      {
+        FoundBlock found_block = (FoundBlock) found_candidates.get(i);
+
+        String date_value = getDateValue(text, found_block);
+        EGIZDate date = EGIZDate.parseFromString(date_value);
+
+        logger.debug("  #" + i + ": " + date + " - " + found_block);
+      }
+    }
+
+    List latest_blocks = filterLastDateEqualBlocks(text, found_candidates);
+    if (logger.isDebugEnabled())
+    {
+      logger.debug("latest blocks:");
+      for (int i = 0; i < latest_blocks.size(); i++)
+      {
+        FoundBlock found_block = (FoundBlock) latest_blocks.get(i);
+
+        String date_value = getDateValue(text, found_block);
+        EGIZDate date = EGIZDate.parseFromString(date_value);
+
+        logger.debug("  #" + i + ": " + date + " - " + found_block);
+      }
+    }
+
+    // The semantic equality check has been outdated by the
+    // advanced choosing algorithm.
+    // boolean semantic_equality =
+    // PdfAS.checkForSemanticEquality(latest_blocks);
+    // logger.debug("semantic_equality = " + semantic_equality);
+    // if (!semantic_equality)
+    // {
+    // throw new SignatureException(314, "The latest blocks weren't semantically
+    // equal.");
+    // }
+
+    FoundBlock latest_block = chooseMostPossibleBlock(latest_blocks);
+
+    logger.debug("latest block = " + latest_block);
+    return latest_block;
+  }
+
+  /**
+   * Finds the List of potential blocks within the given text for the given
+   * profile.
+   * 
+   * @param text
+   *          The text, in which potential block are to be sought.
+   * @param block_type
+   *          The profile for which the text is to be sought.
+   * @return Returns the List of potential FoundBlocks or an empty List if none
+   *         could be found.
+   */
+  public static List findPotentialSignaturesForProfile(String text,
+      SignatureTypeDefinition block_type)
+  {
+    logger.debug("find potential signatures for " + block_type.getType());
+    List found_blocks = new ArrayList();
+
+    final boolean old_style = false;
+
+    Vector keys = block_type.getRevertSortedKeys();
+    Vector captions = block_type.getRevertSortedCaptions();
+
+    String last_key = (String) keys.get(0);
+    logger.debug("last_key = " + last_key);
+    String last_caption = (String) captions.get(0);
+    logger.debug("last_caption = " + last_caption);
+    String current_last_caption= last_caption;
+    List found_last_captions = findIndicesWithStartingNL(text, last_caption);
+    if (last_key.equals(SignatureTypes.SIG_ID))
+    {
+      logger.debug("Last key is SIG_ID, so it may not be present. Searching for the previous to last key.");
+      String prevlast_key = (String) keys.get(1);
+      String prevlast_caption = (String) captions.get(1);
+      current_last_caption = prevlast_caption; 
+      List found_prevlast_captions = findIndicesWithStartingNL(text, prevlast_caption);
+      if (!found_prevlast_captions.isEmpty())
+      {
+        found_last_captions.addAll(found_prevlast_captions);
+      }
+    }
+    if (logger.isDebugEnabled())
+    {
+      logger.debug("found " + found_last_captions.size() + " last captions.");
+      for (int i = 0; i < found_last_captions.size(); i++)
+      {
+        logger.debug("  found last caption at index " + found_last_captions.get(i));
+      }
+    }
+
+    for (int lci = 0; lci < found_last_captions.size(); lci++)
+    {
+      int last_caption_index = ((Integer) found_last_captions.get(lci)).intValue();
+      logger.debug("resolving signature block from last caption index " + last_caption_index);
+      int potential_block_end = findEndOfValue(text, last_caption_index);
+      if (potential_block_end == (last_caption_index + current_last_caption.length()+1))
+      {
+    	  potential_block_end = findEndOfValue(text, potential_block_end);
+      }
+      
+      // FIXME: complete HOTFIX
+      /*
+      int extendedValueEnd = potential_block_end;
+      String cv;
+      do {
+         extendedValueEnd = findEndOfValue(text, extendedValueEnd);
+         cv = text.substring(last_caption_index + current_last_caption.length()+1, extendedValueEnd);
+      } while (extendedValueEnd < text.length());
+      */
+      
+      logger.debug("potential_block_end = " + potential_block_end);
+      List found_keys = PdfAS.findBlockInText(text.substring(0, potential_block_end), block_type, old_style); // findRestKeys(text,
+      // keys,
+      // captions,
+      // last_caption_index);
+
+      if (found_keys == null)
+      {
+        logger.debug("Not all other captions could be found for the last_caption_index " + last_caption_index + " ==> discarding this index.");
+
+        continue;
+      }
+
+      // sort found keys ascendingly
+      PdfAS.sortFoundKeysAscendingly(found_keys);
+
+      boolean reverse_check_ok = reverseCheckFoundKeys(text, found_keys);
+      if (!reverse_check_ok)
+      {
+        logger.debug("The reverse check ruled this list of found keys out ==> they are discarded.");
+
+        continue;
+      }
+
+      logger.debug("The reverse check proved this list of found keys out ==> adding them as potential candidates.");
+
+      FoundBlock found_block = new FoundBlock();
+      found_block.std = block_type;
+      found_block.found_keys = found_keys;
+      found_block.end_index = potential_block_end;//findEndOfValue(text, last_caption_index);
+      found_blocks.add(found_block);
+    }
+
+    logger.debug("found " + found_blocks.size() + " potential signatures for " + block_type.getType());
+    return found_blocks;
+  }
+
+  /**
+   * Finds all indices of the given subtext (starting at a new line) within a
+   * given text.
+   * 
+   * <p>
+   * This is usually used to find the indices of the last captions.
+   * </p>
+   * 
+   * @param text
+   *          The text to be searched.
+   * @param subtext
+   *          The subtext to be sought.
+   * @return Returns the List of found indices.
+   */
+  public static List findIndicesWithStartingNL(String text, String subtext)
+  {
+    List found_indices = new ArrayList();
+
+    // // for some reason "^" + subtext doesn't work as a pattern
+    // String pattern = "\n" + subtext;
+    // Pattern p = Pattern.compile(pattern);
+    // Matcher m = p.matcher(text);
+    //    
+    // while (m.find())
+    // {
+    // int found_index = m.start() + 1; // +1 removes the newline
+    // found_indices.add(new Integer(found_index));
+    // }
+
+    int search_from_index = 0;
+    for (;;)
+    {
+      int found_index = text.indexOf("\n" + subtext, search_from_index);
+      if (found_index < 0)
+      {
+        break;
+      }
+      found_index += 1; // The +1 compensates the "\n"
+      found_indices.add(new Integer(found_index));
+      search_from_index = found_index + subtext.length();
+    }
+    return found_indices;
+  }
+
+  /**
+   * Finds the other keys/captions according to their order starting from the
+   * last_caption index upwards.
+   * 
+   * @param text
+   *          The text.
+   * @param keys
+   *          The list of keys.
+   * @param captions
+   *          The list of captions.
+   * @param last_caption_index
+   *          The index of the last caption.
+   * @return Returns the List of found keys, if all keys could be found, or null
+   *         if not all keys could be found.
+   */
+  public static List findRestKeys(String text, List keys, List captions,
+      int last_caption_index)
+  {
+    List found_keys = new ArrayList();
+
+    FoundKey last_caption_found_key = new FoundKey((String) keys.get(0), (String) captions.get(0), last_caption_index);
+    found_keys.add(last_caption_found_key);
+
+    String rest_text = text.substring(0, last_caption_index);
+
+    for (int i = 1; i < captions.size(); i++)
+    {
+      String sought_caption = (String) captions.get(i);
+      int index = rest_text.lastIndexOf(sought_caption);
+
+      if (index < 0)
+      {
+        return null;
+      }
+      FoundKey found_key = new FoundKey((String) keys.get(i), (String) captions.get(i), index);
+      found_keys.add(0, found_key);
+
+      rest_text = rest_text.substring(0, index);
+    }
+
+    return found_keys;
+  }
+
+  /**
+   * Performs a reverse (top to bottom) search for the found keys and checks
+   * that these indices are the same as those that were found during the regular
+   * (bottom up) search.
+   * <p>
+   * If a reverse check proves that the found keys are not at the same positions
+   * as during regular search, this list of found keys should be discarded.
+   * </p>
+   * 
+   * @param text
+   *          The text.
+   * @param found_keys
+   *          The found keys to be reversely checked.
+   * @return Returns true, if all (also the non required) captions could be
+   *         found at the same indices as during regular search, false
+   *         otherwise.
+   */
+  public static boolean reverseCheckFoundKeys(String text, List found_keys)
+  {
+    int search_from_index = ((FoundKey) found_keys.get(0)).start_index;
+    for (int i = 0; i < found_keys.size(); i++)
+    {
+      FoundKey found_key = (FoundKey) found_keys.get(i);
+
+      int reverse_found_index = text.indexOf(found_key.caption, search_from_index);
+      
+      // dferbas fix #331 ??
+      
+      if (reverse_found_index < 0)
+      {
+        throw new RuntimeException("The caption " + found_key.caption + " wasn't found in the text during reverse checking - there is something wrong.");
+      }
+      
+      if (reverse_found_index != found_key.start_index)
+      {
+        logger.debug("The index for caption " + found_key.caption + " wasn't proved during reverse checking.");
+        return false;
+      }
+      search_from_index = found_key.start_index + found_key.caption.length();
+    }
+    return true;
+  }
+
+  /**
+   * Finds the end of the value in the text.
+   * 
+   * <p>
+   * This simply scans for a '\n' from a given start index. The line up to and
+   * inclusive the '\n' is considered to be the value.
+   * </p>
+   * <p>
+   * Note that this method does NOT find the accurate value, if the value goes
+   * over multiple lines! This may bear a serious problem. Usually this method
+   * is only used to finding the end of the last value in a found block, because
+   * mid- values are exactly determined by their start index and the start of
+   * the next caption. Nevertheless, if the last value spans over multiple
+   * lines, this method will not retrieve it completely.
+   * </p>
+   * 
+   * @param text
+   *          The text.
+   * @param start_index
+   *          The start index from where the end of the value is sought.
+   * @return Returns the end index of the value, which is the index of the first
+   *         character not belonging to the value anymore (the character after
+   *         the '\n').
+   */
+  public static int findEndOfValue(String text, int start_index)
+  {
+    int newline_index = text.indexOf('\n', start_index);
+    if (newline_index < 0)
+    {
+      return text.length();
+    }   
+    return newline_index + 1;
+  }
+ 
+  /**
+   * Checks the integrity of a found block.
+   * 
+   * <p>
+   * This is an assertive function.
+   * </p>
+   * 
+   * @param text
+   *          The text.
+   * @param found_block
+   *          The found block.
+   */
+  public static void checkBlockIntegrity(String text, FoundBlock found_block)
+  {
+    for (int i = 0; i < found_block.found_keys.size() - 1; i++)
+    {
+      FoundKey this_key = (FoundKey) found_block.found_keys.get(i);
+      FoundKey next_key = (FoundKey) found_block.found_keys.get(i + 1);
+
+      int this_end_index = findEndOfValue(text, this_key.start_index);
+      if (this_end_index != next_key.start_index)
+      {
+        logger.debug("multi line value: " + this_key);
+        // throw new RuntimeException("The end index of found key " + this_key +
+        // " doesn't match the start index of found key " + next_key);
+      }
+    }
+
+    FoundKey last_key = (FoundKey) found_block.found_keys.get(found_block.found_keys.size() - 1);
+    int end_of_block =  findEndOfValue(text, last_key.start_index);
+    if (end_of_block == (last_key.start_index+last_key.caption.length()+1))
+    {
+    	end_of_block = findEndOfValue(text,end_of_block);
+    }
+    if (end_of_block != found_block.end_index)
+    {
+        throw new RuntimeException("The end index of last key " + last_key + " doesn't match the end index of the block " + found_block);
+    }
+  }
+
+  /**
+   * Cuts out the given found block from the text.
+   * 
+   * @param text
+   *          The text.
+   * @param block
+   *          The found block.
+   * @return Returns the rest text without the block.
+   */
+  public static String cutOutBlock(String text, FoundBlock block)
+  {
+    int block_start_index = ((FoundKey) block.found_keys.get(0)).getStartIndex();
+    int block_end_index = block.end_index;
+    
+    if (block_start_index == 0 && block_end_index == text.length())
+    {
+      // the block is the whole text - the rest text is empty.
+      // This may happen if a (no-text) empty document contains a binary signature.
+      // Then the "signed text" of the binary signature is empty.
+      return "";
+    }
+
+    if (block_end_index == text.length())
+    {
+      // if the block is at the end of the text, remove the "\n" before the
+      // block as well.
+      String pre = text.substring(0, block_start_index - 1);
+      return pre;
+    }
+
+    String pre = text.substring(0, block_start_index);
+    String post = text.substring(block_end_index);
+
+    String rest_text = pre + post;
+    return rest_text;
+  }
+
+  /**
+   * Returns the value of the date field as String.
+   * 
+   * @param text
+   *          The text.
+   * @param block
+   *          The found block.
+   * @return Returns the date value.
+   */
+  public static String getDateValue(String text, FoundBlock block)
+  {
+    FoundKey date_key = block.getDateFoundKey();
+    int date_value_start_index = date_key.start_index + date_key.caption.length();
+    int date_value_end_index = findEndOfValue(text, date_value_start_index);
+    if (date_value_end_index == (date_value_start_index+1))
+    {
+    	date_value_end_index = findEndOfValue(text, date_value_end_index);	
+    }
+    String date_value = text.substring(date_value_start_index, date_value_end_index).trim();
+    logger.debug("DateString="+date_value);
+    return date_value;
+  }
+
+  /**
+   * Creates a SignatureObject from a found block by extracting the
+   * corresponding values.
+   * 
+   * @param text
+   *          The text.
+   * @param found_block
+   *          The found block.
+   * @return Returns the created SignatureObject.
+   * @throws SignatureTypesException
+   *           F.e.
+   * @throws SignatureException
+   *           F.e.
+   */
+  public static SignatureObject createSignatureObjectFromFoundBlock(
+      String text, FoundBlock found_block) throws SignatureTypesException, SignatureException
+  {
+    SignatureObject signatureObject = new SignatureObject();
+
+    signatureObject.setSigType(found_block.std.getType());
+    signatureObject.initByType();
+
+    int end_index = found_block.end_index;
+    for (int i = found_block.found_keys.size() - 1; i >= 0; i--)
+    {
+      FoundKey cur_key = (FoundKey) found_block.found_keys.get(i);
+      int start_index = cur_key.getStartIndex() + cur_key.caption.length();
+
+      String value = text.substring(start_index, end_index);
+
+      signatureObject.setSigValueCaption(cur_key.getKey(), value, cur_key.caption);
+
+      end_index = cur_key.getStartIndex();
+    }
+
+    return signatureObject;
+
+  }
+
+  /**
+   * Parses the EGIZDate from a found block and the given text.
+   * 
+   * @param text
+   *          The text.
+   * @param found_block
+   *          The found block.
+   * @return Returns the parsed EGIZDate.
+   */
+  public static EGIZDate getDateFromFoundBlock(String text,
+      FoundBlock found_block)
+  {
+    String date_value = getDateValue(text, found_block);
+    EGIZDate date = EGIZDate.parseFromString(date_value);
+    return date;
+  }
+
+  /**
+   * Sorts the List of found blocks by date.
+   * 
+   * @param text
+   *          The text.
+   * @param found_blocks
+   *          The List of found blocks.
+   */
+  public static void sortFoundBlocksByDate(final String text, List found_blocks)
+  {
+    Collections.sort(found_blocks, new Comparator()
+    {
+      public int compare(Object arg0, Object arg1)
+      {
+        FoundBlock fb0 = (FoundBlock) arg0;
+        FoundBlock fb1 = (FoundBlock) arg1;
+
+        EGIZDate date0 = getDateFromFoundBlock(text, fb0);
+        EGIZDate date1 = getDateFromFoundBlock(text, fb1);
+        return date0.compareTo(date1);
+      }
+    });
+  }
+
+  /**
+   * Given a List of FoundBlock objects, this method returns the last blocks of
+   * this list that have the same date.
+   * 
+   * <p>
+   * Usually a date sorted list (earliest first, latest last) will be provided
+   * to this method. Then the last date equal blocks are returned, which are the
+   * last blocks.
+   * </p>
+   * 
+   * @param text
+   *          The text to retrieve the values of the fields from.
+   * @param found_blocks
+   *          The List of FoundBlock objects.
+   * @return Returns the List of the last date equal blocks.
+   */
+  public static List filterLastDateEqualBlocks(String text, List found_blocks)
+  {
+    List latest_blocks = new ArrayList();
+    latest_blocks.add(found_blocks.get(found_blocks.size() - 1));
+    for (int i = found_blocks.size() - 2; i >= 0; i--)
+    {
+      FoundBlock this_block = (FoundBlock) found_blocks.get(i);
+      FoundBlock succ_block = (FoundBlock) found_blocks.get(i + 1);
+
+      EGIZDate this_date = getDateFromFoundBlock(text, this_block);
+      EGIZDate succ_date = getDateFromFoundBlock(text, succ_block);
+
+      if (!this_date.equals(succ_date))
+      {
+        break;
+      }
+      latest_blocks.add(0, this_block);
+    }
+
+    return latest_blocks;
+  }
+  /**
+   * Chooses the most possible (best choice) block of the list of blocks.
+   * 
+   * <p>
+   * The strategy to find the most possible block is to choose the very one
+   * block with the maximum number of captions. This block has extracted most
+   * information from the text.
+   * </p>
+   * <p>
+   * If there are still multiple blocks with the same number of cations, the
+   * blocks are compared caption-wise. The block with all captions being longer
+   * or equal to all other blocks' captions wins.
+   * </p>
+   * 
+   * @param found_blocks
+   *          The List of semantically equal blocks.
+   * @return Returns the best choice FoundBlock.
+   * @throws SignatureException
+   */
+  public static FoundBlock chooseMostPossibleBlock(List found_blocks) throws SignatureException
+  {
+    // int largest_block_index = 0;
+    // FoundBlock largest_block = (FoundBlock) found_blocks.get(0);
+    //
+    // for (int i = 1; i < found_blocks.size(); i++)
+    // {
+    // FoundBlock current_block = (FoundBlock) found_blocks.get(i);
+    //
+    // if (current_block.found_keys.size() > largest_block.found_keys.size())
+    // {
+    // largest_block = current_block;
+    // largest_block_index = i;
+    // }
+    // }
+
+    List vertically_largest = filterVerticallyLargestBlocks(found_blocks);
+    if (logger.isDebugEnabled())
+    {
+      logger.debug("vertically largest blocks:");
+      for (int i = 0; i < vertically_largest.size(); i++)
+      {
+        FoundBlock found_block = (FoundBlock) vertically_largest.get(i);
+        logger.debug("  #" + i + ": " + found_block);
+      }
+    }
+
+    List horizontally_largest = filterHorizontallyLargestBlocks(vertically_largest);
+    if (logger.isDebugEnabled())
+    {
+      logger.debug("horizontally largest blocks:");
+      for (int i = 0; i < horizontally_largest.size(); i++)
+      {
+        FoundBlock found_block = (FoundBlock) horizontally_largest.get(i);
+        logger.debug("  #" + i + ": " + found_block);
+      }
+    }
+    FoundBlock largest_block = (FoundBlock) horizontally_largest.get(0);
+    logger.debug("Chose largest block: " + largest_block);
+    return largest_block;
+  }
+
+  /**
+   * Filters out all blocks but the vertically largest ones.
+   * 
+   * <p>
+   * A vertically largest block has the most found keys.
+   * </p>
+   * 
+   * @param found_blocks
+   *          The List of FoundBlock objects to be filtered.
+   * @return Returns the List of the vertically largest FoundBlock objects.
+   */
+  public static List filterVerticallyLargestBlocks(List found_blocks)
+  {
+    // determine the size of the largest block(s)
+    int largest_size = Integer.MIN_VALUE;
+    for (int i = 0; i < found_blocks.size(); i++)
+    {
+      FoundBlock fb = (FoundBlock) found_blocks.get(i);
+      final int current_size = fb.found_keys.size();
+      if (current_size > largest_size)
+      {
+        largest_size = current_size;
+      }
+    }
+
+    // keep all blocks that have the largest_size
+    List largest_blocks = new ArrayList();
+    for (int i = 0; i < found_blocks.size(); i++)
+    {
+      FoundBlock fb = (FoundBlock) found_blocks.get(i);
+      if (fb.found_keys.size() < largest_size)
+      {
+        continue;
+      }
+      largest_blocks.add(fb);
+    }
+    return largest_blocks;
+  }
+
+  /**
+   * Filters out all blocks but the horizonally largest ones.
+   * 
+   * <p>
+   * A vertically largest block has the most found keys.
+   * </p>
+   * 
+   * @param found_blocks
+   *          The List of FoundBlock objects to be filtered. All of these
+   *          FoundBlock objects must have the same number of found keys.
+   * @return Returns the List of the horizontally largest FoundBlock objects.
+   * @throws SignatureException
+   */
+  public static List filterHorizontallyLargestBlocks(List found_blocks) throws SignatureException
+  {
+    List horizontally_largest = new ArrayList();
+    FoundBlock largest_block = (FoundBlock) found_blocks.get(0);
+    horizontally_largest.add(largest_block);
+
+    for (int i = 1; i < found_blocks.size(); i++)
+    {
+      FoundBlock fb = (FoundBlock) found_blocks.get(i);
+
+      if (isHorizontallyEqual(fb, largest_block))
+      {
+        horizontally_largest.add(fb);
+        continue;
+      }
+
+      if (isHorizontallyLarger(fb, largest_block))
+      {
+        horizontally_largest = new ArrayList();
+        largest_block = fb;
+        horizontally_largest.add(largest_block);
+      }
+      else
+      {
+        if (!isHorizontallyLarger(largest_block, fb))
+        {
+          // The block is neither equal nor larger nor lower.
+          // We cannot exactly determine which one to use.
+          throw new SignatureException(315, "The blocks are neither larger nor lower nor equal. Cannot decide which one to pick. fb = " + fb + ", largest_block = " + largest_block);
+        }
+      }
+
+    }
+
+    return horizontally_largest;
+  }
+
+  protected static boolean isHorizontallyEqual(FoundBlock fb0, FoundBlock fb1)
+  {
+    final int num_keys = fb0.found_keys.size();
+    if (num_keys != fb1.found_keys.size())
+    {
+      throw new IllegalArgumentException("Cannot compare FoundBlock keys: fb0 doesn't have the same number of keys as fb1. " + fb0.found_keys.size() + " vs. " + fb1.found_keys.size());
+    }
+
+    for (int i = 0; i < num_keys; i++)
+    {
+      FoundKey fk0 = (FoundKey) fb0.found_keys.get(i);
+      FoundKey fk1 = (FoundKey) fb1.found_keys.get(i);
+
+      if (fk0.caption.length() != fk1.caption.length())
+      {
+        return false;
+      }
+    }
+
+    return true;
+  }
+
+  protected static boolean isHorizontallyLarger(FoundBlock fb0, FoundBlock fb1)
+  {
+    final int num_keys = fb0.found_keys.size();
+    if (num_keys != fb1.found_keys.size())
+    {
+      throw new IllegalArgumentException("Cannot compare FoundBlock keys: fb0 doesn't have the same number of keys as fb1. " + fb0.found_keys.size() + " vs. " + fb1.found_keys.size());
+    }
+
+    boolean larger = false;
+
+    for (int i = 0; i < num_keys; i++)
+    {
+      FoundKey fk0 = (FoundKey) fb0.found_keys.get(i);
+      FoundKey fk1 = (FoundKey) fb1.found_keys.get(i);
+
+      if (fk0.caption.length() == fk1.caption.length())
+      {
+        continue;
+      }
+
+      if (fk0.caption.length() > fk1.caption.length())
+      {
+        larger = true;
+        continue;
+      }
+
+      // if (fk0.caption.length() < fk1.caption.length())
+      return false;
+    }
+
+    return larger;
+  }
+
+}