/*
   Copyright (c) 2005 The European Bioinformatics Institute, and others.
   All rights reserved. Please see the file LICENSE
   in the root directory of this distribution.
   */
   package uk.ac.ebi.intenz.tools.sib.translator.rules;
   
   import java.io.BufferedReader;
   import java.io.IOException;
  import java.io.InputStream;
  import java.io.InputStreamReader;
  import java.util.ArrayList;
  import java.util.regex.Matcher;
  import java.util.regex.Pattern;
  
  import org.apache.log4j.Logger;
  
  
  /**
   * GrammarRules
   *
   * @author P. de Matos
   * @version $id 15-Jul-2005 14:41:31
   *          <p/>
   *          History:<br>
   *          <table>
   *          <tr><th>Developer</th><th>Date</th><th>Description</th></tr>
   *          <tr><td>P.de Matos</td><td>15-Jul-2005</td><td>Created class</td></tr>
   *          </table>
   */
  public class GrammarRules implements RuleGroup {
  
     private static final Logger LOGGER = Logger.getLogger(GrammarRules.class);
  
    private static GrammarRules INSTANCE = new GrammarRules();
  
     private final String[][] CAPITILISATION_RULES_REVERSED = {
       {"^TRNA", "tRNA"},
       {"^MRNA", "mRNA"},
       {"^RRNA", "rRNA"},
       {"^CDNA", "cDNA"},
       {"^DAMP", "dAMP"},
       {"^DADP", "dADP"},
       {"^DATP", "dATP"},
       {"^DUMP", "dUMP"},
       {"^DUDP", "dUDP"},
       {"^DUTP", "dUTP"},
       {"^DCMP", "dCMP"},
       {"^DCDP", "dCDP"},
       {"^DCTP", "dCTP"},
       {"^DGMP", "dGMP"},
       {"^DGDP", "dGDP"},
       {"^DGTP", "dGTP"},
       {"^DTMP", "dTMP"},
       {"^DTDP", "dTDP"},
       {"^DTTP", "dTTP"},
       {"^CAMP", "cAMP"},
       {"^CGMP", "cGMP"}
  
    };
  
     // These are done elsewhere so we exclude them
     private final String[] CAPITILISATION_EXCLUSIONS = {
        "^p\\-",
        "^m\\-",
        "^o\\-",
        "^n\\-",
        "^tRNA",
        "^tRNase",
        "^mRNA",
        "^rRNA",
        "^cDNA",
        "^dAMP",
        "^dADP",
        "^dATP",
        "^dUMP",
        "^dUDP",
        "^dUTP",
        "^dCMP",
        "^dCDP",
        "^dCTP",
        "^dGMP",
        "^dGDP",
        "^dGTP",
        "^dTMP",
        "^dTDP",
        "^dTTP",
        "^cAMP",
        "^cGMP",
        "^ppGpp",
        "^pppGpp",
        "^\\(ppGpp\\)",
        "^\\(fMet\\)",
        "^fMet",
        "^n\\s",
        "^m\\s",
        "^tPA",
        "^(\\[|\\(|\\s)eIF",
        "^sn-",
       "^\\d*?\\([\\+\\-\\d]+?\\)\\-",
       "^cd-"
    };
 
    private final Pattern[] SNEAKY_CAPIT_PATTERNS = {
        Pattern.compile("^(\\[|\\()(\\p{Lower})(\\p{Lower})")
    };
 
 
    private final Pattern SNEAKY_CAPIT_PATTERN = Pattern.compile("^(\\[|\\()(\\w)");
 
    private final ArrayList DECAPITALIZATION_PATTERNS = new ArrayList();
 
    private GrammarRules () {
         InputStream stream = this.getClass().getClassLoader().getResourceAsStream("decapitalization_patterns.txt");
         BufferedReader br = new BufferedReader(new InputStreamReader(stream));
         String pattern;
         try {
             pattern = br.readLine();
             while (pattern != null){
                 DECAPITALIZATION_PATTERNS.add(pattern);
                 pattern = br.readLine();
             }
         } catch (IOException e) {
             e.printStackTrace();
         } finally {
             try {
                 br.close();
                 stream.close();
             } catch (IOException e){}
         }
     }
 
    /**
      * Returns the sole instance of this class.
      * <p/>
      * If no instance is available yet then it will be created.
      *
      * @return the class's sole instance.
      */
     public static GrammarRules getInstance() {
       return INSTANCE;
     }
 
 
    public String applyRules (String text) {
     return reverseRules(text);
    }
 
    public String reverseRules(String text) {
         if (text == null)
             throw new NullPointerException("Parameter 'text' must not be null.");
         if (text.length() == 0)
             return text;
 
         // Capitalise the first letter of a line.
         boolean excludeFromCapitilisation = false;
         for (int iii = 0; iii < CAPITILISATION_EXCLUSIONS.length; iii++) {
             String exclusion = CAPITILISATION_EXCLUSIONS[iii];
             if (Pattern.compile(exclusion).matcher(text).find()) {
                 excludeFromCapitilisation = true;
                 break;
             }
         }
         if (!excludeFromCapitilisation)
             text = doCapitilisation(text);
         text = doDecapitalisation(text);
 
         return text;
     }
 
    private String doCapitilisation (String text) {
       for (int i = 0; i < SNEAKY_CAPIT_PATTERNS.length; i++){
           Matcher matcher = SNEAKY_CAPIT_PATTERNS[i].matcher(text);
           if (matcher.find()){
               text = matcher.replaceAll(matcher.group(1) + matcher.group(2).toUpperCase() + matcher.group(3));
           }
       }
       text = text.substring(0,1).toUpperCase() + text.substring(1);
       text = translate(text, CAPITILISATION_RULES_REVERSED);
       return text;
    }
 
     private String doDecapitalisation(String text) {
         for (int i = 0; i < DECAPITALIZATION_PATTERNS.size(); i++){
             String s = (String) DECAPITALIZATION_PATTERNS.get(i);
             Pattern pattern = Pattern.compile("(^|\\.\\s)(" + s + ")(\\p{Upper})");
             Matcher matcher = pattern.matcher(text);
             while (matcher.find()){
                 text = matcher.replaceFirst(matcher.group(1) + matcher.group(2) + matcher.group(3).toLowerCase());
                 matcher = pattern.matcher(text);
             }
         }
         return text;
     }
 
    /**
    * This method performs the actual application of the rules.
    *
    * @param text  The text to be translated.
    * @param rules Map of rules to be applied.
    * @return the translated text.
    */
   private String translate(String text, String[][] rules) {
     for (int iii = 0; iii < rules.length; iii++) {
       String[] rule = rules[iii];
       text = text.replaceAll(rule[0], rule[1]);
     }
     return text;
   }
 }