package eu.irreality.age.language; import java.util.StringTokenizer; import eu.irreality.age.Mobile; import eu.irreality.age.NaturalLanguage; import eu.irreality.age.StringMethods; import eu.irreality.age.World; import eu.irreality.age.spell.ReferenceNameCorrector; public class Galician extends NaturalLanguage { public Galician() { super("gl"); } private static String firstWord ( String s ) { StringTokenizer st = new StringTokenizer(s); if ( st.hasMoreTokens() ) return st.nextToken(); else return ""; } private static String restWords ( String s ) { StringTokenizer st = new StringTokenizer(s); if ( !st.hasMoreTokens() ) return ""; else { st.nextToken(); if ( !st.hasMoreTokens() ) return ""; else return st.nextToken(""); } } private String uncontractClitics ( String verbWithPossibleClitics ) { //uncontract clitics for easing processing. //for example, "mo" -> "me + o" -> we represent it as "meo" for processing. String result = verbWithPossibleClitics.replaceAll("mo$","meo"); result = result.replaceAll("ma$","mea"); result = result.replaceAll("mos$","meos"); result = result.replaceAll("mas$","meas"); result = result.replaceAll("llo$","lleo"); result = result.replaceAll("lla$","llea"); result = result.replaceAll("llos$","lleos"); result = result.replaceAll("llas$","lleas"); //the extra ifs are for the special cases coller/recoller/escoller, where the verb ending could be mistaken for a clitic but isn't if ( !result.endsWith("collelo") ) result = result.replaceAll("llelo$","lleslo"); if ( !result.endsWith("collela") ) result = result.replaceAll("llela$","llesla"); if ( !result.endsWith("collelos") ) result = result.replaceAll("llelos$","lleslos"); if ( !result.endsWith("collelas") ) result = result.replaceAll("llelas$","lleslas"); //some clitics on infinitive appear transformed: //abrir + a -> abrila (not abrira) //for ease of processing, we undo this linguistic change of r to l. if ( !result.endsWith("lla") ) result = result.replaceAll("la$","ra"); if ( !result.endsWith("llo") ) result = result.replaceAll("lo$","ro"); if ( !result.endsWith("llas") ) result = result.replaceAll("las$","ras"); if ( !result.endsWith("llos") ) result = result.replaceAll("los$","ros"); return result; } //sustituye los comandos al final del verbo por las ZR's correspondientes //this method NO LONGER checks if the input is actually a verb, that should be checked outside. //however, it now (2013-05-07) checks if the OUTPUT is a verb - it won't return outputs with unrecognized verbs. public String substitutePronouns ( Mobile p , String command , Mentions mentions ) { String thestring = command; boolean doneSomething = false; //uncontract clitics for easier processing. String newFirstWord = uncontractClitics (firstWord(thestring)); thestring = newFirstWord + " " + restWords(thestring); if ( firstWord(thestring).toLowerCase().endsWith ( "as" ) && firstWord(thestring).length() > 2 ) { //Pronombre femenino plural. doneSomething = true; //lo quitamos String cut = firstWord(thestring).substring(0,firstWord(thestring).length()-2); //a�adimos la ZR femenina plural thestring = cut + " " + mentions.getLastMentionedObjectFP() + " " + restWords(thestring); doneSomething = true; } else if ( firstWord(thestring).toLowerCase().endsWith ( "os" ) && firstWord(thestring).length() > 2 ) { //Pronombre masculino o neutro plural. doneSomething = true; //lo quitamos String cut = firstWord(thestring).substring(0,firstWord(thestring).length()-2); //a�adimos la ZR plural thestring = cut + " " + mentions.getLastMentionedObjectP() + " " + restWords(thestring); } else if ( firstWord(thestring).toLowerCase().endsWith ( "o" ) && firstWord(thestring).length() > 1 ) { //Pronombre masculino singular doneSomething = true; //lo quitamos String cut = firstWord(thestring).substring(0,firstWord(thestring).length()-1); //a�adimos la ZR masculina singular thestring = cut + " " + mentions.getLastMentionedObjectMS() + " " + restWords(thestring); } else if ( firstWord(thestring).toLowerCase().endsWith ( "a" ) && firstWord(thestring).length() > 1 ) { //Pronombre femenino singular doneSomething = true; //lo quitamos String cut = firstWord(thestring).substring(0,firstWord(thestring).length()-1); //a�adimos la ZR masculina singular thestring = cut + " " + mentions.getLastMentionedObjectFS() + " " + restWords(thestring); } //intermediate checkpoint: we store the string at this point, and whether the first word in this current string has a recognized verb. //we store this because sometimes a verb may look like it has two clitics, but only actually have one. //for example, in Spanish, suppose the input "m�telo" (or "metelo"). //after we substitute "lo", what we have ("mete") is a verb. //if we keep substituting (see below) thinking that "te" is a clitic, we will reach "me", but this is no longer a verb. "te" was not a clitic, it was a verb termination! //in cases like that, we will fall back to this intermediate checkpoint. boolean intermStringIsMeaningful = false; if ( doneSomething && isVerb(removeAccents(firstWord(thestring))) ) intermStringIsMeaningful = true; //we have done some substitution, and obtained something that can be a verb. String intermString = thestring; //these pronouns can appear with the above ones (in combinations like melo, tela, selo, etc.) They can also appear on its own. if ( firstWord(thestring).toLowerCase().endsWith ( "me" ) || firstWord(thestring).toLowerCase().endsWith ( "te" ) || firstWord(thestring).toLowerCase().endsWith ( "se" ) ) { if ( !firstWord(thestring).toLowerCase().endsWith("este") && !firstWord(thestring).toLowerCase().endsWith("norte") /*&& lenguaje.esVerboComando ( thestring.substring(0,thestring.length()-2) ) */ ) { //Pronombre, se refiere al jugador doneSomething = true; //lo quitamos String cut = firstWord(thestring).substring(0,firstWord(thestring).length()-2); //a�adimos el nombre del jugador String playerRefName = p.getBestReferenceName(false); if ( playerRefName == null ) { p.writeError("Error in player " + p + ": cannot apply pronoun substitution to \"" + firstWord(thestring) + "\" because player has no reference name. Add a singular reference name to fix this.\n" ); return command; } thestring = cut + " " + playerRefName + " " + restWords(thestring); } } //if we haven't done any substitutions, we return the original string. if ( !doneSomething ) return command; if ( !isVerb(removeAccents(firstWord(thestring))) ) { if ( intermStringIsMeaningful ) { //this is the case where we have made two substitutions, so that the first one produced a correct verb, //but the last one produced something that is not a verb (like substituting "te" and "lo" in "m�telo"). //in this case, we go back to the intermediate result (substituting "lo" but not "te"), which is likely to be the correct one according to our verb lists. thestring = intermString; } else { //in this case, neither the first nor the second step of the substitution produced a known verb. //therefore, we return the original string (we could be facing a noun ending with something that looks like a critic but isn't, like //"bote" or "cola" - substitutions shouldn't have an effect in this case). return command; } } //{when we reach this point, we have made some substitution and thestring starts with a recognized verb} StringTokenizer st = new StringTokenizer(thestring.toLowerCase()); String newVerb = st.nextToken().trim(); String unaccentedVerb = this.removeAccents(newVerb); //verbos pierden acentos: c�gelo -> c�ge <tal> -> coge <tal> return unaccentedVerb + st.nextToken(""); } private String doContractions ( String s ) { String result = s.replaceAll(" a o"," ao"); result = result.replaceAll(" de o"," do"); result = result.replaceAll(" con o"," co"); result = result.replaceAll(" con a"," coa"); result = result.replaceAll(" por o"," polo"); result = result.replaceAll(" por a"," pola"); result = result.replaceAll(" en o"," no"); result = result.replaceAll(" en a"," na"); return result; } public String correctMorphology ( String s ) { if ( s == null ) return null; String niceString = super.correctMorphology(s); return doContractions(niceString); } public String correctMorphologyWithoutTrimming ( String s ) { if ( s == null ) return null; String niceString = super.correctMorphologyWithoutTrimming(s); return doContractions(niceString); } /** * Gets a Spanish corrector for the reference names of the given world. * @param w * @return */ public ReferenceNameCorrector initNameCorrector ( World w ) { ReferenceNameCorrector base = super.initNameCorrector(w); //add common synonyms of cardinal directions base.addDictionaryWord("noreste"); base.addDictionaryWord("nordeste"); base.addDictionaryWord("sureste"); base.addDictionaryWord("sudoeste"); //add articles so that, for example, the article "las" is not corrected to the noun "alas" //note: i don't think this is really needed for words of length < 3 //in fact this is all legacy, common.lan should take care of it! base.addDictionaryWord("o"); base.addDictionaryWord("a"); base.addDictionaryWord("os"); base.addDictionaryWord("as"); base.addDictionaryWord("un"); base.addDictionaryWord("unha"); base.addDictionaryWord("uns"); base.addDictionaryWord("unhas"); return base; } /** * Returns the default verb, i.e., verb that will be used by default if a reference name is typed * at the beginning of a game without specifying a verb. */ public String getDefaultVerb() { return "mirar"; } }