/**
 * This file defines the default analyzer used for CPix. 
 * The default analyzer should its behaviour depending of used 
 * locale. 
 * 
 * The codes that are used in switch should follow ISO-892-1 standard
 * or ISO-839-2 standard, when 2 letter codes are not available.
 * 
 * WARNING: It is not guaranteed that in Symbian platform the language
 * codes are translated to ISO code. Also symbian platform codes of form 
 * 's29' for Taiwan Chinese or 's1' for English are supported.  
 * 
 * Refer for CPiX documentation for this file's syntax. 
 */

config_switch {

	/**
	 * The prefiltering is done e.g, when searching "$cat", "$cat.cal" etc.   
	 */ 
	case 'prefix':
		locale_switch {

			// French
			case 'fr': 				stdtokens>stdfilter>lowercase>elision(fr);

			// Default
			default:				stdtokens>stdfilter>lowercase;
		};

	/**
	 * Analyzers used for queries and indexing
	 */
	default:
			
		locale_switch {

			// French
			case 'fr': 				stdtokens>stdfilter>lowercase>elision(fr)>stop(fr);
			
			// Hebrew
			case 'he': 				
				config_switch {
					case 'query': // do not use prefix filter, when searchign 
						stdtokens>stdfilter>lowercase>stop(en);
					default: // use prefix filter only when indexing
						stdtokens>stdfilter>lowercase>prefix(he)>stop(en);
				};
		
			// English
			case 'en':		    	stdtokens>stdfilter>lowercase>stop(en);
			
			// Thai
			case 'th':				stdtokens>stdfilter>lowercase>thai>stop(en);
			
			/* 
			 * Far east asian languages
			 * 
			 *   note: Hong and Taiwanese are not differentiated. 
			 *         What are their language codes?
			 *         
			 *   note: Should we include also ISO-839-2 codes in here 
			 *   
			 *   note: Japan is no more supported. Let's used ngram for it anyway
			 */
			case 'jp', 'zh', 'ch':  
									ngram(1)>lowercase>stop(en);

			/**
			 * Korean
			 * 
			 *    note: Because special optimizations, different analyzers
			 *          are used for queries and indexing
			 *          
			 *          
			 *    WARNING: Korean analyzer is not properly tested (!)
			 */
			case 'ko':
				config_switch {
					case 'query': 	koreanquery>lowercase>stop(en);
					default:  	  	korean>lowercase>stop(en);
				};
				
			/**
			 * Default option; used for most languages and should work 'ok' 
			 * for most alphabetic writing systems. 
			 * 
			 *    note: Should we include english stop word list? 
			 */
			default: 				stdtokens>stdfilter>lowercase; 
		};
}