1/** 2******************************************************************************* 3* Copyright (C) 2005-2012, International Business Machines Corporation and * 4* others. All Rights Reserved. * 5******************************************************************************* 6*/ 7package com.ibm.icu.text; 8 9import java.io.ByteArrayInputStream; 10import java.io.IOException; 11import java.io.InputStream; 12import java.io.InputStreamReader; 13import java.io.Reader; 14 15 16/** 17 * This class represents a charset that has been identified by a CharsetDetector 18 * as a possible encoding for a set of input data. From an instance of this 19 * class, you can ask for a confidence level in the charset identification, 20 * or for Java Reader or String to access the original byte data in Unicode form. 21 * <p/> 22 * Instances of this class are created only by CharsetDetectors. 23 * <p/> 24 * Note: this class has a natural ordering that is inconsistent with equals. 25 * The natural ordering is based on the match confidence value. 26 * 27 * @stable ICU 3.4 28 */ 29public class CharsetMatch implements Comparable<CharsetMatch> { 30 31 32 /** 33 * Create a java.io.Reader for reading the Unicode character data corresponding 34 * to the original byte data supplied to the Charset detect operation. 35 * <p/> 36 * CAUTION: if the source of the byte data was an InputStream, a Reader 37 * can be created for only one matching char set using this method. If more 38 * than one charset needs to be tried, the caller will need to reset 39 * the InputStream and create InputStreamReaders itself, based on the charset name. 40 * 41 * @return the Reader for the Unicode character data. 42 * 43 * @stable ICU 3.4 44 */ 45 public Reader getReader() { 46 InputStream inputStream = fInputStream; 47 48 if (inputStream == null) { 49 inputStream = new ByteArrayInputStream(fRawInput, 0, fRawLength); 50 } 51 52 try { 53 inputStream.reset(); 54 return new InputStreamReader(inputStream, getName()); 55 } catch (IOException e) { 56 return null; 57 } 58 } 59 60 /** 61 * Create a Java String from Unicode character data corresponding 62 * to the original byte data supplied to the Charset detect operation. 63 * 64 * @return a String created from the converted input data. 65 * 66 * @stable ICU 3.4 67 */ 68 public String getString() throws java.io.IOException { 69 return getString(-1); 70 71 } 72 73 /** 74 * Create a Java String from Unicode character data corresponding 75 * to the original byte data supplied to the Charset detect operation. 76 * The length of the returned string is limited to the specified size; 77 * the string will be trunctated to this length if necessary. A limit value of 78 * zero or less is ignored, and treated as no limit. 79 * 80 * @param maxLength The maximium length of the String to be created when the 81 * source of the data is an input stream, or -1 for 82 * unlimited length. 83 * @return a String created from the converted input data. 84 * 85 * @stable ICU 3.4 86 */ 87 public String getString(int maxLength) throws java.io.IOException { 88 String result = null; 89 if (fInputStream != null) { 90 StringBuilder sb = new StringBuilder(); 91 char[] buffer = new char[1024]; 92 Reader reader = getReader(); 93 int max = maxLength < 0? Integer.MAX_VALUE : maxLength; 94 int bytesRead = 0; 95 96 while ((bytesRead = reader.read(buffer, 0, Math.min(max, 1024))) >= 0) { 97 sb.append(buffer, 0, bytesRead); 98 max -= bytesRead; 99 } 100 101 reader.close(); 102 103 return sb.toString(); 104 } else { 105 String name = getName(); 106 /* 107 * getName() may return a name with a suffix 'rtl' or 'ltr'. This cannot 108 * be used to open a charset (e.g. IBM424_rtl). The ending '_rtl' or 'ltr' 109 * should be stripped off before creating the string. 110 */ 111 int startSuffix = name.indexOf("_rtl") < 0 ? name.indexOf("_ltr") : name.indexOf("_rtl"); 112 if (startSuffix > 0) { 113 name = name.substring(0, startSuffix); 114 } 115 result = new String(fRawInput, name); 116 } 117 return result; 118 119 } 120 121 /** 122 * Get an indication of the confidence in the charset detected. 123 * Confidence values range from 0-100, with larger numbers indicating 124 * a better match of the input data to the characteristics of the 125 * charset. 126 * 127 * @return the confidence in the charset match 128 * 129 * @stable ICU 3.4 130 */ 131 public int getConfidence() { 132 return fConfidence; 133 } 134 135 /** 136 * Get the name of the detected charset. 137 * The name will be one that can be used with other APIs on the 138 * platform that accept charset names. It is the "Canonical name" 139 * as defined by the class java.nio.charset.Charset; for 140 * charsets that are registered with the IANA charset registry, 141 * this is the MIME-preferred registerd name. 142 * 143 * @see java.nio.charset.Charset 144 * @see java.io.InputStreamReader 145 * 146 * @return The name of the charset. 147 * 148 * @stable ICU 3.4 149 */ 150 public String getName() { 151 return fCharsetName; 152 } 153 154 /** 155 * Get the ISO code for the language of the detected charset. 156 * 157 * @return The ISO code for the language or <code>null</code> if the language cannot be determined. 158 * 159 * @stable ICU 3.4 160 */ 161 public String getLanguage() { 162 return fLang; 163 } 164 165 /** 166 * Compare to other CharsetMatch objects. 167 * Comparison is based on the match confidence value, which 168 * allows CharsetDetector.detectAll() to order its results. 169 * 170 * @param other the CharsetMatch object to compare against. 171 * @return a negative integer, zero, or a positive integer as the 172 * confidence level of this CharsetMatch 173 * is less than, equal to, or greater than that of 174 * the argument. 175 * @throws ClassCastException if the argument is not a CharsetMatch. 176 * @stable ICU 4.4 177 */ 178 public int compareTo (CharsetMatch other) { 179 int compareResult = 0; 180 if (this.fConfidence > other.fConfidence) { 181 compareResult = 1; 182 } else if (this.fConfidence < other.fConfidence) { 183 compareResult = -1; 184 } 185 return compareResult; 186 } 187 188 /* 189 * Constructor. Implementation internal 190 */ 191 CharsetMatch(CharsetDetector det, CharsetRecognizer rec, int conf) { 192 fConfidence = conf; 193 194 // The references to the original application input data must be copied out 195 // of the charset recognizer to here, in case the application resets the 196 // recognizer before using this CharsetMatch. 197 if (det.fInputStream == null) { 198 // We only want the existing input byte data if it came straight from the user, 199 // not if is just the head of a stream. 200 fRawInput = det.fRawInput; 201 fRawLength = det.fRawLength; 202 } 203 fInputStream = det.fInputStream; 204 fCharsetName = rec.getName(); 205 fLang = rec.getLanguage(); 206 } 207 208 /* 209 * Constructor. Implementation internal 210 */ 211 CharsetMatch(CharsetDetector det, CharsetRecognizer rec, int conf, String csName, String lang) { 212 fConfidence = conf; 213 214 // The references to the original application input data must be copied out 215 // of the charset recognizer to here, in case the application resets the 216 // recognizer before using this CharsetMatch. 217 if (det.fInputStream == null) { 218 // We only want the existing input byte data if it came straight from the user, 219 // not if is just the head of a stream. 220 fRawInput = det.fRawInput; 221 fRawLength = det.fRawLength; 222 } 223 fInputStream = det.fInputStream; 224 fCharsetName = csName; 225 fLang = lang; 226 } 227 228 229 // 230 // Private Data 231 // 232 private int fConfidence; 233 private byte[] fRawInput = null; // Original, untouched input bytes. 234 // If user gave us a byte array, this is it. 235 private int fRawLength; // Length of data in fRawInput array. 236 237 private InputStream fInputStream = null; // User's input stream, or null if the user 238 // gave us a byte array. 239 240 private String fCharsetName; // The name of the charset this CharsetMatch 241 // represents. Filled in by the recognizer. 242 private String fLang; // The language, if one was determined by 243 // the recognizer during the detect operation. 244} 245