2 **********************************************************************
3 * Copyright (C) 2005-2012, International Business Machines
4 * Corporation and others. All Rights Reserved.
5 **********************************************************************
8 #include "unicode/utypes.h"
10 #if !UCONFIG_NO_CONVERSION
19 #define ARRAY_SIZE(array) (sizeof array / sizeof array[0])
22 * Matching function shared among the 2022 detectors JP, CN and KR
23 * Counts up the number of legal and unrecognized escape sequences in
24 * the sample of text, and computes a score based on the total number &
25 * the proportion that fit the encoding.
28 * @param text the byte buffer containing text to analyse
29 * @param textLen the size of the text in the byte.
30 * @param escapeSequences the byte escape sequences to test for.
31 * @return match quality, in the range of 0-100.
33 int32_t CharsetRecog_2022::match_2022(const uint8_t *text, int32_t textLen, const uint8_t escapeSequences[][5], int32_t escapeSequences_length) const
46 while(escN < escapeSequences_length) {
47 const uint8_t *seq = escapeSequences[escN];
48 int32_t seq_length = (int32_t)uprv_strlen((const char *) seq);
50 if (textLen-i >= seq_length) {
52 while(j < seq_length) {
53 if(seq[j] != text[i+j]) {
64 // else we ran out of string to compare this time.
72 if( text[i]== 0x0e || text[i] == 0x0f){
85 // Initial quality is based on relative proportion of recongized vs.
86 // unrecognized escape sequences.
87 // All good: quality = 100;
88 // half or less good: quality = 0;
90 quality = (100*hits - 100*misses) / (hits + misses);
92 // Back off quality if there were too few escape sequences seen.
93 // Include shifts in this computation, so that KR does not get penalized
94 // for having only a single Escape sequence, but many shifts.
95 if (hits+shifts < 5) {
96 quality -= (5-(hits+shifts))*10;
107 static const uint8_t escapeSequences_2022JP[][5] = {
108 {0x1b, 0x24, 0x28, 0x43, 0x00}, // KS X 1001:1992
109 {0x1b, 0x24, 0x28, 0x44, 0x00}, // JIS X 212-1990
110 {0x1b, 0x24, 0x40, 0x00, 0x00}, // JIS C 6226-1978
111 {0x1b, 0x24, 0x41, 0x00, 0x00}, // GB 2312-80
112 {0x1b, 0x24, 0x42, 0x00, 0x00}, // JIS X 208-1983
113 {0x1b, 0x26, 0x40, 0x00, 0x00}, // JIS X 208 1990, 1997
114 {0x1b, 0x28, 0x42, 0x00, 0x00}, // ASCII
115 {0x1b, 0x28, 0x48, 0x00, 0x00}, // JIS-Roman
116 {0x1b, 0x28, 0x49, 0x00, 0x00}, // Half-width katakana
117 {0x1b, 0x28, 0x4a, 0x00, 0x00}, // JIS-Roman
118 {0x1b, 0x2e, 0x41, 0x00, 0x00}, // ISO 8859-1
119 {0x1b, 0x2e, 0x46, 0x00, 0x00} // ISO 8859-7
122 #if !UCONFIG_NO_NON_HTML5_CONVERSION
123 static const uint8_t escapeSequences_2022KR[][5] = {
124 {0x1b, 0x24, 0x29, 0x43, 0x00}
127 static const uint8_t escapeSequences_2022CN[][5] = {
128 {0x1b, 0x24, 0x29, 0x41, 0x00}, // GB 2312-80
129 {0x1b, 0x24, 0x29, 0x47, 0x00}, // CNS 11643-1992 Plane 1
130 {0x1b, 0x24, 0x2A, 0x48, 0x00}, // CNS 11643-1992 Plane 2
131 {0x1b, 0x24, 0x29, 0x45, 0x00}, // ISO-IR-165
132 {0x1b, 0x24, 0x2B, 0x49, 0x00}, // CNS 11643-1992 Plane 3
133 {0x1b, 0x24, 0x2B, 0x4A, 0x00}, // CNS 11643-1992 Plane 4
134 {0x1b, 0x24, 0x2B, 0x4B, 0x00}, // CNS 11643-1992 Plane 5
135 {0x1b, 0x24, 0x2B, 0x4C, 0x00}, // CNS 11643-1992 Plane 6
136 {0x1b, 0x24, 0x2B, 0x4D, 0x00}, // CNS 11643-1992 Plane 7
137 {0x1b, 0x4e, 0x00, 0x00, 0x00}, // SS2
138 {0x1b, 0x4f, 0x00, 0x00, 0x00}, // SS3
142 CharsetRecog_2022JP::~CharsetRecog_2022JP() {}
144 const char *CharsetRecog_2022JP::getName() const {
145 return "ISO-2022-JP";
148 UBool CharsetRecog_2022JP::match(InputText *textIn, CharsetMatch *results) const {
149 int32_t confidence = match_2022(textIn->fInputBytes,
151 escapeSequences_2022JP,
152 ARRAY_SIZE(escapeSequences_2022JP));
153 results->set(textIn, this, confidence);
154 return (confidence > 0);
157 #if !UCONFIG_NO_NON_HTML5_CONVERSION
158 CharsetRecog_2022KR::~CharsetRecog_2022KR() {}
160 const char *CharsetRecog_2022KR::getName() const {
161 return "ISO-2022-KR";
164 UBool CharsetRecog_2022KR::match(InputText *textIn, CharsetMatch *results) const {
165 int32_t confidence = match_2022(textIn->fInputBytes,
167 escapeSequences_2022KR,
168 ARRAY_SIZE(escapeSequences_2022KR));
169 results->set(textIn, this, confidence);
170 return (confidence > 0);
173 CharsetRecog_2022CN::~CharsetRecog_2022CN() {}
175 const char *CharsetRecog_2022CN::getName() const {
176 return "ISO-2022-CN";
179 UBool CharsetRecog_2022CN::match(InputText *textIn, CharsetMatch *results) const {
180 int32_t confidence = match_2022(textIn->fInputBytes,
182 escapeSequences_2022CN,
183 ARRAY_SIZE(escapeSequences_2022CN));
184 results->set(textIn, this, confidence);
185 return (confidence > 0);
189 CharsetRecog_2022::~CharsetRecog_2022() {