| 1 |
/* -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */ |
| 2 |
/* ***** BEGIN LICENSE BLOCK ***** |
| 3 |
* Version: MPL 1.1/GPL 2.0/LGPL 2.1 |
| 4 |
* |
| 5 |
* The contents of this file are subject to the Mozilla Public License Version |
| 6 |
* 1.1 (the "License"); you may not use this file except in compliance with |
| 7 |
* the License. You may obtain a copy of the License at |
| 8 |
* http://www.mozilla.org/MPL/ |
| 9 |
* |
| 10 |
* Software distributed under the License is distributed on an "AS IS" basis, |
| 11 |
* WITHOUT WARRANTY OF ANY KIND, either express or implied. See the License |
| 12 |
* for the specific language governing rights and limitations under the |
| 13 |
* License. |
| 14 |
* |
| 15 |
* The Original Code is Mozilla Universal charset detector code. |
| 16 |
* |
| 17 |
* The Initial Developer of the Original Code is |
| 18 |
* Netscape Communications Corporation. |
| 19 |
* Portions created by the Initial Developer are Copyright (C) 2001 |
| 20 |
* the Initial Developer. All Rights Reserved. |
| 21 |
* |
| 22 |
* Contributor(s): |
| 23 |
* Shy Shalom <[email protected]> |
| 24 |
* |
| 25 |
* Alternatively, the contents of this file may be used under the terms of |
| 26 |
* either the GNU General Public License Version 2 or later (the "GPL"), or |
| 27 |
* the GNU Lesser General Public License Version 2.1 or later (the "LGPL"), |
| 28 |
* in which case the provisions of the GPL or the LGPL are applicable instead |
| 29 |
* of those above. If you wish to allow use of your version of this file only |
| 30 |
* under the terms of either the GPL or the LGPL, and not to allow others to |
| 31 |
* use your version of this file under the terms of the MPL, indicate your |
| 32 |
* decision by deleting the provisions above and replace them with the notice |
| 33 |
* and other provisions required by the GPL or the LGPL. If you do not delete |
| 34 |
* the provisions above, a recipient may use your version of this file under |
| 35 |
* the terms of any one of the MPL, the GPL or the LGPL. |
| 36 |
* |
| 37 |
* ***** END LICENSE BLOCK ***** */ |
| 38 |
|
| 39 |
#include "nscore.h" |
| 40 |
|
| 41 |
#include "nsUniversalDetector.h" |
| 42 |
|
| 43 |
#include "nsMBCSGroupProber.h" |
| 44 |
#include "nsSBCSGroupProber.h" |
| 45 |
#include "nsEscCharsetProber.h" |
| 46 |
#include "nsLatin1Prober.h" |
| 47 |
|
| 48 |
nsUniversalDetector::nsUniversalDetector(PRUint32 aLanguageFilter) |
| 49 |
{ |
| 50 |
mNbspFound = PR_FALSE; |
| 51 |
mDone = PR_FALSE; |
| 52 |
mBestGuess = -1; //illegal value as signal |
| 53 |
mInTag = PR_FALSE; |
| 54 |
mEscCharSetProber = nsnull; |
| 55 |
|
| 56 |
mStart = PR_TRUE; |
| 57 |
mDetectedCharset = nsnull; |
| 58 |
mGotData = PR_FALSE; |
| 59 |
mInputState = ePureAscii; |
| 60 |
mLastChar = '\0'; |
| 61 |
mLanguageFilter = aLanguageFilter; |
| 62 |
|
| 63 |
PRUint32 i; |
| 64 |
for (i = 0; i < NUM_OF_CHARSET_PROBERS; i++) |
| 65 |
mCharSetProbers[i] = nsnull; |
| 66 |
} |
| 67 |
|
| 68 |
nsUniversalDetector::~nsUniversalDetector() |
| 69 |
{ |
| 70 |
for (PRInt32 i = 0; i < NUM_OF_CHARSET_PROBERS; i++) |
| 71 |
delete mCharSetProbers[i]; |
| 72 |
|
| 73 |
delete mEscCharSetProber; |
| 74 |
} |
| 75 |
|
| 76 |
void |
| 77 |
nsUniversalDetector::Reset() |
| 78 |
{ |
| 79 |
mNbspFound = PR_FALSE; |
| 80 |
mDone = PR_FALSE; |
| 81 |
mBestGuess = -1; //illegal value as signal |
| 82 |
mInTag = PR_FALSE; |
| 83 |
|
| 84 |
mStart = PR_TRUE; |
| 85 |
mDetectedCharset = nsnull; |
| 86 |
mGotData = PR_FALSE; |
| 87 |
mInputState = ePureAscii; |
| 88 |
mLastChar = '\0'; |
| 89 |
|
| 90 |
if (mEscCharSetProber) |
| 91 |
mEscCharSetProber->Reset(); |
| 92 |
|
| 93 |
PRUint32 i; |
| 94 |
for (i = 0; i < NUM_OF_CHARSET_PROBERS; i++) |
| 95 |
if (mCharSetProbers[i]) |
| 96 |
mCharSetProbers[i]->Reset(); |
| 97 |
} |
| 98 |
|
| 99 |
//--------------------------------------------------------------------- |
| 100 |
#define SHORTCUT_THRESHOLD (float)0.95 |
| 101 |
#define MINIMUM_THRESHOLD (float)0.20 |
| 102 |
|
| 103 |
nsresult nsUniversalDetector::HandleData(const char* aBuf, PRUint32 aLen) |
| 104 |
{ |
| 105 |
if(mDone) |
| 106 |
return NS_OK; |
| 107 |
|
| 108 |
if (aLen > 0) |
| 109 |
mGotData = PR_TRUE; |
| 110 |
|
| 111 |
/* If the data starts with BOM, we know it is UTF. */ |
| 112 |
if (mStart) |
| 113 |
{ |
| 114 |
mStart = PR_FALSE; |
| 115 |
if (aLen > 2) |
| 116 |
{ |
| 117 |
switch (aBuf[0]) |
| 118 |
{ |
| 119 |
case '\xEF': |
| 120 |
if (('\xBB' == aBuf[1]) && ('\xBF' == aBuf[2])) |
| 121 |
/* EF BB BF: UTF-8 encoded BOM. */ |
| 122 |
mDetectedCharset = "UTF-8"; |
| 123 |
break; |
| 124 |
case '\xFE': |
| 125 |
if ('\xFF' == aBuf[1]) |
| 126 |
/* FE FF: UTF-16, big endian BOM. */ |
| 127 |
mDetectedCharset = "UTF-16"; |
| 128 |
break; |
| 129 |
case '\xFF': |
| 130 |
if ('\xFE' == aBuf[1]) |
| 131 |
{ |
| 132 |
if (aLen > 3 && |
| 133 |
aBuf[2] == '\x00' && |
| 134 |
aBuf[3] == '\x00') |
| 135 |
{ |
| 136 |
/* FF FE 00 00: UTF-32 (LE). */ |
| 137 |
mDetectedCharset = "UTF-32"; |
| 138 |
} |
| 139 |
else |
| 140 |
{ |
| 141 |
/* FF FE: UTF-16, little endian BOM. */ |
| 142 |
mDetectedCharset = "UTF-16"; |
| 143 |
} |
| 144 |
} |
| 145 |
break; |
| 146 |
case '\x00': |
| 147 |
if (aLen > 3 && |
| 148 |
aBuf[1] == '\x00' && |
| 149 |
aBuf[2] == '\xFE' && |
| 150 |
aBuf[3] == '\xFF') |
| 151 |
{ |
| 152 |
/* 00 00 FE FF: UTF-32 (BE). */ |
| 153 |
mDetectedCharset = "UTF-32"; |
| 154 |
} |
| 155 |
break; |
| 156 |
} |
| 157 |
} |
| 158 |
|
| 159 |
if (mDetectedCharset) |
| 160 |
{ |
| 161 |
mDone = PR_TRUE; |
| 162 |
return NS_OK; |
| 163 |
} |
| 164 |
} |
| 165 |
|
| 166 |
PRUint32 i; |
| 167 |
for (i = 0; i < aLen; i++) |
| 168 |
{ |
| 169 |
/* If every other character is ASCII or 0xA0, we don't run charset |
| 170 |
* probers. |
| 171 |
* 0xA0 (NBSP in a few charset) is apparently a rare exception |
| 172 |
* of non-ASCII character often contained in nearly-ASCII text. */ |
| 173 |
if (aBuf[i] & '\x80' && aBuf[i] != '\xA0') |
| 174 |
{ |
| 175 |
/* We got a non-ASCII byte (high-byte) */ |
| 176 |
if (mInputState != eHighbyte) |
| 177 |
{ |
| 178 |
//adjust state |
| 179 |
mInputState = eHighbyte; |
| 180 |
|
| 181 |
//kill mEscCharSetProber if it is active |
| 182 |
if (mEscCharSetProber) { |
| 183 |
delete mEscCharSetProber; |
| 184 |
mEscCharSetProber = nsnull; |
| 185 |
} |
| 186 |
|
| 187 |
//start multibyte and singlebyte charset prober |
| 188 |
if (nsnull == mCharSetProbers[0]) |
| 189 |
{ |
| 190 |
mCharSetProbers[0] = new nsMBCSGroupProber(mLanguageFilter); |
| 191 |
if (nsnull == mCharSetProbers[0]) |
| 192 |
return NS_ERROR_OUT_OF_MEMORY; |
| 193 |
} |
| 194 |
if (nsnull == mCharSetProbers[1] && |
| 195 |
(mLanguageFilter & NS_FILTER_NON_CJK)) |
| 196 |
{ |
| 197 |
mCharSetProbers[1] = new nsSBCSGroupProber; |
| 198 |
if (nsnull == mCharSetProbers[1]) |
| 199 |
return NS_ERROR_OUT_OF_MEMORY; |
| 200 |
} |
| 201 |
if (nsnull == mCharSetProbers[2]) |
| 202 |
{ |
| 203 |
mCharSetProbers[2] = new nsLatin1Prober; |
| 204 |
if (nsnull == mCharSetProbers[2]) |
| 205 |
return NS_ERROR_OUT_OF_MEMORY; |
| 206 |
} |
| 207 |
} |
| 208 |
} |
| 209 |
else |
| 210 |
{ |
| 211 |
/* Just pure ASCII or NBSP so far. */ |
| 212 |
if (aBuf[i] == '\xA0') |
| 213 |
{ |
| 214 |
/* ASCII with the only exception of NBSP seems quite common. |
| 215 |
* I doubt it is really necessary to train a model here, so let's |
| 216 |
* just make an exception. |
| 217 |
*/ |
| 218 |
mNbspFound = PR_TRUE; |
| 219 |
} |
| 220 |
else if (mInputState == ePureAscii && |
| 221 |
(aBuf[i] == '\033' || (aBuf[i] == '{' && mLastChar == '~'))) |
| 222 |
{ |
| 223 |
/* We found an escape character or HZ "~{". */ |
| 224 |
mInputState = eEscAscii; |
| 225 |
} |
| 226 |
mLastChar = aBuf[i]; |
| 227 |
} |
| 228 |
} |
| 229 |
|
| 230 |
nsProbingState st; |
| 231 |
switch (mInputState) |
| 232 |
{ |
| 233 |
case eEscAscii: |
| 234 |
if (nsnull == mEscCharSetProber) { |
| 235 |
mEscCharSetProber = new nsEscCharSetProber(mLanguageFilter); |
| 236 |
if (nsnull == mEscCharSetProber) |
| 237 |
return NS_ERROR_OUT_OF_MEMORY; |
| 238 |
} |
| 239 |
st = mEscCharSetProber->HandleData(aBuf, aLen); |
| 240 |
if (st == eFoundIt) |
| 241 |
{ |
| 242 |
mDone = PR_TRUE; |
| 243 |
mDetectedCharset = mEscCharSetProber->GetCharSetName(); |
| 244 |
} |
| 245 |
break; |
| 246 |
case eHighbyte: |
| 247 |
for (i = 0; i < NUM_OF_CHARSET_PROBERS; i++) |
| 248 |
{ |
| 249 |
if (mCharSetProbers[i]) |
| 250 |
{ |
| 251 |
st = mCharSetProbers[i]->HandleData(aBuf, aLen); |
| 252 |
if (st == eFoundIt) |
| 253 |
{ |
| 254 |
mDone = PR_TRUE; |
| 255 |
mDetectedCharset = mCharSetProbers[i]->GetCharSetName(); |
| 256 |
return NS_OK; |
| 257 |
} |
| 258 |
} |
| 259 |
} |
| 260 |
break; |
| 261 |
|
| 262 |
default: |
| 263 |
break; |
| 264 |
} |
| 265 |
return NS_OK; |
| 266 |
} |
| 267 |
|
| 268 |
|
| 269 |
//--------------------------------------------------------------------- |
| 270 |
void nsUniversalDetector::DataEnd() |
| 271 |
{ |
| 272 |
if (!mGotData) |
| 273 |
{ |
| 274 |
// we haven't got any data yet, return immediately |
| 275 |
// caller program sometimes call DataEnd before anything has been sent to detector |
| 276 |
return; |
| 277 |
} |
| 278 |
|
| 279 |
if (! mDetectedCharset) |
| 280 |
{ |
| 281 |
switch (mInputState) |
| 282 |
{ |
| 283 |
case eEscAscii: |
| 284 |
case ePureAscii: |
| 285 |
if (mNbspFound) |
| 286 |
{ |
| 287 |
/* ISO-8859-1 is a good result candidate for ASCII + NBSP. |
| 288 |
* (though it could have been any ISO-8859 encoding). */ |
| 289 |
mDetectedCharset = "ISO-8859-1"; |
| 290 |
} |
| 291 |
else |
| 292 |
{ |
| 293 |
/* ASCII with the ESC character (or the sequence "~{") is still |
| 294 |
* ASCII until proven otherwise. */ |
| 295 |
mDetectedCharset = "ASCII"; |
| 296 |
} |
| 297 |
default: |
| 298 |
break; |
| 299 |
} |
| 300 |
} |
| 301 |
|
| 302 |
if (mDetectedCharset) |
| 303 |
{ |
| 304 |
mDone = PR_TRUE; |
| 305 |
Report(mDetectedCharset); |
| 306 |
return; |
| 307 |
} |
| 308 |
|
| 309 |
switch (mInputState) |
| 310 |
{ |
| 311 |
case eHighbyte: |
| 312 |
{ |
| 313 |
float proberConfidence; |
| 314 |
float maxProberConfidence = (float)0.0; |
| 315 |
PRInt32 maxProber = 0; |
| 316 |
|
| 317 |
for (PRInt32 i = 0; i < NUM_OF_CHARSET_PROBERS; i++) |
| 318 |
{ |
| 319 |
if (mCharSetProbers[i]) |
| 320 |
{ |
| 321 |
proberConfidence = mCharSetProbers[i]->GetConfidence(); |
| 322 |
if (proberConfidence > maxProberConfidence) |
| 323 |
{ |
| 324 |
maxProberConfidence = proberConfidence; |
| 325 |
maxProber = i; |
| 326 |
} |
| 327 |
} |
| 328 |
} |
| 329 |
//do not report anything because we are not confident of it, that's in fact a negative answer |
| 330 |
if (maxProberConfidence > MINIMUM_THRESHOLD) |
| 331 |
Report(mCharSetProbers[maxProber]->GetCharSetName()); |
| 332 |
} |
| 333 |
break; |
| 334 |
case eEscAscii: |
| 335 |
break; |
| 336 |
default: |
| 337 |
; |
| 338 |
} |
| 339 |
return; |
| 340 |
} |