Aegisub/vendor/universalchardet/CharDistribution.h

243 lines
7.8 KiB
C++

/* -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
/* ***** BEGIN LICENSE BLOCK *****
* Version: MPL 1.1/GPL 2.0/LGPL 2.1
*
* The contents of this file are subject to the Mozilla Public License Version
* 1.1 (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
* http://www.mozilla.org/MPL/
*
* Software distributed under the License is distributed on an "AS IS" basis,
* WITHOUT WARRANTY OF ANY KIND, either express or implied. See the License
* for the specific language governing rights and limitations under the
* License.
*
* The Original Code is Mozilla Communicator client code.
*
* The Initial Developer of the Original Code is
* Netscape Communications Corporation.
* Portions created by the Initial Developer are Copyright (C) 1998
* the Initial Developer. All Rights Reserved.
*
* Contributor(s):
*
* Alternatively, the contents of this file may be used under the terms of
* either the GNU General Public License Version 2 or later (the "GPL"), or
* the GNU Lesser General Public License Version 2.1 or later (the "LGPL"),
* in which case the provisions of the GPL or the LGPL are applicable instead
* of those above. If you wish to allow use of your version of this file only
* under the terms of either the GPL or the LGPL, and not to allow others to
* use your version of this file under the terms of the MPL, indicate your
* decision by deleting the provisions above and replace them with the notice
* and other provisions required by the GPL or the LGPL. If you do not delete
* the provisions above, a recipient may use your version of this file under
* the terms of any one of the MPL, the GPL or the LGPL.
*
* ***** END LICENSE BLOCK ***** */
#ifndef CharDistribution_h__
#define CharDistribution_h__
#include "nscore.h"
#define ENOUGH_DATA_THRESHOLD 1024
#define MINIMUM_DATA_THRESHOLD 4
class CharDistributionAnalysis
{
public:
CharDistributionAnalysis() {Reset(PR_FALSE);}
//feed a block of data and do distribution analysis
void HandleData(const char* aBuf, PRUint32 aLen) {}
//Feed a character with known length
void HandleOneChar(const char* aStr, PRUint32 aCharLen)
{
PRInt32 order;
//we only care about 2-bytes character in our distribution analysis
order = (aCharLen == 2) ? GetOrder(aStr) : -1;
if (order >= 0)
{
mTotalChars++;
//order is valid
if ((PRUint32)order < mTableSize)
{
if (512 > mCharToFreqOrder[order])
mFreqChars++;
}
}
}
//return confidence base on existing data
float GetConfidence(void);
//Reset analyser, clear any state
void Reset(PRBool aIsPreferredLanguage)
{
mDone = PR_FALSE;
mTotalChars = 0;
mFreqChars = 0;
mDataThreshold = aIsPreferredLanguage ? 0 : MINIMUM_DATA_THRESHOLD;
}
//This function is for future extension. Caller can use this function to control
//analyser's behavior
void SetOpion(){}
//It is not necessary to receive all data to draw conclusion. For charset detection,
// certain amount of data is enough
PRBool GotEnoughData() {return mTotalChars > ENOUGH_DATA_THRESHOLD;}
protected:
//we do not handle character base on its original encoding string, but
//convert this encoding string to a number, here called order.
//This allow multiple encoding of a language to share one frequency table
virtual PRInt32 GetOrder(const char* str) {return -1;}
//If this flag is set to PR_TRUE, detection is done and conclusion has been made
PRBool mDone;
//The number of characters whose frequency order is less than 512
PRUint32 mFreqChars;
//Total character encounted.
PRUint32 mTotalChars;
//Number of hi-byte characters needed to trigger detection
PRUint32 mDataThreshold;
//Mapping table to get frequency order from char order (get from GetOrder())
const PRInt16 *mCharToFreqOrder;
//Size of above table
PRUint32 mTableSize;
//This is a constant value varies from language to language, it is used in
//calculating confidence. See my paper for further detail.
float mTypicalDistributionRatio;
};
class EUCTWDistributionAnalysis: public CharDistributionAnalysis
{
public:
EUCTWDistributionAnalysis();
protected:
//for euc-TW encoding, we are interested
// first byte range: 0xc4 -- 0xfe
// second byte range: 0xa1 -- 0xfe
//no validation needed here. State machine has done that
PRInt32 GetOrder(const char* str)
{ if ((unsigned char)*str >= (unsigned char)0xc4)
return 94*((unsigned char)str[0]-(unsigned char)0xc4) + (unsigned char)str[1] - (unsigned char)0xa1;
else
return -1;
}
};
class EUCKRDistributionAnalysis : public CharDistributionAnalysis
{
public:
EUCKRDistributionAnalysis();
protected:
//for euc-KR encoding, we are interested
// first byte range: 0xb0 -- 0xfe
// second byte range: 0xa1 -- 0xfe
//no validation needed here. State machine has done that
PRInt32 GetOrder(const char* str)
{ if ((unsigned char)*str >= (unsigned char)0xb0)
return 94*((unsigned char)str[0]-(unsigned char)0xb0) + (unsigned char)str[1] - (unsigned char)0xa1;
else
return -1;
}
};
class GB2312DistributionAnalysis : public CharDistributionAnalysis
{
public:
GB2312DistributionAnalysis();
protected:
//for GB2312 encoding, we are interested
// first byte range: 0xb0 -- 0xfe
// second byte range: 0xa1 -- 0xfe
//no validation needed here. State machine has done that
PRInt32 GetOrder(const char* str)
{ if ((unsigned char)*str >= (unsigned char)0xb0 && (unsigned char)str[1] >= (unsigned char)0xa1)
return 94*((unsigned char)str[0]-(unsigned char)0xb0) + (unsigned char)str[1] - (unsigned char)0xa1;
else
return -1;
}
};
class Big5DistributionAnalysis : public CharDistributionAnalysis
{
public:
Big5DistributionAnalysis();
protected:
//for big5 encoding, we are interested
// first byte range: 0xa4 -- 0xfe
// second byte range: 0x40 -- 0x7e , 0xa1 -- 0xfe
//no validation needed here. State machine has done that
PRInt32 GetOrder(const char* str)
{ if ((unsigned char)*str >= (unsigned char)0xa4)
if ((unsigned char)str[1] >= (unsigned char)0xa1)
return 157*((unsigned char)str[0]-(unsigned char)0xa4) + (unsigned char)str[1] - (unsigned char)0xa1 +63;
else
return 157*((unsigned char)str[0]-(unsigned char)0xa4) + (unsigned char)str[1] - (unsigned char)0x40;
else
return -1;
}
};
class SJISDistributionAnalysis : public CharDistributionAnalysis
{
public:
SJISDistributionAnalysis();
protected:
//for sjis encoding, we are interested
// first byte range: 0x81 -- 0x9f , 0xe0 -- 0xfe
// second byte range: 0x40 -- 0x7e, 0x81 -- oxfe
//no validation needed here. State machine has done that
PRInt32 GetOrder(const char* str)
{
PRInt32 order;
if ((unsigned char)*str >= (unsigned char)0x81 && (unsigned char)*str <= (unsigned char)0x9f)
order = 188 * ((unsigned char)str[0]-(unsigned char)0x81);
else if ((unsigned char)*str >= (unsigned char)0xe0 && (unsigned char)*str <= (unsigned char)0xef)
order = 188 * ((unsigned char)str[0]-(unsigned char)0xe0 + 31);
else
return -1;
order += (unsigned char)*(str+1) - 0x40;
if ((unsigned char)str[1] > (unsigned char)0x7f)
order--;
return order;
}
};
class EUCJPDistributionAnalysis : public CharDistributionAnalysis
{
public:
EUCJPDistributionAnalysis();
protected:
//for euc-JP encoding, we are interested
// first byte range: 0xa0 -- 0xfe
// second byte range: 0xa1 -- 0xfe
//no validation needed here. State machine has done that
PRInt32 GetOrder(const char* str)
{ if ((unsigned char)*str >= (unsigned char)0xa0)
return 94*((unsigned char)str[0]-(unsigned char)0xa1) + (unsigned char)str[1] - (unsigned char)0xa1;
else
return -1;
}
};
#endif //CharDistribution_h__