michael@0: /*
michael@0:  **********************************************************************
michael@0:  *   Copyright (C) 2005-2012, International Business Machines
michael@0:  *   Corporation and others.  All Rights Reserved.
michael@0:  **********************************************************************
michael@0:  */
michael@0: 
michael@0: #ifndef __CSR2022_H
michael@0: #define __CSR2022_H
michael@0: 
michael@0: #include "unicode/utypes.h"
michael@0: 
michael@0: #if !UCONFIG_NO_CONVERSION
michael@0: 
michael@0: #include "csrecog.h"
michael@0: 
michael@0: U_NAMESPACE_BEGIN
michael@0: 
michael@0: class CharsetMatch;
michael@0: 
michael@0: /**
michael@0:  *  class CharsetRecog_2022  part of the ICU charset detection imlementation.
michael@0:  *                           This is a superclass for the individual detectors for
michael@0:  *                           each of the detectable members of the ISO 2022 family
michael@0:  *                           of encodings.
michael@0:  * 
michael@0:  *                           The separate classes are nested within this class.
michael@0:  * 
michael@0:  * @internal
michael@0:  */
michael@0: class CharsetRecog_2022 : public CharsetRecognizer
michael@0: {
michael@0: 
michael@0: public:    
michael@0:     virtual ~CharsetRecog_2022() = 0;
michael@0: 
michael@0: protected:
michael@0: 
michael@0:     /**
michael@0:      * Matching function shared among the 2022 detectors JP, CN and KR
michael@0:      * Counts up the number of legal an unrecognized escape sequences in
michael@0:      * the sample of text, and computes a score based on the total number &
michael@0:      * the proportion that fit the encoding.
michael@0:      * 
michael@0:      * 
michael@0:      * @param text the byte buffer containing text to analyse
michael@0:      * @param textLen  the size of the text in the byte.
michael@0:      * @param escapeSequences the byte escape sequences to test for.
michael@0:      * @return match quality, in the range of 0-100.
michael@0:      */
michael@0:     int32_t match_2022(const uint8_t *text,
michael@0:                        int32_t textLen,
michael@0:                        const uint8_t escapeSequences[][5],
michael@0:                        int32_t escapeSequences_length) const;
michael@0: 
michael@0: };
michael@0: 
michael@0: class CharsetRecog_2022JP :public CharsetRecog_2022
michael@0: {
michael@0: public:
michael@0:     virtual ~CharsetRecog_2022JP();
michael@0: 
michael@0:     const char *getName() const;
michael@0: 
michael@0:     UBool match(InputText *textIn, CharsetMatch *results) const;
michael@0: };
michael@0: 
michael@0: class CharsetRecog_2022KR :public CharsetRecog_2022 {
michael@0: public:
michael@0:     virtual ~CharsetRecog_2022KR();
michael@0: 
michael@0:     const char *getName() const;
michael@0: 
michael@0:     UBool match(InputText *textIn, CharsetMatch *results) const;
michael@0: 
michael@0: };
michael@0: 
michael@0: class CharsetRecog_2022CN :public CharsetRecog_2022
michael@0: {
michael@0: public:
michael@0:     virtual ~CharsetRecog_2022CN();
michael@0: 
michael@0:     const char* getName() const;
michael@0: 
michael@0:     UBool match(InputText *textIn, CharsetMatch *results) const;
michael@0: };
michael@0: 
michael@0: U_NAMESPACE_END
michael@0: 
michael@0: #endif
michael@0: #endif /* __CSR2022_H */