| | 1 | // © 2016 and later: Unicode, Inc. and others. |
| | 2 | // License & terms of use: http://www.unicode.org/copyright.html |
| | 3 | /* |
| | 4 | ********************************************************************** |
| | 5 | * Copyright (C) 2005-2015, International Business Machines |
| | 6 | * Corporation and others. All Rights Reserved. |
| | 7 | ********************************************************************** |
| | 8 | */ |
| | 9 | |
| | 10 | #ifndef __CSR2022_H |
| | 11 | #define __CSR2022_H |
| | 12 | |
| | 13 | #include "unicode/utypes.h" |
| | 14 | |
| | 15 | #if !UCONFIG_NO_CONVERSION |
| | 16 | |
| | 17 | #include "csrecog.h" |
| | 18 | |
| | 19 | U_NAMESPACE_BEGIN |
| | 20 | |
| | 21 | class CharsetMatch; |
| | 22 | |
| | 23 | /** |
| | 24 | * class CharsetRecog_2022 part of the ICU charset detection imlementation. |
| | 25 | * This is a superclass for the individual detectors for |
| | 26 | * each of the detectable members of the ISO 2022 family |
| | 27 | * of encodings. |
| | 28 | * |
| | 29 | * The separate classes are nested within this class. |
| | 30 | * |
| | 31 | * @internal |
| | 32 | */ |
| | 33 | class CharsetRecog_2022 : public CharsetRecognizer |
| | 34 | { |
| | 35 | |
| | 36 | public: |
| | 37 | virtual ~CharsetRecog_2022() = 0; |
| | 38 | |
| | 39 | protected: |
| | 40 | |
| | 41 | /** |
| | 42 | * Matching function shared among the 2022 detectors JP, CN and KR |
| | 43 | * Counts up the number of legal an unrecognized escape sequences in |
| | 44 | * the sample of text, and computes a score based on the total number & |
| | 45 | * the proportion that fit the encoding. |
| | 46 | * |
| | 47 | * |
| | 48 | * @param text the byte buffer containing text to analyse |
| | 49 | * @param textLen the size of the text in the byte. |
| | 50 | * @param escapeSequences the byte escape sequences to test for. |
| | 51 | * @return match quality, in the range of 0-100. |
| | 52 | */ |
| | 53 | int32_t match_2022(const uint8_t *text, |
| | 54 | int32_t textLen, |
| | 55 | const uint8_t escapeSequences[][5], |
| | 56 | int32_t escapeSequences_length) const; |
| | 57 | |
| | 58 | }; |
| | 59 | |
| | 60 | class CharsetRecog_2022JP :public CharsetRecog_2022 |
| | 61 | { |
| | 62 | public: |
| | 63 | virtual ~CharsetRecog_2022JP(); |
| | 64 | |
| | 65 | const char *getName() const; |
| | 66 | |
| | 67 | UBool match(InputText *textIn, CharsetMatch *results) const; |
| | 68 | }; |
| | 69 | |
| | 70 | #if !UCONFIG_ONLY_HTML_CONVERSION |
| | 71 | class CharsetRecog_2022KR :public CharsetRecog_2022 { |
| | 72 | public: |
| | 73 | virtual ~CharsetRecog_2022KR(); |
| | 74 | |
| | 75 | const char *getName() const; |
| | 76 | |
| | 77 | UBool match(InputText *textIn, CharsetMatch *results) const; |
| | 78 | |
| | 79 | }; |
| | 80 | |
| | 81 | class CharsetRecog_2022CN :public CharsetRecog_2022 |
| | 82 | { |
| | 83 | public: |
| | 84 | virtual ~CharsetRecog_2022CN(); |
| | 85 | |
| | 86 | const char* getName() const; |
| | 87 | |
| | 88 | UBool match(InputText *textIn, CharsetMatch *results) const; |
| | 89 | }; |
| | 90 | #endif |
| | 91 | |
| | 92 | U_NAMESPACE_END |
| | 93 | |
| | 94 | #endif |
| | 95 | #endif /* __CSR2022_H */ |
projects / apple / icu.git / blame_incremental