tesseract
3.03
|
00001 00002 // File: unichar.h 00003 // Description: Unicode character/ligature class. 00004 // Author: Ray Smith 00005 // Created: Wed Jun 28 17:05:01 PDT 2006 00006 // 00007 // (C) Copyright 2006, Google Inc. 00008 // Licensed under the Apache License, Version 2.0 (the "License"); 00009 // you may not use this file except in compliance with the License. 00010 // You may obtain a copy of the License at 00011 // http://www.apache.org/licenses/LICENSE-2.0 00012 // Unless required by applicable law or agreed to in writing, software 00013 // distributed under the License is distributed on an "AS IS" BASIS, 00014 // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 00015 // See the License for the specific language governing permissions and 00016 // limitations under the License. 00017 // 00019 00020 #ifndef TESSERACT_CCUTIL_UNICHAR_H__ 00021 #define TESSERACT_CCUTIL_UNICHAR_H__ 00022 00023 #include <memory.h> 00024 #include <string.h> 00025 00026 // Maximum number of characters that can be stored in a UNICHAR. Must be 00027 // at least 4. Must not exceed 31 without changing the coding of length. 00028 #define UNICHAR_LEN 30 00029 00030 // A UNICHAR_ID is the unique id of a unichar. 00031 typedef int UNICHAR_ID; 00032 00033 // A variable to indicate an invalid or uninitialized unichar id. 00034 static const int INVALID_UNICHAR_ID = -1; 00035 // A special unichar that corresponds to INVALID_UNICHAR_ID. 00036 static const char INVALID_UNICHAR[] = "__INVALID_UNICHAR__"; 00037 00038 enum StrongScriptDirection { 00039 DIR_NEUTRAL = 0, // Text contains only neutral characters. 00040 DIR_LEFT_TO_RIGHT = 1, // Text contains no Right-to-Left characters. 00041 DIR_RIGHT_TO_LEFT = 2, // Text contains no Left-to-Right characters. 00042 DIR_MIX = 3, // Text contains a mixture of left-to-right 00043 // and right-to-left characters. 00044 }; 00045 00046 // The UNICHAR class holds a single classification result. This may be 00047 // a single Unicode character (stored as between 1 and 4 utf8 bytes) or 00048 // multple Unicode characters representing the NFKC expansion of a ligature 00049 // such as fi, ffl etc. These are also stored as utf8. 00050 class UNICHAR { 00051 public: 00052 UNICHAR() { 00053 memset(chars, 0, UNICHAR_LEN); 00054 } 00055 00056 // Construct from a utf8 string. If len<0 then the string is null terminated. 00057 // If the string is too long to fit in the UNICHAR then it takes only what 00058 // will fit. 00059 UNICHAR(const char* utf8_str, int len); 00060 00061 // Construct from a single UCS4 character. 00062 explicit UNICHAR(int unicode); 00063 00064 // Default copy constructor and operator= are OK. 00065 00066 // Get the first character as UCS-4. 00067 int first_uni() const; 00068 00069 // Get the length of the UTF8 string. 00070 int utf8_len() const { 00071 int len = chars[UNICHAR_LEN - 1]; 00072 return len >=0 && len < UNICHAR_LEN ? len : UNICHAR_LEN; 00073 } 00074 00075 // Get a UTF8 string, but NOT NULL terminated. 00076 const char* utf8() const { 00077 return chars; 00078 } 00079 00080 // Get a terminated UTF8 string: Must delete[] it after use. 00081 char* utf8_str() const; 00082 00083 // Get the number of bytes in the first character of the given utf8 string. 00084 static int utf8_step(const char* utf8_str); 00085 00086 // A class to simplify iterating over and accessing elements of a UTF8 00087 // string. Note that unlike the UNICHAR class, const_iterator does NOT COPY or 00088 // take ownership of the underlying byte array. It also does not permit 00089 // modification of the array (as the name suggests). 00090 // 00091 // Example: 00092 // for (UNICHAR::const_iterator it = UNICHAR::begin(str, str_len); 00093 // it != UNICHAR::end(str, len); 00094 // ++it) { 00095 // tprintf("UCS-4 symbol code = %d\n", *it); 00096 // char buf[5]; 00097 // int char_len = it.get_utf8(buf); buf[char_len] = '\0'; 00098 // tprintf("Char = %s\n", buf); 00099 // } 00100 class const_iterator { 00101 typedef const_iterator CI; 00102 00103 public: 00104 // Step to the next UTF8 character. 00105 // If the current position is at an illegal UTF8 character, then print an 00106 // error message and step by one byte. If the current position is at a NULL 00107 // value, don't step past it. 00108 const_iterator& operator++(); 00109 00110 // Return the UCS-4 value at the current position. 00111 // If the current position is at an illegal UTF8 value, return a single 00112 // space character. 00113 int operator*() const; 00114 00115 // Store the UTF-8 encoding of the current codepoint into buf, which must be 00116 // at least 4 bytes long. Return the number of bytes written. 00117 // If the current position is at an illegal UTF8 value, writes a single 00118 // space character and returns 1. 00119 // Note that this method does not null-terminate the buffer. 00120 int get_utf8(char* buf) const; 00121 // Returns the number of bytes of the current codepoint. Returns 1 if the 00122 // current position is at an illegal UTF8 value. 00123 int utf8_len() const; 00124 00125 // Return the pointer into the string at the current position. 00126 const char* utf8_data() const { return it_; } 00127 00128 // Iterator equality operators. 00129 friend bool operator==(const CI& lhs, const CI& rhs) { 00130 return lhs.it_ == rhs.it_; 00131 } 00132 friend bool operator!=(const CI& lhs, const CI& rhs) { 00133 return !(lhs == rhs); 00134 } 00135 00136 private: 00137 friend class UNICHAR; 00138 explicit const_iterator(const char* it) : it_(it) {} 00139 00140 const char* it_; // Pointer into the string. 00141 }; 00142 00143 // Create a start/end iterator pointing to a string. Note that these methods 00144 // are static and do NOT create a copy or take ownership of the underlying 00145 // array. 00146 static const_iterator begin(const char* utf8_str, const int byte_length); 00147 static const_iterator end(const char* utf8_str, const int byte_length); 00148 00149 private: 00150 // A UTF-8 representation of 1 or more Unicode characters. 00151 // The last element (chars[UNICHAR_LEN - 1]) is a length if 00152 // its value < UNICHAR_LEN, otherwise it is a genuine character. 00153 char chars[UNICHAR_LEN]; 00154 }; 00155 00156 #endif // TESSERACT_CCUTIL_UNICHAR_H__