michael@0: // Copyright 2013 Google Inc. All Rights Reserved.
michael@0: //
michael@0: // Licensed under the Apache License, Version 2.0 (the "License");
michael@0: // you may not use this file except in compliance with the License.
michael@0: // You may obtain a copy of the License at
michael@0: //
michael@0: //     http://www.apache.org/licenses/LICENSE-2.0
michael@0: //
michael@0: // Unless required by applicable law or agreed to in writing, software
michael@0: // distributed under the License is distributed on an "AS IS" BASIS,
michael@0: // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
michael@0: // See the License for the specific language governing permissions and
michael@0: // limitations under the License.
michael@0: 
michael@0: //
michael@0: // Author: dsites@google.com (Dick Sites)
michael@0: //
michael@0: 
michael@0: 
michael@0: #ifndef I18N_ENCODINGS_CLD2_INTERNAL_GETONESCRIPTSPAN_H_
michael@0: #define I18N_ENCODINGS_CLD2_INTERNAL_GETONESCRIPTSPAN_H_
michael@0: 
michael@0: #include "integral_types.h"
michael@0: #include "langspan.h"
michael@0: #include "offsetmap.h"
michael@0: 
michael@0: namespace CLD2 {
michael@0: 
michael@0: static const int kMaxScriptBuffer = 40960;
michael@0: static const int kMaxScriptLowerBuffer = (kMaxScriptBuffer * 3) / 2;
michael@0: static const int kMaxScriptBytes = kMaxScriptBuffer - 32;   // Leave some room
michael@0: static const int kWithinScriptTail = 32;    // Stop at word space in last
michael@0:                                             // N bytes of script buffer
michael@0: 
michael@0: 
michael@0: static inline bool IsContinuationByte(char c) {
michael@0:   return static_cast<signed char>(c) < -64;
michael@0: }
michael@0: 
michael@0: // Gets lscript number for letters; always returns
michael@0: //   0 (common script) for non-letters
michael@0: int GetUTF8LetterScriptNum(const char* src);
michael@0: 
michael@0: // Update src pointer to point to next quadgram, +2..+5
michael@0: // Looks at src[0..4]
michael@0: const char* AdvanceQuad(const char* src);
michael@0: 
michael@0: 
michael@0: class ScriptScanner {
michael@0:  public:
michael@0:   ScriptScanner(const char* buffer, int buffer_length, bool is_plain_text);
michael@0:   ScriptScanner(const char* buffer, int buffer_length, bool is_plain_text,
michael@0:                 bool any_text, bool any_script);
michael@0:   ~ScriptScanner();
michael@0: 
michael@0:   // Copy next run of same-script non-tag letters to buffer [NUL terminated]
michael@0:   bool GetOneScriptSpan(LangSpan* span);
michael@0: 
michael@0:   // Force Latin and Cyrillic scripts to be lowercase
michael@0:   void LowerScriptSpan(LangSpan* span);
michael@0: 
michael@0:   // Copy next run of same-script non-tag letters to buffer [NUL terminated]
michael@0:   // Force Latin and Cyrillic scripts to be lowercase
michael@0:   bool GetOneScriptSpanLower(LangSpan* span);
michael@0: 
michael@0:   // Copy next run of non-tag characters to buffer [NUL terminated]
michael@0:   // This just removes tags and removes entities
michael@0:   // Buffer has leading space
michael@0:   bool GetOneTextSpan(LangSpan* span);
michael@0: 
michael@0:   // Maps byte offset in most recent GetOneScriptSpan/Lower
michael@0:   // span->text [0..text_bytes] into an additional byte offset from
michael@0:   // span->offset, to get back to corresponding text in the original
michael@0:   // input buffer.
michael@0:   // text_offset must be the first byte
michael@0:   // of a UTF-8 character, or just beyond the last character. Normally this
michael@0:   // routine is called with the first byte of an interesting range and
michael@0:   // again with the first byte of the following range.
michael@0:   int MapBack(int text_offset);
michael@0: 
michael@0:   const char* GetBufferStart() {return start_byte_;};
michael@0: 
michael@0:  private:
michael@0:   // Skip over tags and non-letters
michael@0:   int SkipToFrontOfSpan(const char* src, int len, int* script);
michael@0: 
michael@0:   const char* start_byte_;        // Starting byte of buffer to scan
michael@0:   const char* next_byte_;         // First unscanned byte
michael@0:   const char* next_byte_limit_;   // Last byte + 1
michael@0:   int byte_length_;               // Bytes left: next_byte_limit_ - next_byte_
michael@0: 
michael@0:   bool is_plain_text_;            // true fo text, false for HTML
michael@0:   char* script_buffer_;           // Holds text with expanded entities
michael@0:   char* script_buffer_lower_;     // Holds lowercased text
michael@0:   bool letters_marks_only_;       // To distinguish scriptspan of one
michael@0:                                   // letters/marks vs. any mixture of text
michael@0:   bool one_script_only_;          // To distinguish scriptspan of one
michael@0:                                   // script vs. any mixture of scripts
michael@0:   int exit_state_;                // For tag parser kTagParseTbl_0, based
michael@0:                                   // on letters_marks_only_
michael@0:  public :
michael@0:   // Expose for debugging
michael@0:   OffsetMap map2original_;    // map from script_buffer_ to buffer
michael@0:   OffsetMap map2uplow_;       // map from script_buffer_lower_ to script_buffer_
michael@0: };
michael@0: 
michael@0: }  // namespace CLD2
michael@0: 
michael@0: #endif  // I18N_ENCODINGS_CLD2_INTERNAL_GETONESCRIPTSPAN_H_
michael@0: