00001 00005 #if !defined(_SA_common_h) 00006 #define _SA_common_h 00007 00008 #include "math.h" 00009 00010 typedef unsigned int IndexType; 00011 typedef unsigned int TextLenType; 00012 typedef unsigned short int SearchLenType; 00013 00014 //constants 00015 const int SIZE_ONE_READ = 16384; //when loading the data, each I/O read in SIZE_ONE_READ data points 00016 const int MAX_TOKEN_LEN = 1024; //length of the longest word 00017 00018 const int NUMBER_OF_RESERVED_WORDS_IN_VOC = 100; 00019 00021 const double SALM_PROB_UNK = 0.00000000023283064365386962890625; // 1/4G 00022 const double SALM_LOG_PROB_UNK = log(SALM_PROB_UNK); 00023 const double SALM_LOG_0 = -20; 00024 00028 typedef struct s_nGramScanningInfoElement 00029 { 00030 IndexType vocId; 00031 TextLenType freqThreshForOutput; 00032 TextLenType freqSoFar; 00033 }S_nGramScanningInfoElement; 00034 00035 #endif 00036
1.5.1