00001
00035
00036
00037
00038
00039
00040
00041
00042 #ifndef __SENT_DFA_H__
00043 #define __SENT_DFA_H__
00044
00045 #include <sent/stddefs.h>
00046
00047 #define DFA_STATESTEP 1000
00048
00049 #define INITIAL_S 0x10000000
00050 #define ACCEPT_S 0x00000001
00051
00053 typedef struct _dfa_arc {
00054 short label;
00055 int to_state;
00056 struct _dfa_arc *next;
00057 } DFA_ARC;
00058
00060 typedef struct {
00061 int number;
00062 unsigned int status;
00063 DFA_ARC *arc;
00064 } DFA_STATE;
00065
00067 typedef struct {
00068 int term_num;
00069 WORD_ID **tw;
00070 int *wnum;
00071 } TERM_INFO;
00072
00074 typedef struct {
00075 DFA_STATE *st;
00076 int maxstatenum;
00077 int state_num;
00078 int arc_num;
00079 int term_num;
00080
00089 unsigned char **cp;
00090 unsigned char *cp_begin;
00091 unsigned char *cp_end;
00092 unsigned char *cp_root;
00093 TERM_INFO term;
00094 boolean *is_sp;
00095 WORD_ID sp_id;
00096 } DFA_INFO;
00097
00098 DFA_INFO *dfa_info_new();
00099 void dfa_info_free(DFA_INFO *dfa);
00100 void dfa_state_init(DFA_INFO *dinfo);
00101 void dfa_state_expand(DFA_INFO *dinfo, int needed);
00102 boolean rddfa(FILE *fp, DFA_INFO *dinfo);
00103 boolean rddfa_fd(int fd, DFA_INFO *dinfo);
00104 boolean rddfa_sd(int sd, DFA_INFO *dinfo);
00105 boolean rddfa_line(char *line, DFA_INFO *dinfo, int *state_max, int *arc_num, int *terminal_max);
00106 void dfa_append(DFA_INFO *dst, DFA_INFO *src, int soffset, int coffset);
00107
00108 boolean init_dfa(DFA_INFO *dinfo, char *filename);
00109 WORD_ID dfa_symbol_lookup(DFA_INFO *dinfo, char *terminalname);
00110 boolean extract_cpair(DFA_INFO *dinfo);
00111 boolean cpair_append(DFA_INFO *dst, DFA_INFO *src, int coffset);
00112 void print_dfa_info(FILE *fp, DFA_INFO *dinfo);
00113 void print_dfa_cp(FILE *fp, DFA_INFO *dinfo);
00114 boolean dfa_cp(DFA_INFO *dfa, int i, int j);
00115 boolean dfa_cp_begin(DFA_INFO *dfa, int i);
00116 boolean dfa_cp_end(DFA_INFO *dfa, int i);
00117 void set_dfa_cp(DFA_INFO *dfa, int i, int j, boolean value);
00118 void set_dfa_cp_begin(DFA_INFO *dfa, int i, boolean value);
00119 void set_dfa_cp_end(DFA_INFO *dfa, int i, boolean value);
00120 void init_dfa_cp(DFA_INFO *dfa);
00121 void malloc_dfa_cp(DFA_INFO *dfa, int term_num);
00122 void realloc_dfa_cp(DFA_INFO *dfa, int old_term_num, int new_term_num);
00123 void free_dfa_cp(DFA_INFO *dfa);
00124
00125
00126
00127 #include <sent/vocabulary.h>
00128 boolean make_dfa_voca_ref(DFA_INFO *dinfo, WORD_INFO *winfo);
00129 void make_terminfo(TERM_INFO *tinfo, DFA_INFO *dinfo, WORD_INFO *winfo);
00130 void free_terminfo(TERM_INFO *tinfo);
00131 void terminfo_append(TERM_INFO *dst, TERM_INFO *src, int coffset, int woffset);
00132 #include <sent/htk_hmm.h>
00133 void dfa_find_pause_word(DFA_INFO *dfa, WORD_INFO *winfo, HTK_HMM_INFO *hmminfo);
00134 boolean dfa_pause_word_append(DFA_INFO *dst, DFA_INFO *src, int coffset);
00135 boolean read_grammar_from_socket(int sd, DFA_INFO **ret_dfa, WORD_INFO **ret_winfo, HTK_HMM_INFO *hmminfo);
00136
00137 #endif