00001
00002
00003
00004
00005
00006
00007
00008
00009
00010
00011
00012
00013
00014
00015
00016
00017
00018
00019
00020
00021
00022
00023
00024
00025
00026
00027
00028
00029
00030
00031
00032
00033
00034
00035
00036
00037
00038
00039
00040
00041
00042
00043
00044
00045
00046
00047
00048
00049
00050
00051
00052
00053
00054
00055
00056
00057
00058
00059
00060
00061
00062
00063
00064
00065
00066
00067
00068
00069
00070
00071
00072
00073
00074
00075
00076
00077 #include "fillpen.h"
00078
00079 fillpen_t *
00080 fillpen_init(s3dict_t * dict, const char *file, float64 silprob, float64 fillprob,
00081 float64 lw, float64 wip, logmath_t *logmath)
00082 {
00083 s3wid_t w, bw;
00084 float64 prob;
00085 FILE *fp;
00086 char line[1024], wd[1024];
00087 int32 k;
00088 fillpen_t *_fillpen;
00089
00090 _fillpen = (fillpen_t *) ckd_calloc(1, sizeof(fillpen_t));
00091
00092 _fillpen->dict = dict;
00093 _fillpen->lw = lw;
00094 _fillpen->wip = wip;
00095 _fillpen->silprob = silprob;
00096 _fillpen->fillerprob = fillprob;
00097 if (dict->filler_end >= dict->filler_start)
00098 _fillpen->prob =
00099 (int32 *) ckd_calloc(dict->filler_end - dict->filler_start + 1,
00100 sizeof(int32));
00101 else
00102 _fillpen->prob = NULL;
00103
00104
00105 prob = fillprob;
00106 for (w = dict->filler_start; w <= dict->filler_end; w++)
00107 _fillpen->prob[w - dict->filler_start] =
00108 (int32) ((logmath_log(logmath, prob) * lw + logmath_log(logmath, wip)));
00109
00110
00111 w = s3dict_wordid(dict, S3_SILENCE_WORD);
00112 if (NOT_S3WID(w) || (w < dict->filler_start) || (w > dict->filler_end))
00113 E_FATAL("%s not a filler word in the given dictionary\n",
00114 S3_SILENCE_WORD);
00115 prob = silprob;
00116 _fillpen->prob[w - dict->filler_start] =
00117 (int32) ((logmath_log(logmath, prob) * lw + logmath_log(logmath, wip)));
00118
00119
00120 if (!file)
00121 return _fillpen;
00122
00123 E_INFO("Reading filler penalty file: %s\n", file);
00124 if ((fp = fopen(file, "r")) == NULL)
00125 E_FATAL("fopen(%s,r) failed\n", file);
00126 while (fgets(line, sizeof(line), fp) != NULL) {
00127 if (line[0] == '#')
00128 continue;
00129
00130 k = sscanf(line, "%s %lf", wd, &prob);
00131 if ((k != 0) && (k != 2))
00132 E_FATAL("Bad input line: %s\n", line);
00133 w = s3dict_wordid(dict, wd);
00134 if (NOT_S3WID(w) || (w < dict->filler_start)
00135 || (w > dict->filler_end))
00136 E_FATAL("%s not a filler word in the given dictionary\n",
00137 S3_SILENCE_WORD);
00138
00139 _fillpen->prob[w - dict->filler_start] =
00140 (int32) ((logmath_log(logmath, prob) * lw + logmath_log(logmath, wip)));
00141 }
00142 fclose(fp);
00143
00144
00145 for (w = dict->filler_start; w <= dict->filler_end; w++) {
00146 bw = s3dict_basewid(dict, w);
00147 if (bw != w)
00148 _fillpen->prob[w - dict->filler_start] =
00149 _fillpen->prob[bw - dict->filler_start];
00150 }
00151
00152 return _fillpen;
00153 }
00154
00155 void
00156 fillpen_report(fillpen_t * f)
00157 {
00158 E_INFO_NOFN("Initialization of fillpen_t, report:\n");
00159 E_INFO_NOFN("Language weight =%f \n", f->lw);
00160 E_INFO_NOFN("Word Insertion Penalty =%f \n", f->wip);
00161 E_INFO_NOFN("Silence probability =%f \n", f->silprob);
00162 E_INFO_NOFN("Filler probability =%f \n", f->fillerprob);
00163 E_INFO_NOFN("\n");
00164
00165 }
00166
00167 int32
00168 fillpen(fillpen_t * f, s3wid_t w)
00169 {
00170 assert((w >= f->dict->filler_start) && (w <= f->dict->filler_end));
00171 return (f->prob[w - f->dict->filler_start]);
00172 }
00173
00174
00175
00176 void
00177 fillpen_free(fillpen_t * f)
00178 {
00179 if (f) {
00180 if (f->prob)
00181 ckd_free((void *) f->prob);
00182 ckd_free((void *) f);
00183 }
00184 }