wapiti 0.0.5 → 0.1.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +7 -0
- data/.simplecov +3 -0
- data/Gemfile +25 -2
- data/HISTORY.md +5 -1
- data/LICENSE +14 -13
- data/README.md +9 -16
- data/Rakefile +38 -8
- data/ext/wapiti/bcd.c +126 -124
- data/ext/wapiti/decoder.c +203 -124
- data/ext/wapiti/decoder.h +6 -4
- data/ext/wapiti/extconf.rb +2 -2
- data/ext/wapiti/gradient.c +491 -320
- data/ext/wapiti/gradient.h +52 -34
- data/ext/wapiti/lbfgs.c +74 -33
- data/ext/wapiti/model.c +47 -37
- data/ext/wapiti/model.h +22 -20
- data/ext/wapiti/native.c +850 -839
- data/ext/wapiti/native.h +1 -1
- data/ext/wapiti/options.c +52 -20
- data/ext/wapiti/options.h +37 -30
- data/ext/wapiti/pattern.c +35 -33
- data/ext/wapiti/pattern.h +12 -11
- data/ext/wapiti/progress.c +14 -13
- data/ext/wapiti/progress.h +3 -2
- data/ext/wapiti/quark.c +14 -16
- data/ext/wapiti/quark.h +6 -5
- data/ext/wapiti/reader.c +83 -69
- data/ext/wapiti/reader.h +11 -9
- data/ext/wapiti/rprop.c +84 -43
- data/ext/wapiti/sequence.h +18 -16
- data/ext/wapiti/sgdl1.c +45 -43
- data/ext/wapiti/thread.c +19 -17
- data/ext/wapiti/thread.h +5 -4
- data/ext/wapiti/tools.c +7 -7
- data/ext/wapiti/tools.h +3 -4
- data/ext/wapiti/trainers.h +1 -1
- data/ext/wapiti/vmath.c +40 -38
- data/ext/wapiti/vmath.h +12 -11
- data/ext/wapiti/wapiti.c +159 -37
- data/ext/wapiti/wapiti.h +18 -4
- data/lib/wapiti.rb +15 -15
- data/lib/wapiti/errors.rb +15 -15
- data/lib/wapiti/model.rb +92 -84
- data/lib/wapiti/options.rb +123 -124
- data/lib/wapiti/utility.rb +14 -14
- data/lib/wapiti/version.rb +2 -2
- data/spec/spec_helper.rb +29 -9
- data/spec/wapiti/model_spec.rb +230 -194
- data/spec/wapiti/native_spec.rb +7 -8
- data/spec/wapiti/options_spec.rb +184 -174
- data/wapiti.gemspec +22 -8
- metadata +38 -42
- data/.gitignore +0 -5
data/ext/wapiti/model.h
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
/*
|
2
2
|
* Wapiti - A linear-chain CRF tool
|
3
3
|
*
|
4
|
-
* Copyright (c) 2009-
|
4
|
+
* Copyright (c) 2009-2013 CNRS
|
5
5
|
* All rights reserved.
|
6
6
|
*
|
7
7
|
* Redistribution and use in source and binary forms, with or without
|
@@ -29,14 +29,15 @@
|
|
29
29
|
#define model_h
|
30
30
|
|
31
31
|
#include <stddef.h>
|
32
|
-
#include <
|
32
|
+
#include <stdint.h>
|
33
|
+
#include <sys/time.h>
|
33
34
|
|
34
|
-
#include "wapiti.h"
|
35
35
|
#include "options.h"
|
36
36
|
#include "sequence.h"
|
37
37
|
#include "reader.h"
|
38
|
+
#include "wapiti.h"
|
38
39
|
|
39
|
-
typedef struct
|
40
|
+
typedef struct timeval tms_t;
|
40
41
|
|
41
42
|
/* mdl_t:
|
42
43
|
* Represent a linear-chain CRF model. The model contain both unigram and
|
@@ -60,34 +61,35 @@ typedef struct tms tms_t;
|
|
60
61
|
*/
|
61
62
|
typedef struct mdl_s mdl_t;
|
62
63
|
struct mdl_s {
|
63
|
-
opt_t
|
64
|
+
opt_t *opt; // options for training
|
65
|
+
int type; // model type
|
64
66
|
|
65
67
|
// Size of various model parameters
|
66
|
-
|
67
|
-
|
68
|
-
|
68
|
+
uint32_t nlbl; // Y number of labels
|
69
|
+
uint64_t nobs; // O number of observations
|
70
|
+
uint64_t nftr; // F number of features
|
69
71
|
|
70
72
|
// Informations about observations
|
71
|
-
char
|
72
|
-
|
73
|
-
|
73
|
+
char *kind; // [O] observations type
|
74
|
+
uint64_t *uoff; // [O] unigram weights offset
|
75
|
+
uint64_t *boff; // [O] bigram weights offset
|
74
76
|
|
75
77
|
// The model itself
|
76
|
-
double
|
78
|
+
double *theta; // [F] features weights
|
77
79
|
|
78
80
|
// Datasets
|
79
|
-
dat_t
|
80
|
-
dat_t
|
81
|
-
rdr_t
|
81
|
+
dat_t *train; // training dataset
|
82
|
+
dat_t *devel; // development dataset
|
83
|
+
rdr_t *reader;
|
82
84
|
|
83
85
|
// Stoping criterion
|
84
|
-
double
|
85
|
-
|
86
|
-
|
86
|
+
double *werr; // Window of error rate of last iters
|
87
|
+
uint32_t wcnt; // Number of iters in the window
|
88
|
+
uint32_t wpos; // Position for the next iter
|
87
89
|
|
88
90
|
// Timing
|
89
|
-
tms_t
|
90
|
-
double
|
91
|
+
tms_t timer; // start time of last iter
|
92
|
+
double total; // total training time
|
91
93
|
};
|
92
94
|
|
93
95
|
mdl_t *mdl_new(rdr_t *rdr);
|
data/ext/wapiti/native.c
CHANGED
@@ -1,13 +1,15 @@
|
|
1
1
|
#include <stdio.h>
|
2
2
|
#include <string.h>
|
3
3
|
|
4
|
-
#include "wapiti.h"
|
5
4
|
#include "options.h"
|
6
5
|
#include "reader.h"
|
6
|
+
#include "decoder.h"
|
7
7
|
#include "model.h"
|
8
8
|
#include "trainers.h"
|
9
|
+
#include "progress.h"
|
9
10
|
#include "quark.h"
|
10
11
|
#include "tools.h"
|
12
|
+
#include "wapiti.h"
|
11
13
|
|
12
14
|
#include "native.h"
|
13
15
|
|
@@ -22,27 +24,36 @@ VALUE cConfigurationError;
|
|
22
24
|
VALUE cLogger;
|
23
25
|
|
24
26
|
|
27
|
+
/* --- Forward declarations --- */
|
28
|
+
|
29
|
+
int wapiti_main(int argc, char *argv[argc]);
|
30
|
+
|
31
|
+
void dolabel(mdl_t *mdl);
|
32
|
+
void dotrain(mdl_t *mdl);
|
33
|
+
void doupdt(mdl_t *mdl);
|
34
|
+
|
35
|
+
|
25
36
|
/* --- Utilities --- */
|
26
37
|
|
27
38
|
static void trn_auto(mdl_t *mdl) {
|
28
|
-
|
29
|
-
|
30
|
-
|
31
|
-
|
32
|
-
|
39
|
+
const int maxiter = mdl->opt->maxiter;
|
40
|
+
mdl->opt->maxiter = 3;
|
41
|
+
trn_sgdl1(mdl);
|
42
|
+
mdl->opt->maxiter = maxiter;
|
43
|
+
trn_lbfgs(mdl);
|
33
44
|
}
|
34
45
|
|
35
46
|
static const struct {
|
36
|
-
|
37
|
-
|
47
|
+
const char *name;
|
48
|
+
void (* train)(mdl_t *mdl);
|
38
49
|
} trn_lst[] = {
|
39
|
-
|
40
|
-
|
41
|
-
|
42
|
-
|
43
|
-
|
44
|
-
|
45
|
-
|
50
|
+
{"l-bfgs", trn_lbfgs},
|
51
|
+
{"sgd-l1", trn_sgdl1},
|
52
|
+
{"bcd", trn_bcd },
|
53
|
+
{"rprop", trn_rprop},
|
54
|
+
{"rprop+", trn_rprop},
|
55
|
+
{"rprop-", trn_rprop},
|
56
|
+
{"auto", trn_auto }
|
46
57
|
};
|
47
58
|
static const int trn_cnt = sizeof(trn_lst) / sizeof(trn_lst[0]);
|
48
59
|
|
@@ -52,78 +63,78 @@ static const int trn_cnt = sizeof(trn_lst) / sizeof(trn_lst[0]);
|
|
52
63
|
// Auxiliary Methods
|
53
64
|
|
54
65
|
static opt_t *get_options(VALUE self) {
|
55
|
-
|
56
|
-
|
57
|
-
|
66
|
+
opt_t *options;
|
67
|
+
Data_Get_Struct(self, opt_t, options);
|
68
|
+
return options;
|
58
69
|
}
|
59
70
|
|
60
71
|
// Copies a Ruby string to the heap and stores it in a pointer.
|
61
72
|
// Frees the pointer before assigning the new value.
|
62
73
|
static void copy_string(char **dst, VALUE rb_string) {
|
63
|
-
|
74
|
+
Check_Type(rb_string, T_STRING);
|
64
75
|
|
65
|
-
|
66
|
-
|
76
|
+
if (*dst) { free(*dst); *dst = (char*)0; }
|
77
|
+
*dst = calloc(RSTRING_LEN(rb_string) + 1, sizeof(char));
|
67
78
|
|
68
|
-
|
79
|
+
memcpy(*dst, StringValuePtr(rb_string), RSTRING_LEN(rb_string) + 1);
|
69
80
|
}
|
70
81
|
|
71
82
|
|
72
83
|
// Constructor / Desctructor
|
73
84
|
|
74
85
|
static void mark_options(opt_t* options __attribute__((__unused__))) {
|
75
|
-
|
86
|
+
// nothing
|
76
87
|
}
|
77
88
|
|
78
89
|
static void deallocate_options(opt_t* options) {
|
79
|
-
|
80
|
-
|
81
|
-
|
82
|
-
|
83
|
-
|
84
|
-
|
85
|
-
|
86
|
-
|
87
|
-
|
88
|
-
|
90
|
+
|
91
|
+
// free string options
|
92
|
+
if (options->input) { free(options->input); }
|
93
|
+
if (options->output) { free(options->output); }
|
94
|
+
if (options->algo) { free((void*)options->algo); }
|
95
|
+
if (options->devel) { free(options->devel); }
|
96
|
+
if (options->pattern) { free((void*)options->pattern); }
|
97
|
+
|
98
|
+
free(options);
|
99
|
+
options = (opt_t*)0;
|
89
100
|
}
|
90
101
|
|
91
102
|
static VALUE allocate_options(VALUE self) {
|
92
|
-
|
93
|
-
|
94
|
-
}
|
95
|
-
|
96
|
-
static VALUE initialize_options(int argc, VALUE *argv, VALUE self) {
|
97
|
-
|
98
|
-
|
99
|
-
|
100
|
-
|
101
|
-
|
102
|
-
|
103
|
-
|
104
|
-
|
105
|
-
|
106
|
-
|
107
|
-
|
108
|
-
|
109
|
-
|
110
|
-
|
111
|
-
|
112
|
-
|
113
|
-
|
114
|
-
|
115
|
-
|
116
|
-
|
117
|
-
|
118
|
-
|
119
|
-
|
120
|
-
|
121
|
-
|
122
|
-
|
123
|
-
|
124
|
-
|
125
|
-
|
126
|
-
|
103
|
+
opt_t* options = malloc(sizeof(opt_t));
|
104
|
+
return Data_Wrap_Struct(self, mark_options, deallocate_options, options);
|
105
|
+
}
|
106
|
+
|
107
|
+
static VALUE initialize_options(int argc, VALUE *argv, VALUE self) {
|
108
|
+
opt_t* options = get_options(self);
|
109
|
+
*options = opt_defaults;
|
110
|
+
|
111
|
+
if (options->maxiter == 0) {
|
112
|
+
options->maxiter = INT_MAX;
|
113
|
+
}
|
114
|
+
|
115
|
+
// copy the default algorithm name to the heap so that all options strings
|
116
|
+
// are on the heap
|
117
|
+
char* tmp = calloc(strlen(options->algo), sizeof(char));
|
118
|
+
memcpy(tmp, options->algo, strlen(options->algo));
|
119
|
+
options->algo = tmp;
|
120
|
+
|
121
|
+
if (argc > 1) {
|
122
|
+
rb_raise(rb_const_get(rb_mKernel, rb_intern("ArgumentError")),
|
123
|
+
"wrong number of arguments (%d for 0..1)", argc);
|
124
|
+
}
|
125
|
+
|
126
|
+
// set defaults
|
127
|
+
if (argc) {
|
128
|
+
Check_Type(argv[0], T_HASH);
|
129
|
+
(void)rb_funcall(self, rb_intern("update"), 1, argv[0]);
|
130
|
+
}
|
131
|
+
|
132
|
+
// yield self if block_given?
|
133
|
+
if (rb_block_given_p()) {
|
134
|
+
rb_yield(self);
|
135
|
+
}
|
136
|
+
|
137
|
+
return self;
|
127
138
|
}
|
128
139
|
|
129
140
|
|
@@ -133,192 +144,192 @@ static VALUE initialize_options(int argc, VALUE *argv, VALUE self) {
|
|
133
144
|
// Fixnum Accessors
|
134
145
|
|
135
146
|
static VALUE options_nbest(VALUE self) {
|
136
|
-
|
147
|
+
return INT2FIX(get_options(self)->nbest);
|
137
148
|
}
|
138
149
|
|
139
150
|
static VALUE options_set_nbest(VALUE self, VALUE rb_fixnum) {
|
140
|
-
|
141
|
-
|
142
|
-
|
143
|
-
|
151
|
+
Check_Type(rb_fixnum, T_FIXNUM);
|
152
|
+
get_options(self)->nbest = FIX2INT(rb_fixnum);
|
153
|
+
|
154
|
+
return rb_fixnum;
|
144
155
|
}
|
145
156
|
|
146
157
|
|
147
158
|
static VALUE options_stopwin(VALUE self) {
|
148
|
-
|
159
|
+
return INT2FIX(get_options(self)->stopwin);
|
149
160
|
}
|
150
161
|
|
151
162
|
static VALUE options_set_stopwin(VALUE self, VALUE rb_fixnum) {
|
152
|
-
|
153
|
-
|
154
|
-
|
155
|
-
|
163
|
+
Check_Type(rb_fixnum, T_FIXNUM);
|
164
|
+
get_options(self)->stopwin = FIX2INT(rb_fixnum);
|
165
|
+
|
166
|
+
return rb_fixnum;
|
156
167
|
}
|
157
168
|
|
158
169
|
static VALUE options_objwin(VALUE self) {
|
159
|
-
|
170
|
+
return INT2FIX(get_options(self)->objwin);
|
160
171
|
}
|
161
172
|
|
162
173
|
static VALUE options_set_objwin(VALUE self, VALUE rb_fixnum) {
|
163
|
-
|
164
|
-
|
165
|
-
|
166
|
-
|
174
|
+
Check_Type(rb_fixnum, T_FIXNUM);
|
175
|
+
get_options(self)->objwin = FIX2INT(rb_fixnum);
|
176
|
+
|
177
|
+
return rb_fixnum;
|
167
178
|
}
|
168
179
|
|
169
180
|
|
170
181
|
static VALUE options_maxiter(VALUE self) {
|
171
|
-
|
182
|
+
return INT2FIX(get_options(self)->maxiter);
|
172
183
|
}
|
173
184
|
|
174
185
|
static VALUE options_set_maxiter(VALUE self, VALUE rb_fixnum) {
|
175
|
-
|
186
|
+
opt_t *options = get_options(self);
|
176
187
|
|
177
|
-
|
178
|
-
|
179
|
-
|
180
|
-
|
188
|
+
Check_Type(rb_fixnum, T_FIXNUM);
|
189
|
+
options->maxiter = FIX2INT(rb_fixnum);
|
190
|
+
|
191
|
+
return rb_fixnum;
|
181
192
|
}
|
182
193
|
|
183
194
|
static VALUE options_jobsize(VALUE self) {
|
184
|
-
|
195
|
+
return INT2FIX(get_options(self)->jobsize);
|
185
196
|
}
|
186
197
|
|
187
198
|
static VALUE options_set_jobsize(VALUE self, VALUE rb_fixnum) {
|
188
|
-
|
199
|
+
opt_t *options = get_options(self);
|
200
|
+
|
201
|
+
Check_Type(rb_fixnum, T_FIXNUM);
|
202
|
+
options->jobsize = FIX2INT(rb_fixnum);
|
189
203
|
|
190
|
-
|
191
|
-
options->jobsize = FIX2INT(rb_fixnum);
|
192
|
-
|
193
|
-
return rb_fixnum;
|
204
|
+
return rb_fixnum;
|
194
205
|
}
|
195
206
|
|
196
207
|
static VALUE options_nthread(VALUE self) {
|
197
|
-
|
208
|
+
return INT2FIX(get_options(self)->nthread);
|
198
209
|
}
|
199
210
|
|
200
211
|
static VALUE options_set_nthread(VALUE self, VALUE rb_fixnum) {
|
201
|
-
|
212
|
+
opt_t *options = get_options(self);
|
202
213
|
|
203
|
-
|
204
|
-
|
205
|
-
|
206
|
-
|
214
|
+
Check_Type(rb_fixnum, T_FIXNUM);
|
215
|
+
options->nthread = FIX2INT(rb_fixnum);
|
216
|
+
|
217
|
+
return rb_fixnum;
|
207
218
|
}
|
208
219
|
|
209
220
|
static VALUE options_histsz(VALUE self) {
|
210
|
-
|
221
|
+
return INT2FIX(get_options(self)->lbfgs.histsz);
|
211
222
|
}
|
212
223
|
|
213
224
|
static VALUE options_set_histsz(VALUE self, VALUE rb_fixnum) {
|
214
|
-
|
215
|
-
|
216
|
-
|
217
|
-
|
225
|
+
Check_Type(rb_fixnum, T_FIXNUM);
|
226
|
+
get_options(self)->lbfgs.histsz = FIX2INT(rb_fixnum);
|
227
|
+
|
228
|
+
return rb_fixnum;
|
218
229
|
}
|
219
230
|
|
220
231
|
static VALUE options_maxls(VALUE self) {
|
221
|
-
|
232
|
+
return INT2FIX(get_options(self)->lbfgs.maxls);
|
222
233
|
}
|
223
234
|
|
224
235
|
static VALUE options_set_maxls(VALUE self, VALUE rb_fixnum) {
|
225
|
-
|
226
|
-
|
227
|
-
|
228
|
-
|
236
|
+
Check_Type(rb_fixnum, T_FIXNUM);
|
237
|
+
get_options(self)->lbfgs.maxls = FIX2INT(rb_fixnum);
|
238
|
+
|
239
|
+
return rb_fixnum;
|
229
240
|
}
|
230
241
|
|
231
242
|
|
232
243
|
// Float Accessors
|
233
244
|
|
234
245
|
static VALUE options_rho1(VALUE self) {
|
235
|
-
|
246
|
+
return rb_float_new(get_options(self)->rho1);
|
236
247
|
}
|
237
248
|
|
238
249
|
static VALUE options_set_rho1(VALUE self, VALUE rb_numeric) {
|
239
|
-
|
240
|
-
|
250
|
+
get_options(self)->rho1 = NUM2DBL(rb_numeric);
|
251
|
+
return rb_numeric;
|
241
252
|
}
|
242
253
|
|
243
254
|
static VALUE options_rho2(VALUE self) {
|
244
|
-
|
255
|
+
return rb_float_new(get_options(self)->rho2);
|
245
256
|
}
|
246
257
|
|
247
258
|
static VALUE options_set_rho2(VALUE self, VALUE rb_numeric) {
|
248
|
-
|
249
|
-
|
259
|
+
get_options(self)->rho2 = NUM2DBL(rb_numeric);
|
260
|
+
return rb_numeric;
|
250
261
|
}
|
251
262
|
|
252
263
|
static VALUE options_stopeps(VALUE self) {
|
253
|
-
|
264
|
+
return rb_float_new(get_options(self)->stopeps);
|
254
265
|
}
|
255
266
|
|
256
267
|
static VALUE options_set_stopeps(VALUE self, VALUE rb_numeric) {
|
257
|
-
|
258
|
-
|
268
|
+
get_options(self)->stopeps = NUM2DBL(rb_numeric);
|
269
|
+
return rb_numeric;
|
259
270
|
}
|
260
271
|
|
261
272
|
static VALUE options_eta0(VALUE self) {
|
262
|
-
|
273
|
+
return rb_float_new(get_options(self)->sgdl1.eta0);
|
263
274
|
}
|
264
275
|
|
265
276
|
static VALUE options_set_eta0(VALUE self, VALUE rb_numeric) {
|
266
|
-
|
267
|
-
|
277
|
+
get_options(self)->sgdl1.eta0 = NUM2DBL(rb_numeric);
|
278
|
+
return rb_numeric;
|
268
279
|
}
|
269
280
|
|
270
281
|
static VALUE options_alpha(VALUE self) {
|
271
|
-
|
282
|
+
return rb_float_new(get_options(self)->sgdl1.alpha);
|
272
283
|
}
|
273
284
|
|
274
285
|
static VALUE options_set_alpha(VALUE self, VALUE rb_numeric) {
|
275
|
-
|
276
|
-
|
286
|
+
get_options(self)->sgdl1.alpha = NUM2DBL(rb_numeric);
|
287
|
+
return rb_numeric;
|
277
288
|
}
|
278
289
|
|
279
290
|
static VALUE options_kappa(VALUE self) {
|
280
|
-
|
291
|
+
return rb_float_new(get_options(self)->bcd.kappa);
|
281
292
|
}
|
282
293
|
|
283
294
|
static VALUE options_set_kappa(VALUE self, VALUE rb_numeric) {
|
284
|
-
|
285
|
-
|
295
|
+
get_options(self)->bcd.kappa = NUM2DBL(rb_numeric);
|
296
|
+
return rb_numeric;
|
286
297
|
}
|
287
298
|
|
288
299
|
static VALUE options_stpmin(VALUE self) {
|
289
|
-
|
300
|
+
return rb_float_new(get_options(self)->rprop.stpmin);
|
290
301
|
}
|
291
302
|
|
292
303
|
static VALUE options_set_stpmin(VALUE self, VALUE rb_numeric) {
|
293
|
-
|
294
|
-
|
304
|
+
get_options(self)->rprop.stpmin = NUM2DBL(rb_numeric);
|
305
|
+
return rb_numeric;
|
295
306
|
}
|
296
307
|
|
297
308
|
static VALUE options_stpmax(VALUE self) {
|
298
|
-
|
309
|
+
return rb_float_new(get_options(self)->rprop.stpmax);
|
299
310
|
}
|
300
311
|
|
301
312
|
static VALUE options_set_stpmax(VALUE self, VALUE rb_numeric) {
|
302
|
-
|
303
|
-
|
313
|
+
get_options(self)->rprop.stpmax = NUM2DBL(rb_numeric);
|
314
|
+
return rb_numeric;
|
304
315
|
}
|
305
316
|
|
306
317
|
static VALUE options_stpinc(VALUE self) {
|
307
|
-
|
318
|
+
return rb_float_new(get_options(self)->rprop.stpinc);
|
308
319
|
}
|
309
320
|
|
310
321
|
static VALUE options_set_stpinc(VALUE self, VALUE rb_numeric) {
|
311
|
-
|
312
|
-
|
322
|
+
get_options(self)->rprop.stpinc = NUM2DBL(rb_numeric);
|
323
|
+
return rb_numeric;
|
313
324
|
}
|
314
325
|
|
315
326
|
static VALUE options_stpdec(VALUE self) {
|
316
|
-
|
327
|
+
return rb_float_new(get_options(self)->rprop.stpdec);
|
317
328
|
}
|
318
329
|
|
319
330
|
static VALUE options_set_stpdec(VALUE self, VALUE rb_numeric) {
|
320
|
-
|
321
|
-
|
331
|
+
get_options(self)->rprop.stpdec = NUM2DBL(rb_numeric);
|
332
|
+
return rb_numeric;
|
322
333
|
}
|
323
334
|
|
324
335
|
|
@@ -326,84 +337,84 @@ static VALUE options_set_stpdec(VALUE self, VALUE rb_numeric) {
|
|
326
337
|
// Boolean Accessors
|
327
338
|
|
328
339
|
static VALUE options_maxent(VALUE self) {
|
329
|
-
|
340
|
+
return get_options(self)->maxent ? Qtrue : Qfalse;
|
330
341
|
}
|
331
342
|
|
332
343
|
static VALUE options_set_maxent(VALUE self, VALUE rb_boolean) {
|
333
|
-
|
334
|
-
|
344
|
+
get_options(self)->maxent = !(TYPE(rb_boolean) == T_NIL || !rb_boolean);
|
345
|
+
return rb_boolean;
|
335
346
|
}
|
336
347
|
|
337
348
|
static VALUE options_compact(VALUE self) {
|
338
|
-
|
349
|
+
return get_options(self)->compact ? Qtrue : Qfalse;
|
339
350
|
}
|
340
351
|
|
341
352
|
static VALUE options_set_compact(VALUE self, VALUE rb_boolean) {
|
342
|
-
|
343
|
-
|
353
|
+
get_options(self)->compact = !(TYPE(rb_boolean) == T_NIL || !rb_boolean);
|
354
|
+
return rb_boolean;
|
344
355
|
}
|
345
356
|
|
346
357
|
static VALUE options_sparse(VALUE self) {
|
347
|
-
|
358
|
+
return get_options(self)->sparse ? Qtrue : Qfalse;
|
348
359
|
}
|
349
360
|
|
350
361
|
static VALUE options_set_sparse(VALUE self, VALUE rb_boolean) {
|
351
|
-
|
352
|
-
|
362
|
+
get_options(self)->sparse = !(TYPE(rb_boolean) == T_NIL || !rb_boolean);
|
363
|
+
return rb_boolean;
|
353
364
|
}
|
354
365
|
|
355
366
|
static VALUE options_check(VALUE self) {
|
356
|
-
|
367
|
+
return get_options(self)->check ? Qtrue : Qfalse;
|
357
368
|
}
|
358
369
|
|
359
370
|
static VALUE options_set_check(VALUE self, VALUE rb_boolean) {
|
360
|
-
|
361
|
-
|
371
|
+
get_options(self)->check = !(TYPE(rb_boolean) == T_NIL || !rb_boolean);
|
372
|
+
return rb_boolean;
|
362
373
|
}
|
363
374
|
|
364
375
|
static VALUE options_label(VALUE self) {
|
365
|
-
|
376
|
+
return get_options(self)->label ? Qtrue : Qfalse;
|
366
377
|
}
|
367
378
|
|
368
379
|
static VALUE options_set_label(VALUE self, VALUE rb_boolean) {
|
369
|
-
|
370
|
-
|
380
|
+
get_options(self)->label = !(TYPE(rb_boolean) == T_NIL || !rb_boolean);
|
381
|
+
return rb_boolean;
|
371
382
|
}
|
372
383
|
|
373
384
|
static VALUE options_outsc(VALUE self) {
|
374
|
-
|
385
|
+
return get_options(self)->outsc ? Qtrue : Qfalse;
|
375
386
|
}
|
376
387
|
|
377
388
|
static VALUE options_set_outsc(VALUE self, VALUE rb_boolean) {
|
378
|
-
|
379
|
-
|
389
|
+
get_options(self)->outsc = !(TYPE(rb_boolean) == T_NIL || !rb_boolean);
|
390
|
+
return rb_boolean;
|
380
391
|
}
|
381
392
|
|
382
393
|
static VALUE options_lblpost(VALUE self) {
|
383
|
-
|
394
|
+
return get_options(self)->lblpost ? Qtrue : Qfalse;
|
384
395
|
}
|
385
396
|
|
386
397
|
static VALUE options_set_lblpost(VALUE self, VALUE rb_boolean) {
|
387
|
-
|
388
|
-
|
398
|
+
get_options(self)->lblpost = !(TYPE(rb_boolean) == T_NIL || !rb_boolean);
|
399
|
+
return rb_boolean;
|
389
400
|
}
|
390
401
|
|
391
402
|
static VALUE options_clip(VALUE self) {
|
392
|
-
|
403
|
+
return get_options(self)->lbfgs.clip ? Qtrue : Qfalse;
|
393
404
|
}
|
394
405
|
|
395
406
|
static VALUE options_set_clip(VALUE self, VALUE rb_boolean) {
|
396
|
-
|
397
|
-
|
407
|
+
get_options(self)->lbfgs.clip = !(TYPE(rb_boolean) == T_NIL || !rb_boolean);
|
408
|
+
return rb_boolean;
|
398
409
|
}
|
399
410
|
|
400
411
|
static VALUE options_cutoff(VALUE self) {
|
401
|
-
|
412
|
+
return get_options(self)->rprop.cutoff ? Qtrue : Qfalse;
|
402
413
|
}
|
403
414
|
|
404
415
|
static VALUE options_set_cutoff(VALUE self, VALUE rb_boolean) {
|
405
|
-
|
406
|
-
|
416
|
+
get_options(self)->rprop.cutoff = !(TYPE(rb_boolean) == T_NIL || !rb_boolean);
|
417
|
+
return rb_boolean;
|
407
418
|
}
|
408
419
|
|
409
420
|
|
@@ -412,201 +423,201 @@ static VALUE options_set_cutoff(VALUE self, VALUE rb_boolean) {
|
|
412
423
|
// String Accessors
|
413
424
|
|
414
425
|
static VALUE options_pattern(VALUE self) {
|
415
|
-
|
416
|
-
|
426
|
+
const char *pattern = get_options(self)->pattern;
|
427
|
+
return rb_str_new2(pattern ? pattern : "");
|
417
428
|
}
|
418
429
|
|
419
430
|
static VALUE options_set_pattern(VALUE self, VALUE rb_string) {
|
420
|
-
|
421
|
-
|
422
|
-
|
423
|
-
|
431
|
+
opt_t *options = get_options(self);
|
432
|
+
copy_string((char**)&(options->pattern), rb_string);
|
433
|
+
|
434
|
+
return rb_string;
|
424
435
|
}
|
425
436
|
|
426
437
|
static VALUE options_model(VALUE self) {
|
427
|
-
|
428
|
-
|
438
|
+
const char *model = get_options(self)->model;
|
439
|
+
return rb_str_new2(model ? model : "");
|
429
440
|
}
|
430
441
|
|
431
442
|
static VALUE options_set_model(VALUE self, VALUE rb_string) {
|
432
|
-
|
433
|
-
|
434
|
-
|
435
|
-
|
443
|
+
opt_t *options = get_options(self);
|
444
|
+
copy_string(&(options->model), rb_string);
|
445
|
+
|
446
|
+
return rb_string;
|
436
447
|
}
|
437
448
|
|
438
449
|
static VALUE options_algorithm(VALUE self) {
|
439
|
-
|
440
|
-
|
450
|
+
const char *algorithm = get_options(self)->algo;
|
451
|
+
return rb_str_new2(algorithm ? algorithm : "");
|
441
452
|
}
|
442
453
|
|
443
454
|
static VALUE options_set_algorithm(VALUE self, VALUE rb_string) {
|
444
|
-
|
445
|
-
|
446
|
-
|
447
|
-
|
455
|
+
opt_t *options = get_options(self);
|
456
|
+
copy_string((char**)&(options->algo), rb_string);
|
457
|
+
|
458
|
+
return rb_string;
|
448
459
|
}
|
449
460
|
|
450
461
|
static VALUE options_development_data(VALUE self) {
|
451
|
-
|
452
|
-
|
462
|
+
char *development_data = get_options(self)->devel;
|
463
|
+
return rb_str_new2(development_data ? development_data : "");
|
453
464
|
}
|
454
465
|
|
455
466
|
static VALUE options_set_development_data(VALUE self, VALUE rb_string) {
|
456
|
-
|
457
|
-
|
458
|
-
|
459
|
-
|
467
|
+
opt_t *options = get_options(self);
|
468
|
+
copy_string(&(options->devel), rb_string);
|
469
|
+
|
470
|
+
return rb_string;
|
460
471
|
}
|
461
472
|
|
462
473
|
|
463
474
|
void Init_options() {
|
464
|
-
|
465
|
-
|
466
|
-
|
467
|
-
|
475
|
+
cOptions = rb_define_class_under(mWapiti, "Options", rb_cObject);
|
476
|
+
rb_define_alloc_func(cOptions, allocate_options);
|
477
|
+
|
478
|
+
rb_define_method(cOptions, "initialize", initialize_options, -1);
|
479
|
+
|
480
|
+
// Option Accessors
|
481
|
+
|
482
|
+
rb_define_method(cOptions, "stopwin", options_stopwin, 0);
|
483
|
+
rb_define_method(cOptions, "stopwin=", options_set_stopwin, 1);
|
484
|
+
|
485
|
+
rb_define_alias(cOptions, "stop_window", "stopwin");
|
486
|
+
rb_define_alias(cOptions, "stop_window=", "stopwin=");
|
468
487
|
|
469
|
-
|
488
|
+
rb_define_method(cOptions, "objwin", options_objwin, 0);
|
489
|
+
rb_define_method(cOptions, "objwin=", options_set_objwin, 1);
|
470
490
|
|
471
|
-
|
472
|
-
|
491
|
+
rb_define_alias(cOptions, "convergence_window", "objwin");
|
492
|
+
rb_define_alias(cOptions, "convergence_window=", "objwin=");
|
473
493
|
|
474
|
-
|
475
|
-
|
494
|
+
rb_define_method(cOptions, "maxiter", options_maxiter, 0);
|
495
|
+
rb_define_method(cOptions, "maxiter=", options_set_maxiter, 1);
|
476
496
|
|
477
|
-
|
478
|
-
|
497
|
+
rb_define_alias(cOptions, "max_iterations", "maxiter");
|
498
|
+
rb_define_alias(cOptions, "max_iterations=", "maxiter=");
|
479
499
|
|
480
|
-
|
481
|
-
|
482
|
-
|
483
|
-
rb_define_method(cOptions, "maxiter", options_maxiter, 0);
|
484
|
-
rb_define_method(cOptions, "maxiter=", options_set_maxiter, 1);
|
500
|
+
rb_define_method(cOptions, "jobsize", options_jobsize, 0);
|
501
|
+
rb_define_method(cOptions, "jobsize=", options_set_jobsize, 1);
|
485
502
|
|
486
|
-
|
487
|
-
|
488
|
-
|
489
|
-
rb_define_method(cOptions, "jobsize", options_jobsize, 0);
|
490
|
-
rb_define_method(cOptions, "jobsize=", options_set_jobsize, 1);
|
503
|
+
rb_define_method(cOptions, "nthread", options_nthread, 0);
|
504
|
+
rb_define_method(cOptions, "nthread=", options_set_nthread, 1);
|
491
505
|
|
492
|
-
|
493
|
-
|
506
|
+
rb_define_alias(cOptions, "threads", "nthread");
|
507
|
+
rb_define_alias(cOptions, "threads=", "nthread=");
|
494
508
|
|
495
|
-
|
496
|
-
|
509
|
+
rb_define_method(cOptions, "rho1", options_rho1, 0);
|
510
|
+
rb_define_method(cOptions, "rho1=", options_set_rho1, 1);
|
497
511
|
|
498
|
-
|
499
|
-
|
512
|
+
rb_define_method(cOptions, "rho2", options_rho2, 0);
|
513
|
+
rb_define_method(cOptions, "rho2=", options_set_rho2, 1);
|
500
514
|
|
501
|
-
|
502
|
-
|
515
|
+
rb_define_method(cOptions, "stopeps", options_stopeps, 0);
|
516
|
+
rb_define_method(cOptions, "stopeps=", options_set_stopeps, 1);
|
503
517
|
|
504
|
-
|
505
|
-
|
518
|
+
rb_define_alias(cOptions, "stop_epsilon", "stopeps");
|
519
|
+
rb_define_alias(cOptions, "stop_epsilon=", "stopeps=");
|
506
520
|
|
507
|
-
|
508
|
-
|
521
|
+
rb_define_method(cOptions, "maxent", options_maxent, 0);
|
522
|
+
rb_define_method(cOptions, "maxent=", options_set_maxent, 1);
|
509
523
|
|
510
|
-
|
511
|
-
rb_define_method(cOptions, "maxent=", options_set_maxent, 1);
|
524
|
+
rb_define_alias(cOptions, "maxent?", "maxent");
|
512
525
|
|
513
|
-
|
526
|
+
rb_define_method(cOptions, "compact", options_compact, 0);
|
527
|
+
rb_define_method(cOptions, "compact=", options_set_compact, 1);
|
514
528
|
|
515
|
-
|
516
|
-
rb_define_method(cOptions, "compact=", options_set_compact, 1);
|
529
|
+
rb_define_alias(cOptions, "compact?", "compact");
|
517
530
|
|
518
|
-
|
531
|
+
rb_define_method(cOptions, "sparse", options_sparse, 0);
|
532
|
+
rb_define_method(cOptions, "sparse=", options_set_sparse, 1);
|
519
533
|
|
520
|
-
|
521
|
-
rb_define_method(cOptions, "sparse=", options_set_sparse, 1);
|
534
|
+
rb_define_alias(cOptions, "sparse?", "sparse");
|
522
535
|
|
523
|
-
|
536
|
+
rb_define_method(cOptions, "skip_tokens", options_label, 0);
|
537
|
+
rb_define_method(cOptions, "skip_tokens=", options_set_label, 1);
|
524
538
|
|
525
|
-
|
526
|
-
rb_define_method(cOptions, "skip_tokens=", options_set_label, 1);
|
539
|
+
rb_define_alias(cOptions, "skip_tokens?", "skip_tokens");
|
527
540
|
|
528
|
-
|
541
|
+
rb_define_method(cOptions, "check", options_check, 0);
|
542
|
+
rb_define_method(cOptions, "check=", options_set_check, 1);
|
529
543
|
|
530
|
-
|
531
|
-
rb_define_method(cOptions, "check=", options_set_check, 1);
|
544
|
+
rb_define_alias(cOptions, "check?", "check");
|
532
545
|
|
533
|
-
|
546
|
+
rb_define_method(cOptions, "lblpost", options_lblpost, 0);
|
547
|
+
rb_define_method(cOptions, "lblpost=", options_set_lblpost, 1);
|
534
548
|
|
535
|
-
|
536
|
-
rb_define_method(cOptions, "lblpost=", options_set_lblpost, 1);
|
549
|
+
rb_define_alias(cOptions, "lblpost?", "lblpost");
|
537
550
|
|
538
|
-
|
551
|
+
rb_define_alias(cOptions, "posterior", "lblpost");
|
552
|
+
rb_define_alias(cOptions, "posterior?", "lblpost");
|
553
|
+
rb_define_alias(cOptions, "posterior=", "lblpost=");
|
539
554
|
|
540
|
-
|
541
|
-
|
542
|
-
rb_define_alias(cOptions, "posterior=", "lblpost=");
|
555
|
+
rb_define_method(cOptions, "outsc", options_outsc, 0);
|
556
|
+
rb_define_method(cOptions, "outsc=", options_set_outsc, 1);
|
543
557
|
|
544
|
-
|
545
|
-
rb_define_method(cOptions, "outsc=", options_set_outsc, 1);
|
558
|
+
rb_define_alias(cOptions, "outsc?", "outsc");
|
546
559
|
|
547
|
-
|
560
|
+
rb_define_alias(cOptions, "score", "outsc");
|
561
|
+
rb_define_alias(cOptions, "score?", "outsc");
|
562
|
+
rb_define_alias(cOptions, "score=", "outsc=");
|
548
563
|
|
549
|
-
|
550
|
-
|
551
|
-
rb_define_alias(cOptions, "score=", "outsc=");
|
564
|
+
rb_define_method(cOptions, "pattern", options_pattern, 0);
|
565
|
+
rb_define_method(cOptions, "pattern=", options_set_pattern, 1);
|
552
566
|
|
553
|
-
|
554
|
-
|
567
|
+
rb_define_alias(cOptions, "template", "pattern");
|
568
|
+
rb_define_alias(cOptions, "template=", "pattern=");
|
555
569
|
|
556
|
-
|
557
|
-
|
570
|
+
rb_define_method(cOptions, "model", options_model, 0);
|
571
|
+
rb_define_method(cOptions, "model=", options_set_model, 1);
|
558
572
|
|
559
|
-
|
560
|
-
|
573
|
+
rb_define_method(cOptions, "algorithm", options_algorithm, 0);
|
574
|
+
rb_define_method(cOptions, "algorithm=", options_set_algorithm, 1);
|
561
575
|
|
562
|
-
|
563
|
-
|
576
|
+
rb_define_alias(cOptions, "algo", "algorithm");
|
577
|
+
rb_define_alias(cOptions, "algo=", "algorithm=");
|
564
578
|
|
565
|
-
|
566
|
-
|
579
|
+
rb_define_method(cOptions, "development_data", options_development_data, 0);
|
580
|
+
rb_define_method(cOptions, "development_data=", options_set_development_data, 1);
|
567
581
|
|
568
|
-
|
569
|
-
|
582
|
+
rb_define_alias(cOptions, "devel", "development_data");
|
583
|
+
rb_define_alias(cOptions, "devel=", "development_data=");
|
570
584
|
|
571
|
-
|
572
|
-
|
585
|
+
rb_define_method(cOptions, "clip", options_clip, 0);
|
586
|
+
rb_define_method(cOptions, "clip=", options_set_clip, 1);
|
573
587
|
|
574
|
-
|
575
|
-
|
588
|
+
rb_define_method(cOptions, "histsz", options_histsz, 0);
|
589
|
+
rb_define_method(cOptions, "histsz=", options_set_histsz, 1);
|
576
590
|
|
577
|
-
|
578
|
-
|
591
|
+
rb_define_method(cOptions, "maxls", options_maxls, 0);
|
592
|
+
rb_define_method(cOptions, "maxls=", options_set_maxls, 1);
|
579
593
|
|
580
|
-
|
581
|
-
|
594
|
+
rb_define_method(cOptions, "eta0", options_eta0, 0);
|
595
|
+
rb_define_method(cOptions, "eta0=", options_set_eta0, 1);
|
582
596
|
|
583
|
-
|
584
|
-
|
597
|
+
rb_define_method(cOptions, "alpha", options_alpha, 0);
|
598
|
+
rb_define_method(cOptions, "alpha=", options_set_alpha, 1);
|
585
599
|
|
586
|
-
|
587
|
-
|
600
|
+
rb_define_method(cOptions, "kappa", options_kappa, 0);
|
601
|
+
rb_define_method(cOptions, "kappa=", options_set_kappa, 1);
|
588
602
|
|
589
|
-
|
590
|
-
|
603
|
+
rb_define_method(cOptions, "stpmin", options_stpmin, 0);
|
604
|
+
rb_define_method(cOptions, "stpmin=", options_set_stpmin, 1);
|
591
605
|
|
592
|
-
|
593
|
-
|
606
|
+
rb_define_method(cOptions, "stpmax", options_stpmax, 0);
|
607
|
+
rb_define_method(cOptions, "stpmax=", options_set_stpmax, 1);
|
594
608
|
|
595
|
-
|
596
|
-
|
609
|
+
rb_define_method(cOptions, "stpinc", options_stpinc, 0);
|
610
|
+
rb_define_method(cOptions, "stpinc=", options_set_stpinc, 1);
|
597
611
|
|
598
|
-
|
599
|
-
|
612
|
+
rb_define_method(cOptions, "stpdec", options_stpdec, 0);
|
613
|
+
rb_define_method(cOptions, "stpdec=", options_set_stpdec, 1);
|
600
614
|
|
601
|
-
|
602
|
-
|
615
|
+
rb_define_method(cOptions, "cutoff", options_cutoff, 0);
|
616
|
+
rb_define_method(cOptions, "cutoff=", options_set_cutoff, 1);
|
617
|
+
|
618
|
+
rb_define_method(cOptions, "nbest", options_nbest, 0);
|
619
|
+
rb_define_method(cOptions, "nbest=", options_set_nbest, 1);
|
603
620
|
|
604
|
-
rb_define_method(cOptions, "cutoff", options_cutoff, 0);
|
605
|
-
rb_define_method(cOptions, "cutoff=", options_set_cutoff, 1);
|
606
|
-
|
607
|
-
rb_define_method(cOptions, "nbest", options_nbest, 0);
|
608
|
-
rb_define_method(cOptions, "nbest=", options_set_nbest, 1);
|
609
|
-
|
610
621
|
}
|
611
622
|
|
612
623
|
|
@@ -615,119 +626,119 @@ void Init_options() {
|
|
615
626
|
// Auxiliary Methods
|
616
627
|
|
617
628
|
static mdl_t *get_model(VALUE self) {
|
618
|
-
|
619
|
-
|
620
|
-
|
629
|
+
mdl_t *model;
|
630
|
+
Data_Get_Struct(self, mdl_t, model);
|
631
|
+
return model;
|
621
632
|
}
|
622
633
|
|
623
634
|
// Constructor / Desctructor
|
624
635
|
|
625
636
|
static void mark_model(mdl_t *model __attribute__((__unused__))) {
|
626
|
-
|
637
|
+
// nothing
|
627
638
|
}
|
628
639
|
|
629
640
|
static void deallocate_model(mdl_t *model) {
|
630
|
-
|
631
|
-
|
632
|
-
|
633
|
-
|
641
|
+
if (model) {
|
642
|
+
mdl_free(model);
|
643
|
+
model = (mdl_t*)0;
|
644
|
+
}
|
634
645
|
}
|
635
646
|
|
636
647
|
static VALUE allocate_model(VALUE self) {
|
637
|
-
|
638
|
-
|
648
|
+
mdl_t *model = mdl_new(rdr_new(false));
|
649
|
+
return Data_Wrap_Struct(self, mark_model, deallocate_model, model);
|
639
650
|
}
|
640
651
|
|
641
|
-
static VALUE model_set_options(VALUE self, VALUE rb_options) {
|
642
|
-
|
643
|
-
|
644
|
-
|
645
|
-
|
646
|
-
|
647
|
-
|
648
|
-
|
649
|
-
|
650
|
-
|
651
|
-
|
652
|
-
|
653
|
-
|
654
|
-
|
655
|
-
|
652
|
+
static VALUE model_set_options(VALUE self, VALUE rb_options) {
|
653
|
+
if (strncmp("Wapiti::Options", rb_obj_classname(rb_options), 15) != 0) {
|
654
|
+
rb_raise(cNativeError, "argument must be a Wapiti::Options instance");
|
655
|
+
}
|
656
|
+
|
657
|
+
mdl_t *model = get_model(self);
|
658
|
+
|
659
|
+
// Store reference to options in model struct
|
660
|
+
model->opt = get_options(rb_options);
|
661
|
+
|
662
|
+
// Update reader
|
663
|
+
model->reader->autouni = model->opt->maxent;
|
664
|
+
|
665
|
+
// Save instance variable
|
666
|
+
rb_ivar_set(self, rb_intern("@options"), rb_options);
|
656
667
|
|
657
|
-
|
668
|
+
return rb_options;
|
658
669
|
}
|
659
670
|
|
660
671
|
static VALUE initialize_model(int argc, VALUE *argv, VALUE self) {
|
661
|
-
|
662
|
-
|
663
|
-
|
664
|
-
|
665
|
-
|
666
|
-
|
667
|
-
|
668
|
-
|
669
|
-
|
670
|
-
|
671
|
-
|
672
|
-
|
673
|
-
|
674
|
-
|
675
|
-
|
676
|
-
|
677
|
-
|
678
|
-
|
679
|
-
|
680
|
-
|
681
|
-
|
682
|
-
|
683
|
-
|
684
|
-
|
685
|
-
|
686
|
-
|
687
|
-
|
688
|
-
|
689
|
-
|
690
|
-
|
691
|
-
|
692
|
-
|
693
|
-
|
694
|
-
|
695
|
-
|
696
|
-
|
697
|
-
|
698
|
-
|
672
|
+
VALUE options;
|
673
|
+
|
674
|
+
if (argc > 1) {
|
675
|
+
rb_raise(rb_const_get(rb_mKernel, rb_intern("ArgumentError")),
|
676
|
+
"wrong number of arguments (%d for 0..1)", argc);
|
677
|
+
}
|
678
|
+
|
679
|
+
if (argc) {
|
680
|
+
if (TYPE(argv[0]) == T_HASH) {
|
681
|
+
options = rb_funcall(cOptions, rb_intern("new"), 1, argv[0]);
|
682
|
+
}
|
683
|
+
else {
|
684
|
+
if (strncmp("Wapiti::Options", rb_obj_classname(argv[0]), 15) != 0) {
|
685
|
+
rb_raise(cNativeError, "argument must be a hash or an options instance");
|
686
|
+
}
|
687
|
+
options = argv[0];
|
688
|
+
}
|
689
|
+
}
|
690
|
+
else {
|
691
|
+
options = rb_funcall(cOptions, rb_intern("new"), 0);
|
692
|
+
}
|
693
|
+
|
694
|
+
// yield options if block_given?
|
695
|
+
if (rb_block_given_p()) {
|
696
|
+
rb_yield(options);
|
697
|
+
}
|
698
|
+
|
699
|
+
model_set_options(self, options);
|
700
|
+
|
701
|
+
// Load a previous model if specified by options
|
702
|
+
if (get_options(options)->model) {
|
703
|
+
rb_funcall(self, rb_intern("load"), 0);
|
704
|
+
}
|
705
|
+
|
706
|
+
// initialize counters
|
707
|
+
rb_funcall(self, rb_intern("clear_counters"), 0);
|
708
|
+
|
709
|
+
return self;
|
699
710
|
}
|
700
711
|
|
701
712
|
|
702
713
|
// Native accessors
|
703
714
|
|
704
715
|
static VALUE model_nlbl(VALUE self) {
|
705
|
-
|
716
|
+
return INT2FIX(get_model(self)->nlbl);
|
706
717
|
}
|
707
718
|
|
708
719
|
static VALUE model_nobs(VALUE self) {
|
709
|
-
|
720
|
+
return INT2FIX(get_model(self)->nobs);
|
710
721
|
}
|
711
722
|
|
712
723
|
static VALUE model_nftr(VALUE self) {
|
713
|
-
|
724
|
+
return INT2FIX(get_model(self)->nftr);
|
714
725
|
}
|
715
726
|
|
716
727
|
static VALUE model_total(VALUE self) {
|
717
|
-
|
728
|
+
return rb_float_new(get_model(self)->total);
|
718
729
|
}
|
719
730
|
|
720
731
|
|
721
732
|
// Instance methods
|
722
733
|
|
723
734
|
static VALUE model_sync(VALUE self) {
|
724
|
-
|
725
|
-
|
735
|
+
mdl_sync(get_model(self));
|
736
|
+
return self;
|
726
737
|
}
|
727
738
|
|
728
739
|
static VALUE model_compact(VALUE self) {
|
729
|
-
|
730
|
-
|
740
|
+
mdl_compact(get_model(self));
|
741
|
+
return self;
|
731
742
|
}
|
732
743
|
|
733
744
|
// call-seq:
|
@@ -737,400 +748,398 @@ static VALUE model_compact(VALUE self) {
|
|
737
748
|
// Saves the model to a file. Uses the Model's path if no argument given,
|
738
749
|
// otherwise uses the passed-in argument as the Model's path.
|
739
750
|
static VALUE model_save(int argc, VALUE *argv, VALUE self) {
|
740
|
-
|
741
|
-
|
742
|
-
|
743
|
-
|
744
|
-
|
745
|
-
|
746
|
-
|
747
|
-
|
748
|
-
|
749
|
-
|
750
|
-
|
751
|
-
|
752
|
-
|
753
|
-
|
754
|
-
|
755
|
-
|
756
|
-
|
757
|
-
|
758
|
-
|
759
|
-
|
760
|
-
|
761
|
-
|
762
|
-
|
763
|
-
|
764
|
-
|
765
|
-
|
766
|
-
|
767
|
-
|
768
|
-
|
751
|
+
if (argc > 1) {
|
752
|
+
rb_raise(rb_const_get(rb_mKernel, rb_intern("ArgumentError")),
|
753
|
+
"wrong number of arguments (%d for 0..1)", argc);
|
754
|
+
}
|
755
|
+
|
756
|
+
mdl_t *model = get_model(self);
|
757
|
+
|
758
|
+
// save passed-in path in options
|
759
|
+
if (argc) {
|
760
|
+
Check_Type(argv[0], T_STRING);
|
761
|
+
rb_ivar_set(self, rb_intern("@path"), argv[0]);
|
762
|
+
}
|
763
|
+
|
764
|
+
// open the output file
|
765
|
+
FILE *file = 0;
|
766
|
+
VALUE path = rb_ivar_get(self, rb_intern("@path"));
|
767
|
+
|
768
|
+
if (NIL_P(path)) {
|
769
|
+
rb_raise(cNativeError, "failed to save model: no path given");
|
770
|
+
}
|
771
|
+
|
772
|
+
if (!(file = fopen(StringValueCStr(path), "w"))) {
|
773
|
+
rb_raise(cNativeError, "failed to save model: failed to open model file");
|
774
|
+
}
|
775
|
+
|
776
|
+
mdl_save(model, file);
|
777
|
+
fclose(file);
|
778
|
+
|
779
|
+
return self;
|
769
780
|
}
|
770
781
|
|
771
782
|
static VALUE model_load(int argc, VALUE *argv, VALUE self) {
|
772
|
-
|
773
|
-
|
774
|
-
|
775
|
-
|
776
|
-
|
777
|
-
|
778
|
-
|
779
|
-
|
780
|
-
|
781
|
-
|
782
|
-
|
783
|
-
|
784
|
-
|
785
|
-
|
786
|
-
|
787
|
-
|
788
|
-
|
789
|
-
|
790
|
-
|
791
|
-
|
792
|
-
|
793
|
-
|
794
|
-
|
795
|
-
|
796
|
-
|
797
|
-
|
798
|
-
|
799
|
-
|
800
|
-
|
783
|
+
if (argc > 1) {
|
784
|
+
rb_raise(rb_const_get(rb_mKernel, rb_intern("ArgumentError")),
|
785
|
+
"wrong number of arguments (%d for 0..1)", argc);
|
786
|
+
}
|
787
|
+
|
788
|
+
mdl_t *model = get_model(self);
|
789
|
+
|
790
|
+
// save passed-in argument in options
|
791
|
+
if (argc) {
|
792
|
+
Check_Type(argv[0], T_STRING);
|
793
|
+
rb_ivar_set(self, rb_intern("@path"), argv[0]);
|
794
|
+
}
|
795
|
+
|
796
|
+
// open the model file
|
797
|
+
FILE *file = 0;
|
798
|
+
VALUE path = rb_ivar_get(self, rb_intern("@path"));
|
799
|
+
|
800
|
+
if (NIL_P(path)) {
|
801
|
+
rb_raise(cNativeError, "failed to load model: no path given");
|
802
|
+
}
|
803
|
+
|
804
|
+
if (!(file = fopen(StringValueCStr(path), "r"))) {
|
805
|
+
rb_raise(cNativeError, "failed to load model: failed to open model file");
|
806
|
+
}
|
807
|
+
|
808
|
+
mdl_load(model, file);
|
809
|
+
fclose(file);
|
810
|
+
|
811
|
+
return self;
|
801
812
|
}
|
802
813
|
|
803
814
|
static dat_t *to_dat(rdr_t *reader, VALUE data, bool labelled) {
|
804
|
-
|
805
|
-
|
806
|
-
|
807
|
-
|
808
|
-
|
809
|
-
|
810
|
-
|
811
|
-
|
812
|
-
|
813
|
-
|
814
|
-
|
815
|
-
|
816
|
-
|
817
|
-
|
818
|
-
|
819
|
-
|
820
|
-
|
821
|
-
|
822
|
-
|
823
|
-
|
824
|
-
|
825
|
-
|
826
|
-
|
827
|
-
|
828
|
-
|
829
|
-
|
830
|
-
|
831
|
-
|
832
|
-
|
833
|
-
|
834
|
-
|
835
|
-
|
836
|
-
|
837
|
-
|
838
|
-
|
839
|
-
|
840
|
-
|
841
|
-
|
842
|
-
|
843
|
-
|
844
|
-
|
845
|
-
|
846
|
-
|
847
|
-
|
848
|
-
|
849
|
-
|
815
|
+
Check_Type(data, T_ARRAY);
|
816
|
+
|
817
|
+
const unsigned int n = RARRAY_LEN(data);
|
818
|
+
unsigned int i, j, k;
|
819
|
+
|
820
|
+
dat_t *dat = wapiti_xmalloc(sizeof(dat_t));
|
821
|
+
dat->nseq = 0;
|
822
|
+
dat->mlen = 0;
|
823
|
+
dat->lbl = labelled;
|
824
|
+
dat->seq = wapiti_xmalloc(sizeof(seq_t*) * n);
|
825
|
+
|
826
|
+
for (i = 0; i < n; ++i) {
|
827
|
+
VALUE sequence = rb_ary_entry(data, i);
|
828
|
+
Check_Type(sequence, T_ARRAY);
|
829
|
+
|
830
|
+
k = RARRAY_LEN(sequence);
|
831
|
+
raw_t *raw = wapiti_xmalloc(sizeof(raw_t) + sizeof(char*) * k);
|
832
|
+
|
833
|
+
for (j = 0; j < k; ++j) {
|
834
|
+
VALUE line = rb_ary_entry(sequence, j);
|
835
|
+
Check_Type(line, T_STRING);
|
836
|
+
raw->lines[j] = StringValueCStr(line);
|
837
|
+
}
|
838
|
+
|
839
|
+
raw->len = k;
|
840
|
+
|
841
|
+
seq_t *seq = rdr_raw2seq(reader, raw, labelled);
|
842
|
+
xfree(raw);
|
843
|
+
|
844
|
+
if (seq == 0) { break; }
|
845
|
+
|
846
|
+
// and store the sequence
|
847
|
+
dat->seq[dat->nseq++] = seq;
|
848
|
+
dat->mlen = max(dat->mlen, seq->len);
|
849
|
+
|
850
|
+
}
|
851
|
+
|
852
|
+
// if no sequence was read, free memory
|
853
|
+
if (dat->nseq == 0) {
|
854
|
+
xfree(dat->seq);
|
855
|
+
xfree(dat);
|
856
|
+
|
857
|
+
return 0;
|
858
|
+
}
|
859
|
+
|
860
|
+
return dat;
|
850
861
|
}
|
851
862
|
|
852
863
|
|
853
864
|
static VALUE model_train(VALUE self, VALUE data) {
|
854
|
-
|
855
|
-
|
856
|
-
|
857
|
-
|
858
|
-
|
859
|
-
|
860
|
-
|
861
|
-
|
862
|
-
|
863
|
-
|
864
|
-
|
865
|
-
|
866
|
-
|
867
|
-
|
868
|
-
|
869
|
-
|
870
|
-
|
871
|
-
|
872
|
-
|
873
|
-
|
874
|
-
|
875
|
-
|
876
|
-
|
877
|
-
|
878
|
-
|
879
|
-
|
880
|
-
|
881
|
-
|
882
|
-
|
883
|
-
|
884
|
-
|
885
|
-
|
886
|
-
|
887
|
-
|
888
|
-
|
889
|
-
|
890
|
-
|
891
|
-
|
892
|
-
|
893
|
-
|
894
|
-
|
895
|
-
|
896
|
-
|
897
|
-
|
898
|
-
|
899
|
-
|
900
|
-
|
901
|
-
|
902
|
-
|
903
|
-
|
904
|
-
|
905
|
-
|
906
|
-
|
907
|
-
|
908
|
-
|
909
|
-
|
910
|
-
|
911
|
-
|
912
|
-
|
913
|
-
|
914
|
-
|
915
|
-
|
916
|
-
|
917
|
-
|
918
|
-
|
919
|
-
|
920
|
-
|
921
|
-
|
922
|
-
|
923
|
-
|
924
|
-
|
925
|
-
|
926
|
-
|
927
|
-
|
928
|
-
|
929
|
-
|
930
|
-
|
931
|
-
|
932
|
-
|
933
|
-
|
934
|
-
|
935
|
-
|
936
|
-
|
937
|
-
|
938
|
-
|
939
|
-
|
940
|
-
|
941
|
-
|
942
|
-
|
943
|
-
return self;
|
865
|
+
|
866
|
+
mdl_t* model = get_model(self);
|
867
|
+
|
868
|
+
int trn;
|
869
|
+
for (trn = 0; trn < trn_cnt; trn++) {
|
870
|
+
if (!strcmp(model->opt->algo, trn_lst[trn].name)) break;
|
871
|
+
}
|
872
|
+
|
873
|
+
if (trn == trn_cnt) {
|
874
|
+
rb_raise(cNativeError, "failed to train model: unknown algorithm '%s'", model->opt->algo);
|
875
|
+
}
|
876
|
+
|
877
|
+
FILE *file;
|
878
|
+
|
879
|
+
// Load the pattern file. This will unlock the database if previously
|
880
|
+
// locked by loading a model.
|
881
|
+
if (model->opt->pattern) {
|
882
|
+
file = fopen(model->opt->pattern, "r");
|
883
|
+
|
884
|
+
if (!file) {
|
885
|
+
rb_raise(cNativeError, "failed to train model: failed to load pattern file '%s'", model->opt->pattern);
|
886
|
+
}
|
887
|
+
|
888
|
+
rdr_loadpat(model->reader, file);
|
889
|
+
fclose(file);
|
890
|
+
}
|
891
|
+
else {
|
892
|
+
// rb_raise(cNativeError, "failed to train model: no pattern given");
|
893
|
+
}
|
894
|
+
|
895
|
+
qrk_lock(model->reader->obs, false);
|
896
|
+
|
897
|
+
|
898
|
+
// Load the training data. When this is done we lock the quarks as we
|
899
|
+
// don't want to put in the model, informations present only in the
|
900
|
+
// devlopment set.
|
901
|
+
|
902
|
+
switch (TYPE(data)) {
|
903
|
+
case T_STRING:
|
904
|
+
if (!(file = fopen(StringValuePtr(data), "r"))) {
|
905
|
+
rb_raise(cNativeError, "failed to train model: failed to open training data '%s", StringValuePtr(data));
|
906
|
+
}
|
907
|
+
|
908
|
+
model->train = rdr_readdat(model->reader, file, true);
|
909
|
+
fclose(file);
|
910
|
+
|
911
|
+
break;
|
912
|
+
case T_ARRAY:
|
913
|
+
model->train = to_dat(model->reader, data, true);
|
914
|
+
|
915
|
+
break;
|
916
|
+
default:
|
917
|
+
rb_raise(cNativeError, "failed to train model: invalid training data type (expected instance of String or Array)");
|
918
|
+
}
|
919
|
+
|
920
|
+
qrk_lock(model->reader->lbl, true);
|
921
|
+
qrk_lock(model->reader->obs, true);
|
922
|
+
|
923
|
+
if (!model->train || model->train->nseq == 0) {
|
924
|
+
rb_raise(cNativeError, "failed to train model: no training data loaded");
|
925
|
+
}
|
926
|
+
|
927
|
+
// If present, load the development set in the model. If not specified,
|
928
|
+
// the training dataset will be used instead.
|
929
|
+
if (model->opt->devel) {
|
930
|
+
if (!(file = fopen(model->opt->devel, "r"))) {
|
931
|
+
rb_raise(cNativeError, "failed to train model: cannot open development file '%s'", model->opt->devel);
|
932
|
+
}
|
933
|
+
|
934
|
+
model->devel = rdr_readdat(model->reader, file, true);
|
935
|
+
fclose(file);
|
936
|
+
}
|
937
|
+
|
938
|
+
// Initialize the model. If a previous model was loaded, this will be
|
939
|
+
// just a resync, else the model structure will be created.
|
940
|
+
rb_funcall(self, rb_intern("sync"), 0);
|
941
|
+
|
942
|
+
// Train the model.
|
943
|
+
uit_setup(model);
|
944
|
+
trn_lst[trn].train(model);
|
945
|
+
uit_cleanup(model);
|
946
|
+
|
947
|
+
// If requested compact the model.
|
948
|
+
if (model->opt->compact) {
|
949
|
+
rb_funcall(self, rb_intern("compact"), 0);
|
950
|
+
}
|
951
|
+
|
952
|
+
return self;
|
944
953
|
}
|
945
954
|
|
946
955
|
// Returns a sorted list of all labels in the Model's label database.
|
947
956
|
static VALUE model_labels(VALUE self) {
|
948
|
-
|
949
|
-
|
950
|
-
|
951
|
-
|
952
|
-
|
953
|
-
|
954
|
-
|
955
|
-
|
956
|
-
|
957
|
-
|
958
|
-
|
959
|
-
|
960
|
-
|
961
|
-
|
957
|
+
mdl_t *model = get_model(self);
|
958
|
+
const uint32_t Y = model->nlbl;
|
959
|
+
|
960
|
+
qrk_t *lp = model->reader->lbl;
|
961
|
+
|
962
|
+
VALUE labels = rb_ary_new2(Y);
|
963
|
+
|
964
|
+
for (unsigned int i = 0; i < Y; ++i) {
|
965
|
+
rb_ary_push(labels, rb_str_new2(qrk_id2str(lp, i)));
|
966
|
+
}
|
967
|
+
|
968
|
+
rb_funcall(labels, rb_intern("sort!"), 0);
|
969
|
+
|
970
|
+
return labels;
|
962
971
|
}
|
963
972
|
|
964
973
|
static VALUE decode_sequence(VALUE self, mdl_t *model, raw_t *raw) {
|
965
|
-
|
966
|
-
|
967
|
-
|
968
|
-
|
969
|
-
|
970
|
-
|
971
|
-
|
972
|
-
|
973
|
-
|
974
|
-
|
975
|
-
|
976
|
-
|
977
|
-
|
978
|
-
|
979
|
-
|
980
|
-
|
981
|
-
|
982
|
-
|
983
|
-
|
984
|
-
|
985
|
-
|
986
|
-
|
987
|
-
|
988
|
-
|
989
|
-
|
990
|
-
|
991
|
-
|
992
|
-
|
993
|
-
|
994
|
-
|
995
|
-
|
996
|
-
|
997
|
-
|
998
|
-
|
999
|
-
|
1000
|
-
|
1001
|
-
|
1002
|
-
|
1003
|
-
|
1004
|
-
|
1005
|
-
|
1006
|
-
|
1007
|
-
|
1008
|
-
|
1009
|
-
|
1010
|
-
|
1011
|
-
|
1012
|
-
|
1013
|
-
|
1014
|
-
|
1015
|
-
|
1016
|
-
|
1017
|
-
|
1018
|
-
|
1019
|
-
|
1020
|
-
|
1021
|
-
|
1022
|
-
|
1023
|
-
|
1024
|
-
|
1025
|
-
|
1026
|
-
|
1027
|
-
|
1028
|
-
|
1029
|
-
|
1030
|
-
|
1031
|
-
|
1032
|
-
|
1033
|
-
|
1034
|
-
|
1035
|
-
|
1036
|
-
|
1037
|
-
|
1038
|
-
|
1039
|
-
|
1040
|
-
|
1041
|
-
|
1042
|
-
|
1043
|
-
|
1044
|
-
|
1045
|
-
|
1046
|
-
|
1047
|
-
|
1048
|
-
|
1049
|
-
|
1050
|
-
|
1051
|
-
|
1052
|
-
|
1053
|
-
|
1054
|
-
|
1055
|
-
|
1056
|
-
|
1057
|
-
|
1058
|
-
|
1059
|
-
|
1060
|
-
|
1061
|
-
|
1062
|
-
|
1063
|
-
|
1064
|
-
|
1065
|
-
|
1066
|
-
|
1067
|
-
|
974
|
+
qrk_t *lbls = model->reader->lbl;
|
975
|
+
|
976
|
+
const unsigned int Y = model->nlbl;
|
977
|
+
const unsigned int N = model->opt->nbest;
|
978
|
+
|
979
|
+
seq_t *seq = rdr_raw2seq(model->reader, raw, model->opt->check);
|
980
|
+
|
981
|
+
const unsigned int T = seq->len;
|
982
|
+
unsigned int n, t, tcnt = 0, terr = 0, scnt = 0, serr = 0, stat[3][Y];
|
983
|
+
|
984
|
+
uint32_t *out = wapiti_xmalloc(sizeof(uint32_t) * T * N);
|
985
|
+
double *psc = wapiti_xmalloc(sizeof(double) * T * N);
|
986
|
+
double *scs = wapiti_xmalloc(sizeof(double) * N);
|
987
|
+
|
988
|
+
VALUE sequence, tokens;
|
989
|
+
|
990
|
+
if (N == 1) {
|
991
|
+
tag_viterbi(model, seq, out, scs, psc);
|
992
|
+
}
|
993
|
+
else {
|
994
|
+
tag_nbviterbi(model, seq, N, (void*)out, scs, (void*)psc);
|
995
|
+
}
|
996
|
+
|
997
|
+
sequence = rb_ary_new();
|
998
|
+
|
999
|
+
for (t = 0; t < T; ++t) {
|
1000
|
+
tokens = rb_ary_new();
|
1001
|
+
|
1002
|
+
if (!model->opt->label) {
|
1003
|
+
VALUE token = rb_str_new2(raw->lines[t]);
|
1004
|
+
|
1005
|
+
#ifdef HAVE_RUBY_ENCODING_H
|
1006
|
+
int enc = rb_enc_find_index("UTF-8");
|
1007
|
+
rb_enc_associate_index(token, enc);
|
1008
|
+
#endif
|
1009
|
+
|
1010
|
+
rb_ary_push(tokens, token);
|
1011
|
+
}
|
1012
|
+
|
1013
|
+
for (n = 0; n < N; ++n) {
|
1014
|
+
|
1015
|
+
uint64_t lbl = out[t * N + n];
|
1016
|
+
rb_ary_push(tokens, rb_str_new2(qrk_id2str(lbls, lbl)));
|
1017
|
+
|
1018
|
+
// output individual score
|
1019
|
+
if (model->opt->outsc) {
|
1020
|
+
rb_ary_push(tokens, rb_float_new(psc[t * N + n]));
|
1021
|
+
}
|
1022
|
+
|
1023
|
+
}
|
1024
|
+
|
1025
|
+
// yield token/label pair to block if given
|
1026
|
+
if (rb_block_given_p()) {
|
1027
|
+
tokens = rb_yield(tokens);
|
1028
|
+
}
|
1029
|
+
|
1030
|
+
rb_ary_push(sequence, tokens);
|
1031
|
+
|
1032
|
+
|
1033
|
+
// TODO output sequence score: scs[n] (float)
|
1034
|
+
|
1035
|
+
}
|
1036
|
+
|
1037
|
+
// Statistics
|
1038
|
+
if (model->opt->check) {
|
1039
|
+
int err = 0;
|
1040
|
+
|
1041
|
+
for (t = 0; t < T; ++t) {
|
1042
|
+
stat[0][seq->pos[t].lbl]++;
|
1043
|
+
stat[1][out[t * N]]++;
|
1044
|
+
|
1045
|
+
if (seq->pos[t].lbl != out[t * N]) {
|
1046
|
+
terr++;
|
1047
|
+
err = 1;
|
1048
|
+
}
|
1049
|
+
else {
|
1050
|
+
stat[2][out[t * N]]++;
|
1051
|
+
}
|
1052
|
+
}
|
1053
|
+
|
1054
|
+
tcnt = FIX2INT(rb_ivar_get(self, rb_intern("@token_count")));
|
1055
|
+
rb_ivar_set(self, rb_intern("@token_count"), INT2FIX(tcnt + (unsigned int)T));
|
1056
|
+
|
1057
|
+
terr += FIX2INT(rb_ivar_get(self, rb_intern("@token_errors")));
|
1058
|
+
rb_ivar_set(self, rb_intern("@token_errors"), INT2FIX(terr));
|
1059
|
+
|
1060
|
+
scnt = FIX2INT(rb_ivar_get(self, rb_intern("@sequence_count")));
|
1061
|
+
rb_ivar_set(self, rb_intern("@sequence_count"), INT2FIX(++scnt));
|
1062
|
+
|
1063
|
+
serr = FIX2INT(rb_ivar_get(self, rb_intern("@sequence_errors")));
|
1064
|
+
rb_ivar_set(self, rb_intern("@sequence_errors"), INT2FIX(serr + err));
|
1065
|
+
|
1066
|
+
}
|
1067
|
+
|
1068
|
+
|
1069
|
+
// Cleanup memory used for this sequence
|
1070
|
+
xfree(scs);
|
1071
|
+
xfree(psc);
|
1072
|
+
xfree(out);
|
1073
|
+
|
1074
|
+
rdr_freeseq(seq);
|
1075
|
+
|
1076
|
+
return sequence;
|
1068
1077
|
}
|
1069
1078
|
|
1070
1079
|
static VALUE decode_sequence_array(VALUE self, VALUE array) {
|
1071
|
-
|
1072
|
-
|
1073
|
-
|
1074
|
-
|
1075
|
-
|
1076
|
-
|
1077
|
-
|
1078
|
-
|
1079
|
-
|
1080
|
-
|
1081
|
-
|
1082
|
-
|
1083
|
-
|
1084
|
-
|
1085
|
-
|
1086
|
-
|
1087
|
-
|
1088
|
-
|
1089
|
-
|
1090
|
-
|
1091
|
-
|
1092
|
-
|
1093
|
-
|
1094
|
-
|
1095
|
-
|
1096
|
-
|
1097
|
-
|
1098
|
-
|
1099
|
-
|
1100
|
-
|
1101
|
-
|
1102
|
-
|
1080
|
+
Check_Type(array, T_ARRAY);
|
1081
|
+
const unsigned int n = RARRAY_LEN(array);
|
1082
|
+
|
1083
|
+
mdl_t *model = get_model(self);
|
1084
|
+
raw_t *raw;
|
1085
|
+
|
1086
|
+
const unsigned int N = model->opt->nbest;
|
1087
|
+
unsigned int i, j;
|
1088
|
+
|
1089
|
+
VALUE result = rb_ary_new2(n * N), sequence;
|
1090
|
+
|
1091
|
+
for (i = 0; i < n; ++i) {
|
1092
|
+
sequence = rb_ary_entry(array, i);
|
1093
|
+
Check_Type(sequence, T_ARRAY);
|
1094
|
+
|
1095
|
+
const unsigned int k = RARRAY_LEN(sequence);
|
1096
|
+
raw = wapiti_xmalloc(sizeof(raw_t) + sizeof(char*) * k);
|
1097
|
+
raw->len = k;
|
1098
|
+
|
1099
|
+
for (j = 0; j < k; ++j) {
|
1100
|
+
VALUE line = rb_ary_entry(sequence, j);
|
1101
|
+
Check_Type(line, T_STRING);
|
1102
|
+
|
1103
|
+
raw->lines[j] = StringValueCStr(line);
|
1104
|
+
}
|
1105
|
+
|
1106
|
+
rb_ary_push(result, decode_sequence(self, model, raw));
|
1107
|
+
|
1108
|
+
xfree(raw);
|
1109
|
+
}
|
1110
|
+
|
1111
|
+
return result;
|
1103
1112
|
}
|
1104
1113
|
|
1105
1114
|
static VALUE decode_sequence_file(VALUE self, VALUE path) {
|
1106
|
-
|
1107
|
-
|
1108
|
-
|
1109
|
-
|
1110
|
-
|
1111
|
-
|
1112
|
-
|
1113
|
-
|
1114
|
-
|
1115
|
-
|
1116
|
-
|
1117
|
-
|
1118
|
-
|
1119
|
-
|
1120
|
-
|
1121
|
-
|
1122
|
-
|
1123
|
-
|
1124
|
-
|
1125
|
-
|
1126
|
-
|
1127
|
-
|
1128
|
-
|
1129
|
-
|
1130
|
-
|
1131
|
-
|
1132
|
-
|
1133
|
-
|
1115
|
+
Check_Type(path, T_STRING);
|
1116
|
+
FILE *file;
|
1117
|
+
|
1118
|
+
if (!(file = fopen(StringValueCStr(path), "r"))) {
|
1119
|
+
rb_raise(cNativeError, "failed to label data: could not open file '%s'", StringValueCStr(path));
|
1120
|
+
}
|
1121
|
+
|
1122
|
+
mdl_t *model = get_model(self);
|
1123
|
+
raw_t *raw;
|
1124
|
+
|
1125
|
+
VALUE result = rb_ary_new();
|
1126
|
+
|
1127
|
+
// Next read the input file sequence by sequence and label them, we have
|
1128
|
+
// to take care of not discarding the raw input as we want to send it
|
1129
|
+
// back to the output with the additional predicted labels.
|
1130
|
+
while (!feof(file)) {
|
1131
|
+
|
1132
|
+
// So, first read an input sequence keeping the raw_t object
|
1133
|
+
// available, and label it with Viterbi.
|
1134
|
+
if ((raw = rdr_readraw(model->reader, file)) == 0) {
|
1135
|
+
break;
|
1136
|
+
}
|
1137
|
+
|
1138
|
+
rb_ary_push(result, decode_sequence(self, model, raw));
|
1139
|
+
rdr_freeraw(raw);
|
1140
|
+
}
|
1141
|
+
|
1142
|
+
return result;
|
1134
1143
|
}
|
1135
1144
|
|
1136
1145
|
// cal-seq:
|
@@ -1138,144 +1147,146 @@ static VALUE decode_sequence_file(VALUE self, VALUE path) {
|
|
1138
1147
|
// m.label(filename, options = {}) # => array of labelled tokens
|
1139
1148
|
//
|
1140
1149
|
static VALUE model_label(VALUE self, VALUE data) {
|
1141
|
-
|
1142
|
-
|
1143
|
-
|
1144
|
-
|
1145
|
-
|
1146
|
-
|
1147
|
-
|
1148
|
-
|
1149
|
-
|
1150
|
-
|
1151
|
-
|
1152
|
-
|
1153
|
-
|
1154
|
-
|
1150
|
+
VALUE result;
|
1151
|
+
|
1152
|
+
switch (TYPE(data)) {
|
1153
|
+
case T_STRING:
|
1154
|
+
result = decode_sequence_file(self, data);
|
1155
|
+
break;
|
1156
|
+
case T_ARRAY:
|
1157
|
+
result = decode_sequence_array(self, data);
|
1158
|
+
break;
|
1159
|
+
default:
|
1160
|
+
rb_raise(cNativeError, "failed to label data: invalid data (expected type String or Array)");
|
1161
|
+
}
|
1162
|
+
|
1163
|
+
return result;
|
1155
1164
|
}
|
1156
1165
|
|
1157
1166
|
static void Init_model() {
|
1158
|
-
|
1159
|
-
|
1160
|
-
|
1161
|
-
|
1167
|
+
cModel = rb_define_class_under(mWapiti, "Model", rb_cObject);
|
1168
|
+
rb_define_alloc_func(cModel, allocate_model);
|
1169
|
+
|
1170
|
+
rb_define_method(cModel, "initialize", initialize_model, -1);
|
1171
|
+
|
1172
|
+
rb_define_attr(cModel, "options", 1, 0);
|
1173
|
+
|
1162
1174
|
|
1163
|
-
|
1175
|
+
rb_define_method(cModel, "nlbl", model_nlbl, 0);
|
1176
|
+
rb_define_method(cModel, "labels", model_labels, 0);
|
1164
1177
|
|
1165
|
-
|
1166
|
-
|
1167
|
-
rb_define_method(cModel, "labels", model_labels, 0);
|
1168
|
-
|
1169
|
-
rb_define_method(cModel, "nobs", model_nobs, 0);
|
1170
|
-
rb_define_alias(cModel, "observations", "nobs");
|
1178
|
+
rb_define_method(cModel, "nobs", model_nobs, 0);
|
1179
|
+
rb_define_alias(cModel, "observations", "nobs");
|
1171
1180
|
|
1172
|
-
|
1173
|
-
|
1181
|
+
rb_define_method(cModel, "nftr", model_nftr, 0);
|
1182
|
+
rb_define_alias(cModel, "features", "nftr");
|
1174
1183
|
|
1175
|
-
|
1184
|
+
rb_define_method(cModel, "total", model_total, 0);
|
1176
1185
|
|
1177
|
-
|
1178
|
-
|
1179
|
-
|
1180
|
-
|
1186
|
+
rb_define_method(cModel, "sync", model_sync, 0);
|
1187
|
+
rb_define_method(cModel, "compact", model_compact, 0);
|
1188
|
+
rb_define_method(cModel, "save", model_save, -1);
|
1189
|
+
rb_define_method(cModel, "load", model_load, -1);
|
1181
1190
|
|
1182
|
-
|
1183
|
-
|
1191
|
+
rb_define_method(cModel, "train", model_train, 1);
|
1192
|
+
rb_define_method(cModel, "label", model_label, 1);
|
1184
1193
|
}
|
1185
1194
|
|
1186
1195
|
/* --- Top-Level Utility Methods --- */
|
1187
1196
|
|
1188
1197
|
|
1189
1198
|
static VALUE label(VALUE self __attribute__((__unused__)), VALUE rb_options) {
|
1190
|
-
|
1191
|
-
|
1192
|
-
|
1199
|
+
if (strncmp("Wapiti::Options", rb_obj_classname(rb_options), 15) != 0) {
|
1200
|
+
rb_raise(cNativeError, "argument must be a native options instance");
|
1201
|
+
}
|
1193
1202
|
|
1194
|
-
|
1203
|
+
opt_t *options = get_options(rb_options);
|
1195
1204
|
|
1196
|
-
|
1197
|
-
|
1198
|
-
|
1205
|
+
if (options->mode != 1) {
|
1206
|
+
rb_raise(cNativeError, "invalid options argument: mode should be set to 1 for labelling");
|
1207
|
+
}
|
1199
1208
|
|
1200
|
-
|
1201
|
-
|
1209
|
+
mdl_t *model = mdl_new(rdr_new(options->maxent));
|
1210
|
+
model->opt = options;
|
1202
1211
|
|
1203
|
-
|
1204
|
-
|
1205
|
-
mdl_free(model);
|
1212
|
+
dolabel(model);
|
1206
1213
|
|
1207
|
-
|
1214
|
+
mdl_free(model);
|
1215
|
+
|
1216
|
+
return Qnil;
|
1208
1217
|
}
|
1209
1218
|
|
1219
|
+
#if defined EXTRA
|
1210
1220
|
static VALUE dump(VALUE self __attribute__((__unused__)), VALUE rb_options) {
|
1211
|
-
|
1212
|
-
|
1213
|
-
|
1221
|
+
if (strncmp("Wapiti::Options", rb_obj_classname(rb_options), 15) != 0) {
|
1222
|
+
rb_raise(cNativeError, "argument must be a native options instance");
|
1223
|
+
}
|
1224
|
+
|
1225
|
+
opt_t *options = get_options(rb_options);
|
1214
1226
|
|
1215
|
-
|
1227
|
+
if (options->mode != 2) {
|
1228
|
+
rb_raise(cNativeError, "invalid options argument: mode should be set to 2 for training");
|
1229
|
+
}
|
1216
1230
|
|
1217
|
-
|
1218
|
-
|
1219
|
-
}
|
1231
|
+
mdl_t *model = mdl_new(rdr_new(options->maxent));
|
1232
|
+
model->opt = options;
|
1220
1233
|
|
1221
|
-
|
1222
|
-
model->opt = options;
|
1234
|
+
dodump(model);
|
1223
1235
|
|
1224
|
-
|
1225
|
-
|
1226
|
-
mdl_free(model);
|
1236
|
+
mdl_free(model);
|
1227
1237
|
|
1228
|
-
|
1238
|
+
return Qnil;
|
1229
1239
|
}
|
1230
1240
|
|
1231
1241
|
// This function is a proxy for Wapiti's main entry point.
|
1232
1242
|
static VALUE wapiti(VALUE self __attribute__((__unused__)), VALUE arguments) {
|
1233
|
-
|
1234
|
-
|
1235
|
-
|
1236
|
-
|
1237
|
-
|
1238
|
-
|
1239
|
-
|
1240
|
-
|
1241
|
-
|
1242
|
-
|
1243
|
-
|
1244
|
-
|
1245
|
-
|
1246
|
-
|
1247
|
-
|
1248
|
-
|
1249
|
-
|
1250
|
-
|
1251
|
-
|
1252
|
-
|
1253
|
-
|
1254
|
-
|
1255
|
-
|
1256
|
-
|
1257
|
-
|
1258
|
-
|
1259
|
-
|
1260
|
-
|
1261
|
-
|
1243
|
+
int result = -1, argc = 0;
|
1244
|
+
char **ap, *argv[18], *input, *tmp;
|
1245
|
+
|
1246
|
+
Check_Type(arguments, T_STRING);
|
1247
|
+
tmp = StringValueCStr(arguments);
|
1248
|
+
|
1249
|
+
// allocate space for argument vector
|
1250
|
+
input = (char*)malloc(strlen(tmp) + 8);
|
1251
|
+
|
1252
|
+
// prepend command name
|
1253
|
+
strncpy(input, "wapiti ", 8);
|
1254
|
+
strncat(input, tmp, strlen(input) - 8);
|
1255
|
+
|
1256
|
+
// remember allocation pointer
|
1257
|
+
tmp = input;
|
1258
|
+
|
1259
|
+
// turn input string into argument vector (using
|
1260
|
+
// only the first seventeen tokens from input)
|
1261
|
+
for (ap = argv; (*ap = strsep(&input, " \t")) != (char*)0; ++argc) {
|
1262
|
+
if ((**ap != '\0') && (++ap >= &argv[18])) break;
|
1263
|
+
}
|
1264
|
+
|
1265
|
+
// call main entry point
|
1266
|
+
result = wapiti_main(argc, argv);
|
1267
|
+
|
1268
|
+
// free allocated memory
|
1269
|
+
free(tmp);
|
1270
|
+
|
1271
|
+
return INT2FIX(result);
|
1262
1272
|
}
|
1273
|
+
#endif
|
1263
1274
|
|
1264
1275
|
/* --- Wapiti Extension Entry Point --- */
|
1265
1276
|
|
1266
1277
|
void Init_native() {
|
1267
|
-
|
1268
|
-
|
1269
|
-
|
1270
|
-
|
1271
|
-
|
1272
|
-
|
1273
|
-
|
1274
|
-
|
1275
|
-
|
1276
|
-
|
1277
|
-
|
1278
|
-
|
1279
|
-
|
1280
|
-
|
1281
|
-
}
|
1278
|
+
mWapiti = rb_const_get(rb_mKernel, rb_intern("Wapiti"));
|
1279
|
+
mNative = rb_define_module_under(mWapiti, "Native");
|
1280
|
+
|
1281
|
+
cNativeError = rb_const_get(mWapiti, rb_intern("NativeError"));
|
1282
|
+
cConfigurationError = rb_const_get(mWapiti, rb_intern("ConfigurationError"));
|
1283
|
+
cLogger = rb_funcall(mWapiti, rb_intern("log"), 0);
|
1284
|
+
|
1285
|
+
rb_define_singleton_method(mNative, "label", label, 1);
|
1286
|
+
// rb_define_singleton_method(mNative, "wapiti", wapiti, 1);
|
1287
|
+
|
1288
|
+
rb_define_const(mNative, "VERSION", rb_str_new2(VERSION));
|
1289
|
+
|
1290
|
+
Init_options();
|
1291
|
+
Init_model();
|
1292
|
+
}
|