1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
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
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
|
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;<--IFD-->;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
;; ;;
;; Istituto di Fonetica e Dialettologia ;;
;; Consiglio Nazionale delle Ricerche ;;
;; Padova Italy ;;
;; Copyright (c) 2000 ;;
;; ;;
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;<--IFD-->;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
;; ;;
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;<--IRST-->;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
;; ;;
;; Centro per la ricerca scientifica e tecnologica ;;
;; Povo (TN) - Italy ;;
;; Copyright (c) 2000 ;;
;; ;;
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;<--IRST-->;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
;;;
;;; Definizione di alcuni moduli scritti appositamente per l'italiano
;;;
;original da rimettere dopo i training SOLEML
(define (Function_Word utt)
"Crea la features FW (Function_Word) nella relation Word"
(let (w w1)
(set! w (utt.relation.first utt 'Word))
(while w
;(print (item.name w))
(if (string-matches (item.name w) ".+'.+") ;caso parole con apostrofi
(let ()
(set! w1 (string-append (string-before (item.name w) "'") "'"))
(item.set_feat w 'FW (italian_fw_gpos w1 italian_guess_pos))
(item.set_feat w 'apo "1")) ;parola con apostrofo
(item.set_feat w 'FW (italian_fw_gpos (item.name w) italian_guess_pos)))
(set! w (item.next w))))
utt)
(define (Phrase_Type utt)
"Crea la features type nella relation Phrase"
(let (p pu)
(set! p (utt.relation.first utt 'Phrase))
(while p
(set! pu (item.feat p "R:Phrase.daughtern.R:Token.parent.punc"))
(if (string-equal pu ".")
(item.set_feat p 'type "dichiarativa"))
(if (string-equal pu ";")
(item.set_feat p 'type "completativa"))
(if (string-equal pu ":")
(item.set_feat p 'type "esplicativa"))
(if (string-equal pu ",")
(item.set_feat p 'type "pausativa"))
(if (string-equal pu "?")
(item.set_feat p 'type "interrogativa"))
(if (string-equal pu "!")
(item.set_feat p 'type "esclamativa"))
(if (or (string-equal pu "..") (string-equal pu "..."))
(item.set_feat p 'type "sospensiva"))
(if (or (string-equal pu "\"") (string-equal pu ")") (string-equal pu "\]") (string-equal pu "\}"))
(item.set_feat p 'type "riportativa"))
(set! p (item.next p))))
utt)
(define (Position_FW utt)
"Crea la features position FW (Function_Word) nella relation IntEvent"
(let ( P_gr n P_grP S I )
(utt.relation.create utt 'GrRespiro)
(utt.copy_relation utt 'IntEvent 'GrRespiro) ;;questa copia in realt deve prendere solo primafuncionW, prima content e il primo accentede....
;(print "fase 1")
(set! P_gr (utt.relation.first utt 'GrRespiro))
(while P_gr
(set! n (item.name P_gr))
(if (or (string-equal n "Accented") (string-equal n "altreFW"))
(let ()
(set! P_grP (item.prev P_gr))
(item.relation.remove P_gr 'GrRespiro)
(set! P_gr P_grP)
)
)
(set! P_gr (item.next P_gr))
)
;(print "fase 2")
(set! P_gr (utt.relation.first utt 'GrRespiro))
(set! S (utt.relation.first utt 'Syllable))
(set! I (if P_gr (item.relation.parent P_gr 'Intonation) nil))
(if P_gr (item.append_daughter P_gr S))
(if (item.next P_gr)
(set! I (item.relation.parent (item.next P_gr) 'Intonation))
;nel caso che sia l'utimo item non fa nulla (print "fine....")
)
(set! S (item.next S))
(while S
(if (equal? (item.relation I 'Intonation) (item.relation S 'Intonation))
;;Vecchia condizione non compatibile con 131:
;;(string-equal (item.feat I 'id) (item.feat S 'id))
(let ()
(set! P_gr (item.next P_gr))
(item.append_daughter P_gr S)
(if (item.next P_gr)
(set! I (item.relation.parent (item.next P_gr) 'Intonation))
;nel caso che sia l'utimo item non fa nulla (print "fine....")
)
(set! S (item.next S))
)
(while (and S (not (equal? (item.relation I 'Intonation) (item.relation S 'Intonation))))
;(print "f3")
(item.append_daughter P_gr S)
(set! S (item.next S))))))
utt)
(define (Word_dopo_num utt)
"Dopo il modulo numerico (e solo in questo caso) che salva anche la trascrizione sposta le relation in modo appropiato"
(let (p_token w wn word_str true_word_str trasc_word_str)
(set! p_token (utt.relation.first utt 'Token))
(while p_token
(if (string-equal (item.feat p_token 'pos) "Num+Alfa")
(let ((w (item.daughter1 p_token))(wn (item.next (item.daughtern p_token))))
;(print "Token Num+Alfa")
(while (not (equal? w wn))
;(print "non uguale")
(set! word_str (item.name w))
(if (string-matches word_str ".* -T.*")
(let ()
;(print "c' il -T")
(set! true_word_str (string-before word_str " -T"))
(set! trasc_word_str (string-after word_str " -T "))
;(print true_word_str)
(item.set_name w true_word_str)
(item.set_feat w "TRA_NUM" trasc_word_str);da togliere in futuro
)
)
(set! w (item.next w))
)
))
(set! p_token (item.next p_token))))
utt)
(define (Word_split_pos utt)
"Prende il pos complesso della relation word (lessico) e lo divide in tPOS=True POS e GEN=Genere"
(let (p_word)
(set! p_word (utt.relation.first utt 'Word))
(while p_word
(if (string-matches (item.feat p_word 'pos) ".+\-.+")
(let ()
(item.set_feat p_word "tPOS" (string-before (item.feat p_word 'pos) "\-"))
(item.set_feat p_word "GEN" (string-after (item.feat p_word 'pos) "\-"))
)
(item.set_feat p_word "tPOS" (item.feat p_word 'pos))
)
(set! p_word (item.next p_word))))
utt)
(define (Token_punct utt)
"Mette a posto la punteggiatura degli acronimi"
(let (p_token)
(set! p_token (utt.relation.first utt 'Token))
(while p_token
(if (string-equal (item.feat p_token 'pos) "Acronimo") ;;mette a posto la punteggiatura degli acronimi togliendo il primo punto
(let (punt)
(set! punt (string-after (item.feat p_token 'punc) "."))
(item.set_feat p_token 'punc punt)
))
(set! p_token (item.next p_token))))
utt)
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
(define (My_Pauses utt)
"Inserisce pausa prima della prima FW"
(print (utt.relation_tree utt 'Segment))
(set! s (utt.relation.first utt 'Syllable))
(while s
;(print s)
;(print (item.feat s 'R:GrRespiro.parent.name))
;(print (item.feat s 'id))
;;;DA sostituire id se vuoi la compatibilit con 131
(if (and (not (eq? (item.feat s 'syl_in ) 0)) (equal? (item.feat s 'R:GrRespiro.parent.daughter1.id) (item.feat s 'id)) (equal? (item.feat s 'R:GrRespiro.parent.name) "primaFW"))
(let ()
(print "si!")
(insert_pause utt (item.relation.parent (item.prev s) 'SylStructure))
)
)
(set! s (item.next s))
)
)
(define (MAP_segments utt)
"(MAP_segments utt UTT)
Mappa difoni prendendo come riferimento il file ita_map.scm dove se vi sono i difoni da mappare."
(let ((item_segs_list (utt.relation.items utt 'Segment)) )
(while (cdr item_segs_list)
(set! s1 (car item_segs_list))
(set! s2 (car (cdr item_segs_list)))
(let ((ita_maps ita_map))
;Modifica per mappare coppie di foni
(while ita_maps
(if (and (equal? (item.feat s1 'name) (string-append (car (car ita_maps)))) (equal? (item.feat s2 'name) (string-append (car (cdr (car ita_maps))))))
(let ()
(item.set_name s1 (car (car (cdr (cdr (car ita_maps))))))
(item.set_name s2 (car (cdr (car (cdr (cdr (car ita_maps)))))))
))
(set! ita_maps (cdr ita_maps))
))
(set! item_segs_list (cdr item_segs_list))
)
)
utt)
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
(define (string-remove string1 string2)
"(string-remove STRING1 STRING2)
Remove all instances of STRING2 in STRING1 and return the result."
(begin
(while (and (not (equal? string1 string2))
(or (not (equal? (string-before string1 string2) ""))
(not (equal? (string-after string1 string2) ""))
)
)
(let (before after)
(set! before (string-before string1 string2))
(set! after (string-after string1 string2))
(set! string1 (string-append before after))
)
)
(if (equal? string1 string2) "" string1)
)
)
(define (Classic_Pauses_mB utt)
"(Pauses UTT)
Predict pause insertion."
(let ((words (utt.relation.items utt 'Word)) lastword tpname)
(if words
(begin
(insert_initial_pause utt) ;; always have a start pause
(set! lastword (car (last words)))
(mapcar
(lambda (w)
(let ((pbreak (item.feat w "pbreak"))
(emph (item.feat w "R:Token.parent.EMPH")))
(cond
((or (string-equal "mB" pbreak)
(string-equal "B" pbreak)
(string-equal "BB" pbreak))
(insert_pause utt w))
; ((string-equal emph "1")
; (insert_pause utt w))
((equal? w lastword)
(insert_pause utt w)))))
words)
;; The embarassing bit. Remove any words labelled as punc or fpunc
(mapcar
(lambda (w)
(let ((pos (item.feat w "pos")))
(if (or (string-equal "punc" pos)
(string-equal "fpunc" pos))
(let ((pbreak (item.feat w "pbreak"))
(wp (item.relation w 'Phrase)))
(if (and (string-matches pbreak "BB?")
(item.relation.prev w 'Word))
(item.set_feat
(item.relation.prev w 'Word) "pbreak" pbreak))
(item.relation.remove w 'Word)
;; can't refer to w as we've just deleted it
(item.relation.remove wp 'Phrase)))))
words)))
utt))
(provide 'italian_module)
|
