PereLluis13 commited on
Commit
7859f42
1 Parent(s): cdfa72c

add num. verbalizer

Browse files
text/LICENSE ADDED
@@ -0,0 +1,19 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Copyright (c) 2017 Keith Ito
2
+
3
+ Permission is hereby granted, free of charge, to any person obtaining a copy
4
+ of this software and associated documentation files (the "Software"), to deal
5
+ in the Software without restriction, including without limitation the rights
6
+ to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
7
+ copies of the Software, and to permit persons to whom the Software is
8
+ furnished to do so, subject to the following conditions:
9
+
10
+ The above copyright notice and this permission notice shall be included in
11
+ all copies or substantial portions of the Software.
12
+
13
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
14
+ IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
15
+ FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
16
+ AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
17
+ LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
18
+ OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
19
+ THE SOFTWARE.
text/__init__.py ADDED
@@ -0,0 +1,74 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ """ from https://github.com/keithito/tacotron """
2
+ import re
3
+ from text import cleaners
4
+ from text.symbols import symbols
5
+
6
+
7
+ # Mappings from symbol to numeric ID and vice versa:
8
+ _symbol_to_id = {s: i for i, s in enumerate(symbols)}
9
+ _id_to_symbol = {i: s for i, s in enumerate(symbols)}
10
+
11
+ # Regular expression matching text enclosed in curly braces:
12
+ _curly_re = re.compile(r'(.*?)\{(.+?)\}(.*)')
13
+
14
+
15
+ def text_to_sequence(text, cleaner_names):
16
+ '''Converts a string of text to a sequence of IDs corresponding to the symbols in the text.
17
+
18
+ The text can optionally have ARPAbet sequences enclosed in curly braces embedded
19
+ in it. For example, "Turn left on {HH AW1 S S T AH0 N} Street."
20
+
21
+ Args:
22
+ text: string to convert to a sequence
23
+ cleaner_names: names of the cleaner functions to run the text through
24
+
25
+ Returns:
26
+ List of integers corresponding to the symbols in the text
27
+ '''
28
+ sequence = []
29
+
30
+ # Check for curly braces and treat their contents as ARPAbet:
31
+ while len(text):
32
+ m = _curly_re.match(text)
33
+ if not m:
34
+ sequence += _symbols_to_sequence(_clean_text(text, cleaner_names))
35
+ break
36
+ sequence += _symbols_to_sequence(_clean_text(m.group(1), cleaner_names))
37
+ sequence += _arpabet_to_sequence(m.group(2))
38
+ text = m.group(3)
39
+
40
+ return sequence
41
+
42
+
43
+ def sequence_to_text(sequence):
44
+ '''Converts a sequence of IDs back to a string'''
45
+ result = ''
46
+ for symbol_id in sequence:
47
+ if symbol_id in _id_to_symbol:
48
+ s = _id_to_symbol[symbol_id]
49
+ # Enclose ARPAbet back in curly braces:
50
+ if len(s) > 1 and s[0] == '@':
51
+ s = '{%s}' % s[1:]
52
+ result += s
53
+ return result.replace('}{', ' ')
54
+
55
+
56
+ def _clean_text(text, cleaner_names):
57
+ for name in cleaner_names:
58
+ cleaner = getattr(cleaners, name)
59
+ if not cleaner:
60
+ raise Exception('Unknown cleaner: %s' % name)
61
+ text = cleaner(text)
62
+ return text
63
+
64
+
65
+ def _symbols_to_sequence(symbols):
66
+ return [_symbol_to_id[s] for s in symbols if _should_keep_symbol(s)]
67
+
68
+
69
+ def _arpabet_to_sequence(text):
70
+ return _symbols_to_sequence(['@' + s for s in text.split()])
71
+
72
+
73
+ def _should_keep_symbol(s):
74
+ return s in _symbol_to_id and s is not '_' and s is not '~'
text/ca.sor ADDED
@@ -0,0 +1,485 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ^0 zero
2
+ 1$ u
3
+ 1 un
4
+ 2 dos
5
+ 3 tres
6
+ 4 quatre
7
+ 5 cinc
8
+ 6 sis
9
+ 7 set
10
+ 8 huit # [:ca-valencia:] [:ca-ES-valencia:]
11
+ 8 vuit
12
+ 9 nou
13
+ #10-19
14
+ 10 deu
15
+ 11 onze
16
+ 12 dotze
17
+ 13 tretze
18
+ 14 catorze
19
+ 15 quinze
20
+ 16 setze
21
+ 17 dèsset # [:ca-valencia:] [:ca-ES-valencia:]
22
+ 17 desset # [:ca-balear:] [:ca-ES-balear:]
23
+ 17 disset
24
+ 18 devuit # [:ca-balear:] [:ca-ES-balear:]
25
+ 18 díhuit # [:ca-valencia:] [:ca-ES-valencia:]
26
+ 19 denou # [:ca-balear:] [:ca-ES-balear:]
27
+ 19 dènou # [:ca-valencia:] [:ca-ES-valencia:]
28
+ 1(\d) di$1
29
+ # 20-29
30
+ 20 vint
31
+ 2(\d) vint-i-$1
32
+ # 30, 40, 50, 60, 70, 80, 90
33
+ 30 trenta
34
+ 40 quaranta
35
+ 50 cinquanta
36
+ 60 seixanta
37
+ 70 setanta
38
+ 80 huitanta # [:ca-valencia:] [:ca-ES-valencia:]
39
+ 80 vuitanta
40
+ 90 noranta
41
+ (\d)(\d) $(\10)-$2
42
+
43
+ #100-199
44
+ 100 cent
45
+ 1(\d\d) cent $1
46
+ #200-999
47
+ (\d)00 $1-cents
48
+ (\d)(\d\d) $1-cents $2
49
+
50
+ #1000-1999
51
+ 1000 mil
52
+ 1(\d{3}) mil $1
53
+
54
+ #2000-999999
55
+ (\d{1,3})000 $1 mil
56
+ (\d{1,3})(\d{3}) $1 mil $2
57
+
58
+ # our limit is number <10^606
59
+ (\d{606,}) ""
60
+
61
+ # x-lions
62
+ # 10000000=10^6 -> un milió
63
+ 1((0{6})+) un $(pre:$(count:\1))lió
64
+ 1((\d{6})+) un $(pre:$(count:\1))lió $1
65
+ # 2000000=2·10^6 -> dos milions
66
+ (\d{1,6})((0{6})+) $1 $(pre:$(count:\2))lions
67
+ (\d{1,6})((\d{6})+) $1 $(pre:$(count:\2))lions $2
68
+
69
+
70
+ # count number of 10^6, usefull for x-lions, and x-liards prefixes.
71
+ count:.{0,5}? 0
72
+ count:.{6}.{0,5} 1
73
+ count:(.{12}).{0,5} 2
74
+ count:(.{18}).{0,5} 3
75
+ count:(.{24}).{0,5} 4
76
+ count:(.{30}).{0,5} 5
77
+ count:(.{36}).{0,5} 6
78
+ count:(.{42}).{0,5} 7
79
+ count:(.{48}).{0,5} 8
80
+ count:(.{54}).{0,5} 9
81
+ count:(.{60})(.{0,59}) 1|$(count:\2)
82
+ count:(.{120})(.{0,59}) 2|$(count:\2)
83
+ count:(.{180})(.{0,59}) 3|$(count:\2)
84
+ count:(.{240})(.{0,59}) 4|$(count:\2)
85
+ count:(.{300})(.{0,59}) 5|$(count:\2)
86
+ count:(.{360})(.{0,59}) 6|$(count:\2)
87
+ count:(.{420})(.{0,59}) 7|$(count:\2)
88
+ count:(.{480})(.{0,59}) 8|$(count:\2)
89
+ count:(.{540})(.{0,59}) 9|$(count:\2)
90
+ count:(.{600})(.{0,5}) 10|$(count:\2) # our limit is 10^606-1
91
+
92
+ # prefixes needed for x-lions and x-liards, up to 10^606-1
93
+ pre:1 mi
94
+ pre:2 bi
95
+ pre:3 tri
96
+ pre:4 quadri
97
+ pre:5 quinti
98
+ pre:6 sexti
99
+ pre:7 septi
100
+ pre:8 octi
101
+ pre:9 noni
102
+ pre:10 deci
103
+ pre:1(\d) $(pre2:\1)|deci
104
+ pre:(\d)0 $(pre3:\1)
105
+ pre:(\d)(\d) $(pre2:\2)|$(pre3:\1)
106
+ pre:100 centi
107
+
108
+ pre2:1 uno
109
+ pre2:2 duo
110
+ pre2:3 tre
111
+ pre2:4 quattour
112
+ pre2:5 quin
113
+ pre2:6 sex
114
+ pre2:7 septen
115
+ pre2:8 octo
116
+ pre2:9 novem
117
+
118
+ pre3:1 deci
119
+ pre3:2 viginti
120
+ pre3:3 triginti
121
+ pre3:4 quadraginti
122
+ pre3:5 quinquaginti
123
+ pre3:6 sexaginti
124
+ pre3:7 septuaginti
125
+ pre3:8 octoginti
126
+ pre3:9 nonoginti
127
+ pre3:10 centi
128
+
129
+ # negative number
130
+ [--](\d+) menys |$1
131
+
132
+ # decimals
133
+ "([^,]*\d)[.]((\d{3})+)([,][^,.]*)?" $(\1\2\4)
134
+ "([--]?\d+)([,]0*)?" $1
135
+ "([--]?\d+)[,](\d*)" $(\1·\2)
136
+ "([--]?\d+·0*)([^0]00?)0*" $1| |$2
137
+ "([--]?\d+·0*)([^0])" $1| |$2
138
+ "([--]?\d+·0*)([^0]\d)" $1| |$2
139
+ "([--]?\d+·0*)([^0]\d\d)" $1| |$2
140
+ "([--]?\d+·0*)([^0]\d\d)0*" $1| |$2
141
+
142
+ "([--]?\d+·0*)(([^0]|[^0]\d*[^0]))0*" $1| $(read:\2)
143
+ "([--]?\d+)·(\d*)(\d)" $(\1·\2)| |$3
144
+ "([--]?\d+)·" $1| coma
145
+
146
+ # used for decimal part
147
+ #read:(\d*[^0])0*$ $(read:\1)
148
+ read:(\d*[1-9])(00+)([1-9]\d*) $(read:\1)| |$(read:\2) |$(read:\3)
149
+ read:(\d$) $1
150
+ read:0(\d+) $(read:0)| |$(read:\1)
151
+ read:([1-9]\d) $1
152
+ read:([1-9]\d\d) $1
153
+ read:(\d\d\d) $1
154
+ read:(\d\d)((\d\d)+) $(read:\1)| |$(read:\2)
155
+ read:(\d\d)((\d\d)*)(\d\d\d) $(read:\1)| |$(read:\2)| |$(read:\4)
156
+
157
+
158
+ # convert masculine forms to feminine forms
159
+ # it can be run after: standard number conversion; and after ordinal, partitive functions.
160
+ ## runned with feminine function.
161
+ f:(.*iliard)(.*) \1$(f:\2) # convert only <1,000,000,000
162
+ f:(.*ili)(.*) \1$(f:\2) # convert only <100,0000
163
+ f:(.*d)o(s[^èé]*) $(f:\1ue\2) # 2 -> dos -> dues
164
+ f:(.*cent)(s.*) $(f:\1e\2) # cents -> centes
165
+ f:(((.*)[^a-zèé]|))u$ \1una # vint-i-u -> vint-i-una
166
+ ## runned after ord function.
167
+ f:(.*[^0-9])n$ \1na # segon -> segona
168
+ f:(.*[^0-9]r)$ \1a # tercer -> tercera
169
+ f:(.*[^0-9]r)t$ \1ta # quart -> quarta
170
+ f:(.*[^0-9])è$ \1ena # sisè -> sisena
171
+ f:(.*[^0-9])é$ \1ena # sisé -> sisena
172
+ ## runned after ord2 function.
173
+ f:(.*[0-9])[nrtè]$ \1a # 2n -> 2a
174
+ ## runnded after part function.
175
+ f:(.*ter)ç$ \1cera # terç -> tercera
176
+ f:(.*è[sc]i)m$ \1ma # milionèsim -> milionèsima
177
+ f:(.*[^0-9]i)g$ \1tja # mig -> mitja
178
+
179
+
180
+ no-centes:(.*)centes(.*) \1cents\2
181
+ no-centes:(.*) \1
182
+
183
+ # convert ordinal numbers (1st, 2nd, 3rd,... nth) to partitive (1, 1/2, 1/3, .... 1/n)
184
+ p:(.*)primer$ \1unitat
185
+ p:(.*)segon$ \1mig
186
+ p:(.*)tercer$ \1terç
187
+ p:(.*quart)$ \1
188
+ p:(.*)des[èé]$ \1dècim
189
+ p:((.*)cent)[èé]$ \1èsim
190
+ p:((.*)mil)[èé]$ \1·lèsim
191
+ p:((.*)ilion)[èé]$ \1èsim
192
+ p:((.*)iliard)[èé]$ \1èsim
193
+
194
+
195
+ # fallback, ignore 1-letter not-defined fuctions
196
+ .:(.*) \1
197
+
198
+ # runned after ordinal and partitive fuctions
199
+ pl:(.*[^\d][nrtnec])$ \1s
200
+ pl:(.*[^\d])ig$ \1igs # mig -> mitjos
201
+ pl:(.*[^\d])ja$ \1ges
202
+ pl:(.*[^\d])a$ \1es
203
+ pl:(.*[^\d])[èé]$ \1ens
204
+ # after ord2: 1r->1rs, 2n->2ns, 5è->5ns, ...
205
+ pl:(\d+[rnrt])$ \1s # 1r -> 1rs, 2n -> 2ns, 4t -> 4ts
206
+ pl:(\d+)[èé]$ \1ns # 5è -> 5ns
207
+ pl:(\d+)a$ \1es # 2a -> 2es
208
+ # after partitive
209
+ pl:([^[0-9]*[sç])$ \1os # dos -> dosos, terç > terços
210
+ pl:([^[0-9]*è[sc]im)$ \1s # dècim -> dècims
211
+ #fallback
212
+ pl:(.*) \1
213
+
214
+
215
+ # unit/subunit singular/plural
216
+ # million or greater part of the number name separated by "ili" pattern
217
+ # before masculine to feminine conversion
218
+ us(.).:([^,]*),([^,]*),([^,]*),([^,]*),([^,]*),([^,]*) $(\1:\7)| \2
219
+ up(.).:([^,]*),([^,]*),([^,]*),([^,]*),([^,]*),([^,]*) $(\1:\7)| \3
220
+ ud(.).:([^,]*),([^,]*),([^,]*),([^,]*),([^,]*),([^,]*) $(\1:\7)| \4
221
+ ss.(.):([^,]*),([^,]*),([^,]*),([^,]*),([^,]*),([^,]*) $(\1:\7)| \5
222
+ sp.(.):([^,]*),([^,]*),([^,]*),([^,]*),([^,]*),([^,]*) $(\1:\7)| \6
223
+
224
+ # "mm" means masculine unit and masculine subunit
225
+ # Usually used by Catalan users
226
+ CHF:(.+),(.+) $(\2mm: franc suís, francs suïssos, de francs suïssos, cèntim, cèntims, \1)
227
+ EUR:(.+),(.+) $(\2mm: euro, euros, d'euros, cèntim, cèntims, \1)
228
+ GBP:(.+),(.+) $(\2fm: lliura esterlina, lliures esterlines, de lliures esterlines, penic, penics, \1)
229
+ JPY:(.+),(.+) $(\2mm: ien, iens, de iens, sen, sen, \1)
230
+ USD:(.+),(.+) $(\2mm: dòlar dels EUA, dòlars dels EUA, de dòlars dels EUA, centau, centaus, \1)
231
+ # ACTIVE ISO 4217 CODES--A--
232
+ AED:(.+),(.+) $(\2mm: dírham dels Emirats Àrabs Units, dírhams dels Emirats Àrabs Units, de dírhams dels Emirats Àrabs Units, fils, fulús, \1)
233
+ AFN:(.+),(.+) $(\2mm: afgani, afganis, d'afganis, puli, puli, \1)
234
+ ALL:(.+),(.+) $(\2mm: lek, lekë, de lekë, qindarka, qindarka, \1)
235
+ AMD:(.+),(.+) $(\2mm: dram, drams, de drams, luma, luma, \1)
236
+ ANG:(.+),(.+) $(\2mm: florí de les Antilles Neerlandeses, florins de les Antilles Neerlandeses, de florins de les Antilles Neerlandeses, cèntim, cèntims, \1)
237
+ AOA:(.+),(.+) $(\2fm: kwanza, kwanzes, de kwanzes, cèntim cèntims, \1)
238
+ ARS:(.+),(.+) $(\2mm: peso argentí, pesos argentins, de pesos argentins, centau, centaus, \1)
239
+ AUD(.+),(.+) $(\2mm: dòlar australià, dòlars australians, de dòlars australians, centau, centaus, \1)
240
+ AWG:(.+),(.+) $(\2mm: florí d'Aruba, florins d'Aruba, de florins d'Aruba, cèntim, cèntims, \1)
241
+ AZN:(.+),(.+) $(\2mm: manat azerbaidjanès, manats azerbaidjanesos, de manats azerbaidjanesos, qəpik, qəpik, \1)
242
+ # ACTIVE ISO 4217 CODES --X--
243
+ #XAF Franc CFA emès pel BEAC (Banc dels Estats de l'Àfrica Central)
244
+ XAG:(.+),(.+) $(\2fm: unça de plata, unces de plata, d'unces de plata, cèntim, cèntims, \1)
245
+ XAU:(.+),(.+) $(\2fm: unça d'or, unces d'or, d'unces d'or, cèntim, cèntims, \1)
246
+ #XBA Unitat compensatòria europea (EURCO) (unitat per al mercat d'obligacions)
247
+ #XBB Unitat monetària europea (EMU-6) (unitat per al mercat d'obligacions)
248
+ #XBC Unitat de compte europea 9 (EUA-9) (unitat per al mercat d'obligacions)
249
+ #XBD Unitat de compte europea 17 (EUA-17) (unitat per al mercat d'obligacions)
250
+ #XCD Dòlar del Carib Oriental
251
+ #XDR Drets especials de gir (del Fons Monetari Internacional)
252
+ #XFU Franc UIC (divisa especial)
253
+ #XOF Franc CFA emès pel BCEAO (Banc Central dels Estats de l'Àfrica Occidental)
254
+ XPD:(.+),(.+) $(\2fm: unça de pal·ladi, unces de pal·ladi, d'unces de pal·ladi, cèntim, cèntims, \1)
255
+ #XPF Franc CFP (per als territoris francesos del Pacífic)
256
+ XPT:(.+),(.+) $(\2fm: unça de platí, unces de platí, d'unces de platí, cèntim, cèntims, \1)
257
+ #XTS Codi reservat per a proves
258
+ #XXX Sense moneda, sense transacció monetària
259
+ # OBSOLETE ISO 4217 CODES --Replaced by EUR--
260
+ ADF:(.+),(.+) $(\2mm: franc andorrà, francs andorrans, de francs andorrans, cèntim, cèntims, \1)
261
+ ADP:(.+),(.+) $(\2fm: pesseta andorrana, pessetes andorranes, de pessetes andorranes, cèntim, cèntims, \1)
262
+ ATS:(.+),(.+) $(\2mm: xíling austríac, xílings austríacs, de xílings austríacs, groschen, groschen, \1)
263
+ BEF:(.+),(.+) $(\2mm: franc belga, francs belgues, de francs belgues, cèntim, cèntims, \1)
264
+ CYP:(.+),(.+) $(\2mm: lliura xipriota, lliures xipriotes, de lliures xipriotes, cèntim, cèntims, \1)
265
+ DEM:(.+),(.+) $(\2mm: marc alemany, marcs alemanys, de marcs alemanys, penic, penics, \1)
266
+ ESP:(.+),(.+) $(\2fm: pesseta, pessetes, de pessetes, cèntim, cèntims, \1)
267
+ FIM:(.+),(.+) $(\2mm: marc finlandès, marcs finlandesos, de marcs finlandesos, penic, penics, \1)
268
+ FRF:(.+),(.+) $(\2mm: franc francès, francs francesos, de francs francesos, cèntim, cèntims, \1)
269
+ GRD:(.+),(.+) $(\2fm: dracma grega, dracmes gregues, leptó, leptà, \1)
270
+ IEP:(.+),(.+) $(\2fm: lliura irlandesa, lliures irlandeses, de lliures irlandeses, penic, penics, \1)
271
+ ITL:(.+),(.+) $(\2fm: lira italiana, lires italianes, de lires italianes, cèntim, cèntims, \1)
272
+ LUF:(.+),(.+) $(\2mm: franc luxemburguès, francs luxemburguesos, de francs luxemburguesos, cèntim, cèntims, \1)
273
+ MCF:(.+),(.+) $(\2mm: franc monegasc, francs monegascs, de francs monegascs, cèntim, cèntims, \1)
274
+ MTL:(.+),(.+) $(\2fm: lira maltesa, lires malteses, de lires malteses, cèntim, cèntims, \1)
275
+ NLG:(.+),(.+) $(\2mm: florí neerlandès, florins neerlandesos, de florins neerlandesos, cèntim, cèntims, \1)
276
+ PTE:(.+),(.+) $(\2mm: escut portuguès, escuts portuguesos, de escuts portuguesos, centau, centaus, \1)
277
+ SIT:(.+),(.+) $(\2mm: tolar eslovè, tolars eslovens, de tolars eslovens, stotin, stotinov, \1)
278
+ SKK:(.+),(.+) $(\2fm: corona eslovaca, corones eslovaques, de corones eslovaques, halier, halierov, \1)
279
+ SML:(.+),(.+) $(\2fm: lira de San Marino, lires de San Marino, de lires de San Marino, cèntim, cèntims, \1)
280
+ VAL:(.+),(.+) $(\2fm: lira vaticana, lires vaticanes, de lires vaticanes, cèntim, cèntims, \1)
281
+ XEU:(.+),(.+) $(\2mm: ecu, ecus, d'ecus, cèntim, cèntims, \1)
282
+
283
+ #crypto-currencies
284
+ XMR:(.+),(.+) $(\2mm: monero, moneros, de moneros, piconero, piconeros, \1) #TODO: 1,000,000,000,000 piconeros = 1 monero
285
+ XBT:(.+),(.+) $(\2mm: bitcoin, bitcoins, de bitcoins, satoshi, satoshis, \1) # TODO: 10,000,000 satoshis = 1,000 millibitcoin = 1 bitcoin
286
+
287
+ # unknow currency
288
+ [A-Z]{3}:.* ""
289
+
290
+
291
+ "([A-Z]{3}) ([-−]?1)([.,]00?)?"$(\1:|$2,us)
292
+ "([A-Z]{3}) ([-−]?\d+0{6,})([.,]00?)?"$(\1:|$2,ud)
293
+ "([A-Z]{3}) ([-−]?\d+)([.,]00?)?"$(\1:|$2,up)
294
+ "(([A-Z]{3}) [-−]?\d+)[.,](01)" $1 amb$(\2:un,ss)
295
+ "(([A-Z]{3}) [-−]?\d+)[.,](\d)" $1 amb$(\2:|$(\30),sp)
296
+ "(([A-Z]{3}) [-−]?\d+)[.,](\d\d)" $1 amb$(\2:|$3,sp)
297
+
298
+
299
+ # detects number followed by currency code
300
+ "([-−]?\d+)([.,]\d+)? ([A-Z]{3})" $(\3 \1\2)
301
+
302
+
303
+ # currency symbols
304
+ "€[ ]?([^ ]*)" $(EUR \1)
305
+ "£[ ]?([^ ]*)" $(GBP \1)
306
+ "\$[ ]?([^ ]*)" $(USD \1)
307
+ "¥[ ]?([^ ]*)" $(JPY \1)
308
+ "₩[ ]?([^ ]*)" $(KRW \1)
309
+ "₽[ ]?([^ ]*)" $(RUB \1)
310
+ "ɱ[ ]?([^ ]*)" $(XMR \1)
311
+ "₿[ ]?([^ ]*)" $(XBT \1)
312
+
313
+ "([^ ]+)[ ]?€$" $(EUR \1)
314
+ "([^ ]+)[ ]?£$" $(GBP \1)
315
+ "([^ ]+)[ ]?\$$" $(USD \1)
316
+ "([^ ]+)[ ]?¥$" $(JPY \1)
317
+ "([^ ]+)[ ]?₩$" $(KRW \1)
318
+ "([^ ]+)[ ]?₽$" $(RUB \1)
319
+ "([^ ]+)[ ]?ɱ$" $(XMR \1)
320
+ "([^ ]+)[ ]?₿$" $(XBT \1)
321
+
322
+ == feminine ==
323
+
324
+ 1 una
325
+ (.*) $(f:|$1)
326
+
327
+ == masculine ==
328
+
329
+ 1 un
330
+ (.*) $1
331
+
332
+ == ordinal(-masculine)? ==
333
+
334
+ ([-−]\d+) ""
335
+ \d+[,.] ""
336
+ 0 zeroé # [:ca-valencia:] [:ca-ES-valencia:]
337
+ 0 zeroè
338
+ 1 primer
339
+ 2 segon
340
+ 3 tercer
341
+ 4 quart
342
+ (\d+)$ $(ordinal $2)
343
+ "un ([^ ]*(ilió|iliard))$" $(ordinal \2)
344
+ (.*li)ó$ \2oné # [:ca-valencia:] [:ca-ES-valencia:]
345
+ (.*li)ó$ \2onè
346
+ (.*(cent|mil|ion|iliard))s?$ \2é # [:ca-valencia:] [:ca-ES-valencia:]
347
+ (.*(cent|mil|ion|iliard))s?$ \2è
348
+ "(.* )u$" \2uné # [:ca-valencia:] [:ca-ES-valencia:]
349
+ "(.* )u$" \2unè
350
+ (.*-)u$ \2uné # [:ca-valencia:] [:ca-ES-valencia:]
351
+ (.*-)u$ \2unè
352
+ "u" primer
353
+ "un" primer
354
+ "dos" segon
355
+ "tres" terç
356
+ "quatre" quart
357
+ (.*)cinc$ \2cinqué # [:ca-valencia:] [:ca-ES-valencia:]
358
+ (.*)cinc$ \2cinquè
359
+ (.*)dènou$ \2denové # [:ca-valencia:] [:ca-ES-valencia:]
360
+ (.*)nou$ \2nové # [:ca-valencia:] [:ca-ES-valencia:]
361
+ (.*)nou$ \2novè
362
+ (.*)deu$ \2desé # [:ca-valencia:] [:ca-ES-valencia:]
363
+ (.*)deu$ \2desè
364
+ (.*)dèsset$ \2desseté # [:ca-valencia:] [:ca-ES-valencia:]
365
+ (.*)díhuit$ \2dihuité # [:ca-valencia:] [:ca-ES-valencia:]
366
+ (.*)[ae]$ \2é # [:ca-valencia:] [:ca-ES-valencia:]
367
+ (.*)[ae]$ \2è
368
+ (.*\D)$ \2é # [:ca-valencia:] [:ca-ES-valencia:]
369
+ (.*\D)$ \2è
370
+
371
+ == ordinal-feminine ==
372
+ ([-−]\d+) ""
373
+ \d+[,.] ""
374
+ (\d+)$ $(no-centes:$(f:$(ordinal \1)))
375
+
376
+ == ordinal-masculine-plural ==
377
+
378
+ ([-−]?\d+) $(ordinal-masculine-plural $(ordinal \1))
379
+ primer primers
380
+ segon segons
381
+ (.*)è \1ens
382
+ (.*)er \1ers
383
+
384
+ == ordinal-feminine-plural ==
385
+
386
+ ([-−]?\d+) $(ordinal-feminine-plural $(ordinal-feminine \1))
387
+ (.*)a \1es
388
+
389
+ == ordinal-number(-masculine)? ==
390
+
391
+ #(\d+) $(o:\2)
392
+ 1$ 1r
393
+ 2$ 2n
394
+ 3$ 3r
395
+ 4$ 4t
396
+ (\d+)$ \2é # [:ca-valencia:] [:ca-ES-valencia:]
397
+ (\d+)$ \2è
398
+
399
+ == ordinal-number-feminine ==
400
+ (\d+)$ \1a
401
+
402
+ == partitive(-masculine)? ==
403
+ ([--]?\d+) $(p:$(ordinal \2))
404
+
405
+ == partitive-feminine ==
406
+ ([--]?\d+) $(no-centes:$(f:$(p:$(ordinal \1))))
407
+
408
+
409
+ == partitive(-masculine)?-plural ==
410
+ ([--]?\d+) $(pl:$(p:$(ordinal $2)))
411
+
412
+ == partitive-feminine-plural ==
413
+ ([--]?\d+) $(no-centes:$(pl:$(f:$(p:$(ordinal $1)))))
414
+
415
+ == fraction(-masculine)? ==
416
+ ([--]?1)(/1)? $2
417
+ ([--]?1)/2 mig
418
+ ([--]?1)/([3-9]\d*) $(masculine \2)| $(partitive \3)
419
+ ([--]?\d+)(/1)? $2
420
+ ([--]?\d+)/([1-9]\d*) $2| $(partitive-plural \3)
421
+
422
+ == fraction-feminine ==
423
+ ([--]?1)(/1)? $(f:$1)| unitat
424
+ ([--]?1)/([1-9]\d*) $(f:$1)| $(partitive-feminine \2)| part
425
+ ([--]?\d+)(/1)? $(f:$1)| unitats
426
+ ([--]?\d+)/([1-9]\d*) $(f:$1)| $(partitive-feminine-plural \2)| parts
427
+
428
+ == collective ==
429
+ 2 parell, parella o duo
430
+ 3 tern, terna, tercet, trio, tríada o treset
431
+ 4 qüern, tètrada, quartet, quarteta o quàdruple
432
+ 5 quintern, quintet, cinquet o quíntuple
433
+ 6 sextet, siset o sèxtuple
434
+ 7 septet, setet o sèptuple
435
+ 8 octet o òctuple
436
+ 9 nònuple
437
+ 10 dècada o dècuple
438
+ 12 dotzena
439
+ 100 centenar
440
+ 144 grossa
441
+ 1000 miler
442
+ 10000 miríada
443
+
444
+ == years ==
445
+ 2 bienni
446
+ 3 trienni
447
+ 4 quadrienni
448
+ 5 quinquenni o lustre
449
+ 6 sesenni
450
+ 7 septenni
451
+ 10 dècada o decenni
452
+ 12 duodecenni
453
+ 15 quindecenni
454
+ 20 vintenni o vicenni
455
+ 30 trentenni o tricenni
456
+ 40 quarantenni
457
+ 50 cinquantenni
458
+ 60 seixantenni
459
+ 70 setantenni
460
+ 80 huitantenni [:ca-valencia:] [:ca-ES-valencia:]
461
+ 80 vuitantenni
462
+ 90 norantenni
463
+ 100 segle o centenni
464
+ 1000 mil·lenni
465
+
466
+ == multiplicative ==
467
+ 2 doble o duple
468
+ 3 triple
469
+ 4 quàdruple
470
+ 5 quíntuple
471
+ 6 sèxtuple
472
+ 7 sèptuple
473
+ 8 òctuple
474
+ 9 nònuple
475
+ 10 dècuple
476
+ 12 duodècuple
477
+ 100 cèntuple
478
+ 1/10 subdècuple
479
+ 1/2 súbduple
480
+
481
+ == help ==
482
+
483
+ "" $(1)|, $(2), $(3)\n$(help feminine)$(help masculine)$(help ordinal-number-masculine)$(help ordinal-number-feminine)$(help ordinal-feminine)$(help ordinal-masculine)
484
+ (feminine|masculine|ordinal(-number)?(-feminine|-masculine)?) \1: $(\1 1), $(\1 2), $(\1 3)\n
485
+
text/cleaners.py ADDED
@@ -0,0 +1,150 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ """ from https://github.com/keithito/tacotron """
2
+
3
+ '''
4
+ Cleaners are transformations that run over the input text at both training and eval time.
5
+
6
+ Cleaners can be selected by passing a comma-delimited list of cleaner names as the "cleaners"
7
+ hyperparameter. Some cleaners are English-specific. You'll typically want to use:
8
+ 1. "english_cleaners" for English text
9
+ 2. "transliteration_cleaners" for non-English text that can be transliterated to ASCII using
10
+ the Unidecode library (https://pypi.python.org/pypi/Unidecode)
11
+ 3. "basic_cleaners" if you do not want to transliterate (in this case, you should also update
12
+ the symbols in symbols.py to match your data).
13
+ '''
14
+
15
+ import re
16
+ from unidecode import unidecode
17
+ from text.numbers import normalize_numbers
18
+ from text.numbers_ca import normalize_numbers_ca
19
+ from text.symbols import symbols
20
+
21
+ # Regular expression matching whitespace:
22
+ _whitespace_re = re.compile(r'\s+')
23
+
24
+ # List of (regular expression, replacement) pairs for abbreviations:
25
+ _abbreviations_en = [(re.compile('\\b%s\\.' % x[0], re.IGNORECASE), x[1]) for x in [
26
+ ('mrs', 'misess'),
27
+ ('mr', 'mister'),
28
+ ('dr', 'doctor'),
29
+ ('st', 'saint'),
30
+ ('co', 'company'),
31
+ ('jr', 'junior'),
32
+ ('maj', 'major'),
33
+ ('gen', 'general'),
34
+ ('drs', 'doctors'),
35
+ ('rev', 'reverend'),
36
+ ('lt', 'lieutenant'),
37
+ ('hon', 'honorable'),
38
+ ('sgt', 'sergeant'),
39
+ ('capt', 'captain'),
40
+ ('esq', 'esquire'),
41
+ ('ltd', 'limited'),
42
+ ('col', 'colonel'),
43
+ ('ft', 'fort'),
44
+ ]]
45
+
46
+ # List of (regular expression, replacement) pairs for catalan abbreviations:
47
+ _abbreviations_ca = [(re.compile('\\b%s\\b' % x[0], re.IGNORECASE), x[1]) for x in [
48
+ ('tv3', 't v tres'),
49
+ ('8tv', 'vuit t v'),
50
+ ('pp', 'p p'),
51
+ ('psoe', 'p soe'),
52
+ ('sr.?', 'senyor'),
53
+ ('sra.?', 'senyora'),
54
+ ('srta.?', 'senyoreta')
55
+ ]]
56
+
57
+ _replacements_ca = [(re.compile('%s' % x[0], re.IGNORECASE), x[1]) for x in [
58
+ (';', ','),
59
+ (':', '\.'),
60
+ ('\.\.\.,', ','),
61
+ ('\.\.\.', '…'),
62
+ ('ñ','ny')
63
+ ]]
64
+
65
+
66
+ def expand_abbreviations(text, lang='ca'):
67
+ if lang == 'en':
68
+ _abbreviations = _abbreviations_en
69
+ elif lang == 'ca':
70
+ _abbreviations = _abbreviations_ca
71
+ else:
72
+ raise ValueError('no %s language for abbreviations'%lang)
73
+ for regex, replacement in _abbreviations:
74
+ text = re.sub(regex, replacement, text)
75
+ return text
76
+
77
+
78
+ def convert_characters(text, lang='ca'):
79
+ if lang == 'ca':
80
+ _replacements = _replacements_ca
81
+ else:
82
+ raise ValueError('no %s language for punctuation conversion'%lang)
83
+ for regex, replacement in _replacements_ca:
84
+ text = re.sub(regex, replacement, text)
85
+ return text
86
+
87
+
88
+ def expand_numbers(text, lang="ca"):
89
+ if lang == 'ca':
90
+ return normalize_numbers_ca(text)
91
+ else:
92
+ return normalize_numbers(text)
93
+
94
+
95
+ def lowercase(text):
96
+ return text.lower()
97
+
98
+
99
+ def collapse_whitespace(text):
100
+ return re.sub(_whitespace_re, ' ', text)
101
+
102
+
103
+ def convert_to_ascii(text, lang="ca"):
104
+ if lang == 'en':
105
+ return unidecode(text)
106
+ elif lang == 'ca':
107
+ char_replace = []
108
+ for t in set(list(text)):
109
+ if t not in symbols:
110
+ char_replace.append([t, unidecode(t)])
111
+ for target, replace in char_replace:
112
+ text = text.replace(target, replace)
113
+ return text
114
+ else:
115
+ raise ValueError('no %s language for punctuation conversion'%lang)
116
+
117
+
118
+ def basic_cleaners(text):
119
+ '''Basic pipeline that lowercases and collapses whitespace without transliteration.'''
120
+ text = lowercase(text)
121
+ text = collapse_whitespace(text)
122
+ return text
123
+
124
+
125
+ def transliteration_cleaners(text):
126
+ '''Pipeline for non-English text that transliterates to ASCII.'''
127
+ text = convert_to_ascii(text)
128
+ text = lowercase(text)
129
+ text = collapse_whitespace(text)
130
+ return text
131
+
132
+
133
+ def english_cleaners(text):
134
+ '''Pipeline for English text, including number and abbreviation expansion.'''
135
+ text = convert_to_ascii(text)
136
+ text = lowercase(text)
137
+ text = expand_numbers(text, lang='en')
138
+ text = expand_abbreviations(text, lang='en')
139
+ text = collapse_whitespace(text)
140
+ return text
141
+
142
+
143
+ def catalan_cleaners(text):
144
+ text = lowercase(text)
145
+ text = expand_numbers(text, lang="ca")
146
+ text = convert_characters(text, lang="ca")
147
+ text = convert_to_ascii(text, lang="ca")
148
+ text = expand_abbreviations(text, lang="ca")
149
+ text = collapse_whitespace(text)
150
+ return text
text/cmudict.py ADDED
@@ -0,0 +1,65 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ """ from https://github.com/keithito/tacotron """
2
+
3
+ import re
4
+
5
+
6
+ valid_symbols = [
7
+ 'AA', 'AA0', 'AA1', 'AA2', 'AE', 'AE0', 'AE1', 'AE2', 'AH', 'AH0', 'AH1', 'AH2',
8
+ 'AO', 'AO0', 'AO1', 'AO2', 'AW', 'AW0', 'AW1', 'AW2', 'AY', 'AY0', 'AY1', 'AY2',
9
+ 'B', 'CH', 'D', 'DH', 'EH', 'EH0', 'EH1', 'EH2', 'ER', 'ER0', 'ER1', 'ER2', 'EY',
10
+ 'EY0', 'EY1', 'EY2', 'F', 'G', 'HH', 'IH', 'IH0', 'IH1', 'IH2', 'IY', 'IY0', 'IY1',
11
+ 'IY2', 'JH', 'K', 'L', 'M', 'N', 'NG', 'OW', 'OW0', 'OW1', 'OW2', 'OY', 'OY0',
12
+ 'OY1', 'OY2', 'P', 'R', 'S', 'SH', 'T', 'TH', 'UH', 'UH0', 'UH1', 'UH2', 'UW',
13
+ 'UW0', 'UW1', 'UW2', 'V', 'W', 'Y', 'Z', 'ZH'
14
+ ]
15
+
16
+ _valid_symbol_set = set(valid_symbols)
17
+
18
+
19
+ class CMUDict:
20
+ '''Thin wrapper around CMUDict data. http://www.speech.cs.cmu.edu/cgi-bin/cmudict'''
21
+ def __init__(self, file_or_path, keep_ambiguous=True):
22
+ if isinstance(file_or_path, str):
23
+ with open(file_or_path, encoding='latin-1') as f:
24
+ entries = _parse_cmudict(f)
25
+ else:
26
+ entries = _parse_cmudict(file_or_path)
27
+ if not keep_ambiguous:
28
+ entries = {word: pron for word, pron in entries.items() if len(pron) == 1}
29
+ self._entries = entries
30
+
31
+
32
+ def __len__(self):
33
+ return len(self._entries)
34
+
35
+
36
+ def lookup(self, word):
37
+ '''Returns list of ARPAbet pronunciations of the given word.'''
38
+ return self._entries.get(word.upper())
39
+
40
+
41
+
42
+ _alt_re = re.compile(r'\([0-9]+\)')
43
+
44
+
45
+ def _parse_cmudict(file):
46
+ cmudict = {}
47
+ for line in file:
48
+ if len(line) and (line[0] >= 'A' and line[0] <= 'Z' or line[0] == "'"):
49
+ parts = line.split(' ')
50
+ word = re.sub(_alt_re, '', parts[0])
51
+ pronunciation = _get_pronunciation(parts[1])
52
+ if pronunciation:
53
+ if word in cmudict:
54
+ cmudict[word].append(pronunciation)
55
+ else:
56
+ cmudict[word] = [pronunciation]
57
+ return cmudict
58
+
59
+
60
+ def _get_pronunciation(s):
61
+ parts = s.strip().split(' ')
62
+ for part in parts:
63
+ if part not in _valid_symbol_set:
64
+ return None
65
+ return ' '.join(parts)
text/numbers.py ADDED
@@ -0,0 +1,71 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ """ from https://github.com/keithito/tacotron """
2
+
3
+ import inflect
4
+ import re
5
+
6
+
7
+ _inflect = inflect.engine()
8
+ _comma_number_re = re.compile(r'([0-9][0-9\,]+[0-9])')
9
+ _decimal_number_re = re.compile(r'([0-9]+\.[0-9]+)')
10
+ _pounds_re = re.compile(r'£([0-9\,]*[0-9]+)')
11
+ _dollars_re = re.compile(r'\$([0-9\.\,]*[0-9]+)')
12
+ _ordinal_re = re.compile(r'[0-9]+(st|nd|rd|th)')
13
+ _number_re = re.compile(r'[0-9]+')
14
+
15
+
16
+ def _remove_commas(m):
17
+ return m.group(1).replace(',', '')
18
+
19
+
20
+ def _expand_decimal_point(m):
21
+ return m.group(1).replace('.', ' point ')
22
+
23
+
24
+ def _expand_dollars(m):
25
+ match = m.group(1)
26
+ parts = match.split('.')
27
+ if len(parts) > 2:
28
+ return match + ' dollars' # Unexpected format
29
+ dollars = int(parts[0]) if parts[0] else 0
30
+ cents = int(parts[1]) if len(parts) > 1 and parts[1] else 0
31
+ if dollars and cents:
32
+ dollar_unit = 'dollar' if dollars == 1 else 'dollars'
33
+ cent_unit = 'cent' if cents == 1 else 'cents'
34
+ return '%s %s, %s %s' % (dollars, dollar_unit, cents, cent_unit)
35
+ elif dollars:
36
+ dollar_unit = 'dollar' if dollars == 1 else 'dollars'
37
+ return '%s %s' % (dollars, dollar_unit)
38
+ elif cents:
39
+ cent_unit = 'cent' if cents == 1 else 'cents'
40
+ return '%s %s' % (cents, cent_unit)
41
+ else:
42
+ return 'zero dollars'
43
+
44
+
45
+ def _expand_ordinal(m):
46
+ return _inflect.number_to_words(m.group(0))
47
+
48
+
49
+ def _expand_number(m):
50
+ num = int(m.group(0))
51
+ if num > 1000 and num < 3000:
52
+ if num == 2000:
53
+ return 'two thousand'
54
+ elif num > 2000 and num < 2010:
55
+ return 'two thousand ' + _inflect.number_to_words(num % 100)
56
+ elif num % 100 == 0:
57
+ return _inflect.number_to_words(num // 100) + ' hundred'
58
+ else:
59
+ return _inflect.number_to_words(num, andword='', zero='oh', group=2).replace(', ', ' ')
60
+ else:
61
+ return _inflect.number_to_words(num, andword='')
62
+
63
+
64
+ def normalize_numbers(text):
65
+ text = re.sub(_comma_number_re, _remove_commas, text)
66
+ text = re.sub(_pounds_re, r'\1 pounds', text)
67
+ text = re.sub(_dollars_re, _expand_dollars, text)
68
+ text = re.sub(_decimal_number_re, _expand_decimal_point, text)
69
+ text = re.sub(_ordinal_re, _expand_ordinal, text)
70
+ text = re.sub(_number_re, _expand_number, text)
71
+ return text
text/numbers_ca.py ADDED
@@ -0,0 +1,54 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import re
2
+ import io
3
+ import pathlib
4
+ from text.soros import compile
5
+
6
+ filepath = pathlib.Path(__file__).parent.absolute()
7
+ with io.open(f"{filepath}/ca.sor", 'r', encoding="utf-8") as prg:
8
+ num2text = compile(prg.read(), 'ca')
9
+
10
+ _separador_milers_re = re.compile(r'([0-9][0-9\.]+[0-9]{3})')
11
+ _decimal_re = re.compile(r'([0-9]+\,[0-9]+)')
12
+ _ordinal_ms_re = re.compile(r'([0-9]+)(r|er|n|on|t|rt|è|e|ne|nè)+(\b)')
13
+ _ordinal_mp_re = re.compile(r'([0-9]+)(rs|ns|ts|ns)+(\b)')
14
+ _ordinal_fs_re = re.compile(r'([0-9]+)(a|ra|na|ta)+(\b)')
15
+ _ordinal_fp_re = re.compile(r'([0-9]+)(es)+(\b)')
16
+ _cardinal_re = re.compile(r'[0-9]+')
17
+ _fraccions_re = re.compile(r'(\b)([0-9]+\/[0-9]+)(\b)')
18
+ _hores_re = re.compile(r'(\b)([0-9]{1,2}):([0-9]{2})(\b)')
19
+
20
+ def _esborra_separador_milers(m):
21
+ return m.group(1).replace('.', '')
22
+
23
+ def _num2text(m):
24
+ return num2text.run(m.group(0))
25
+
26
+ def _ordinal_ms(m):
27
+ return num2text.run(f"ordinal {m.group(1)}") + m.group(3)
28
+
29
+ def _ordinal_mp(m):
30
+ return num2text.run(f"ordinal-masculine-plural {m.group(1)}") + m.group(3)
31
+
32
+ def _ordinal_fs(m):
33
+ return num2text.run(f"ordinal-feminine {m.group(1)}") + m.group(3)
34
+
35
+ def _ordinal_fp(m):
36
+ return num2text.run(f"ordinal-feminine-plural {m.group(1)}") + m.group(3)
37
+
38
+ def _fraccions(m):
39
+ return m.group(1) + num2text.run(f"fraction {m.group(2)}") + m.group(3)
40
+
41
+ def _hores(m):
42
+ return m.group(1) + num2text.run(m.group(2)) + " i " + num2text.run(m.group(3)) + m.group(4)
43
+
44
+ def normalize_numbers_ca(text):
45
+ text = re.sub(_separador_milers_re, _esborra_separador_milers, text)
46
+ text = re.sub(_decimal_re, _num2text, text)
47
+ text = re.sub(_ordinal_ms_re, _ordinal_ms, text)
48
+ text = re.sub(_ordinal_mp_re, _ordinal_mp, text)
49
+ text = re.sub(_ordinal_fs_re, _ordinal_fs, text)
50
+ text = re.sub(_ordinal_fp_re, _ordinal_fp, text)
51
+ text = re.sub(_fraccions_re, _fraccions, text)
52
+ text = re.sub(_hores_re, _hores, text)
53
+ text = re.sub(_cardinal_re, _num2text, text)
54
+ return text
text/numbers_ca_test.py ADDED
@@ -0,0 +1,106 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import unittest
2
+
3
+ from text.numbers_ca import normalize_numbers_ca
4
+
5
+
6
+ class NumbersCa(unittest.TestCase):
7
+ def test_cardinals(self):
8
+ """
9
+ Converteix cardinals simples en una frase
10
+ """
11
+ self.assertEqual(normalize_numbers_ca("Va nèixer el 23 de desembre de 1988"), "Va nèixer el vint-i-tres de desembre de mil nou-cents vuitanta-vuit")
12
+ self.assertEqual(normalize_numbers_ca("tinc 3 preguntes"), "tinc tres preguntes")
13
+
14
+ def test_separador_milers(self):
15
+ """
16
+ Ignora separadors de milers
17
+ """
18
+ self.assertEqual(normalize_numbers_ca("1.000"), "mil")
19
+ self.assertEqual(normalize_numbers_ca("323.400"), "tres-cents vint-i-tres mil quatre-cents")
20
+ self.assertEqual(normalize_numbers_ca("900.323.400"), "nou-cents milions tres-cents vint-i-tres mil quatre-cents")
21
+
22
+ def test_decimals(self):
23
+ """
24
+ Converteix decimals
25
+ """
26
+ self.assertEqual(normalize_numbers_ca("1,33"), "u coma trenta-tres")
27
+ self.assertEqual(normalize_numbers_ca("75,5"), "setanta-cinc coma cinc")
28
+ self.assertEqual(normalize_numbers_ca("75,555"), "setanta-cinc coma cinc-cents cinquanta-cinc")
29
+ self.assertEqual(normalize_numbers_ca("999.999.999,99"), "nou-cents noranta-nou milions nou-cents noranta-nou mil nou-cents noranta-nou coma noranta-nou")
30
+ self.assertEqual(normalize_numbers_ca("1,12345678900"), "u coma dotze trenta-quatre cinquanta-sis set-cents vuitanta-nou")
31
+
32
+ def test_decimals_2(self):
33
+ """
34
+ Ignora comes que no pertànyen a un número decimal
35
+ """
36
+ self.assertEqual(normalize_numbers_ca("Va comprar pa, vi i llonganisses"), "Va comprar pa, vi i llonganisses")
37
+ self.assertEqual(normalize_numbers_ca("El número guanyador és 1, 23, 55, 34"), "El número guanyador és u, vint-i-tres, cinquanta-cinc, trenta-quatre")
38
+
39
+ def test_ordinals_ms(self):
40
+ """
41
+ Converteix ordinals masculins singulars
42
+ """
43
+ self.assertEqual(normalize_numbers_ca("Va arribar 4t de 5"), "Va arribar quart de cinc")
44
+ self.assertEqual(normalize_numbers_ca("el 1r va ser ell"), "el primer va ser ell")
45
+ self.assertEqual(normalize_numbers_ca("el 3er, no va aguantar"), "el tercer, no va aguantar")
46
+ self.assertEqual(normalize_numbers_ca("2n"), "segon")
47
+ self.assertEqual(normalize_numbers_ca("2on"), "segon")
48
+ self.assertEqual(normalize_numbers_ca("4t"), "quart")
49
+ self.assertEqual(normalize_numbers_ca("4rt"), "quart")
50
+ self.assertEqual(normalize_numbers_ca("5è: remogueu la barreja"), "cinquè: remogueu la barreja")
51
+ self.assertEqual(normalize_numbers_ca("6e"), "sisè")
52
+ self.assertEqual(normalize_numbers_ca("6e"), "sisè")
53
+ self.assertEqual(normalize_numbers_ca("21nè"), "vint-i-unè")
54
+ self.assertEqual(normalize_numbers_ca("un 81ne de Palamós"), "un vuitanta-unè de Palamós")
55
+
56
+ def test_ordinals_fs(self):
57
+ """
58
+ Converteix ordinals femenins singulars
59
+ """
60
+ self.assertEqual(normalize_numbers_ca("1a"), "primera")
61
+ self.assertEqual(normalize_numbers_ca("3ra"), "tercera")
62
+ self.assertEqual(normalize_numbers_ca("2a"), "segona")
63
+ self.assertEqual(normalize_numbers_ca("2na"), "segona")
64
+ self.assertEqual(normalize_numbers_ca("4a."), "quarta.")
65
+ self.assertEqual(normalize_numbers_ca("pugi a la 4ta, després giri a l'esquerra"), "pugi a la quarta, després giri a l'esquerra")
66
+ self.assertEqual(normalize_numbers_ca("va quedar 5a en la classificació"), "va quedar cinquena en la classificació")
67
+ self.assertEqual(normalize_numbers_ca("la 5na vegada"), "la cinquena vegada")
68
+
69
+ def test_ordinals_mp(self):
70
+ """
71
+ Converteix ordinals masculins plurals
72
+ """
73
+ self.assertEqual(normalize_numbers_ca("1rs"), "primers")
74
+ self.assertEqual(normalize_numbers_ca("van arribar 2ns"), "van arribar segons")
75
+
76
+ def test_ordinals_fp(self):
77
+ """
78
+ Converteix ordinals femenins plurals
79
+ """
80
+ self.assertEqual(normalize_numbers_ca("1es"), "primeres")
81
+
82
+ def test_fraccions_s(self):
83
+ """
84
+ Converteix fraccions singulars
85
+ """
86
+ self.assertEqual(normalize_numbers_ca("1/2 got de vi"), "mig got de vi")
87
+ self.assertEqual(normalize_numbers_ca("1/3 de farina"), "un terç de farina")
88
+ self.assertEqual(normalize_numbers_ca("1/8"), "un vuitè")
89
+
90
+ def test_fraccions_p(self):
91
+ """
92
+ Converteix fraccions plurals
93
+ """
94
+ self.assertEqual(normalize_numbers_ca("4/2 gots de vi"), "quatre migs gots de vi")
95
+ self.assertEqual(normalize_numbers_ca("2/3 de farina"), "dos terços de farina")
96
+ self.assertEqual(normalize_numbers_ca("3/8"), "tres vuitens")
97
+
98
+ def test_hores(self):
99
+ """
100
+ Converteix hores de manera simplificada
101
+ """
102
+ self.assertEqual(normalize_numbers_ca("a les 11:45"), "a les onze i quaranta-cinc")
103
+ self.assertEqual(normalize_numbers_ca("a partir de les 23:12"), "a partir de les vint-i-tres i dotze")
104
+
105
+ if __name__ == '__main__':
106
+ unittest.main()
text/soros.py ADDED
@@ -0,0 +1,140 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ "Soros interpreter (see http://numbertext.org)"
2
+ from __future__ import unicode_literals
3
+ from __future__ import print_function
4
+ import re
5
+ import sys
6
+
7
+
8
+ def run(program, data, lang):
9
+ return compile(program, lang).run(data)
10
+
11
+
12
+ def compile(program, lang):
13
+ return _Soros(program, lang)
14
+
15
+ # conversion function
16
+
17
+
18
+ def _tr(text, chars, chars2, delim):
19
+ for i in range(0, len(chars)):
20
+ text = text.replace(delim + chars[i], chars2[i])
21
+ return text
22
+
23
+
24
+ # string literals for metacharacter encoding
25
+ _m = "\\\";#$()|[]"
26
+ # Unicode private area
27
+ _c = u"\uE000\uE001\uE002\uE003\uE004\uE005\uE006\uE007\uE008\uE009"
28
+ _pipe = u"\uE003"
29
+ # separator prefix = \uE00A
30
+
31
+ # pattern to recognize function calls in the replacement string
32
+ _func = re.compile(_tr(r"""(?:\|?(?:\$\()+)? # optional nested calls
33
+ (\|?\$\(([^\(\)]*)\)\|?) # inner call (2 subgroups)
34
+ (?:\)+\|?)?""", # optional nested calls
35
+ _m[4:8], _c[:4], "\\"), re.X) # \$, \(, \), \| -> \uE000..\uE003
36
+
37
+
38
+ class _Soros:
39
+ def __init__(self, prg, lang):
40
+ self.lines = []
41
+ if prg.find("__numbertext__") == -1:
42
+ prg = "__numbertext__;" + prg
43
+ # default left zero deletion
44
+ # and separator function (no separation, if subcall returns with empty string)
45
+ prg = prg.replace("__numbertext__", u"""0+(0|[1-9]\\d*) $1
46
+ \"([a-z][-a-z]* )0+(0|[1-9]\\d*)\" $(\\1\\2)
47
+ \"\uE00A(.*)\uE00A(.+)\uE00A(.*)\" \\1\\2\\3
48
+ \"\uE00A.*\uE00A\uE00A.*\"
49
+ """)
50
+ prg = _tr(prg, _m[:4], _c[:4],
51
+ "\\") # \\, \", \;, \# -> \uE000..\uE003
52
+ # switch off all country-dependent lines, and switch on the requested ones
53
+ prg = re.sub(
54
+ r"(^|[\n;])([^\n;#]*#[^\n]*[\[]:[^\n:\]]*:][^\n]*)", r"\1#\2", prg)
55
+ prg = re.sub(r"(^|[\n;])#([^\n;#]*#[^\n]*[\[]:" +
56
+ lang.replace("_", "-") + r":][^\n]*)", r"\1\2", prg)
57
+ matchline = re.compile("^\s*(\"[^\"]*\"|[^\s]*)\s*(.*[^\s])?\s*$")
58
+ prefix = ""
59
+ for s in re.sub("(#[^\n]*)?(\n|$)", ";", prg).split(";"):
60
+ macro = re.match("== *(.*[^ ]?) ==", s)
61
+ if macro != None:
62
+ prefix = macro.group(1)
63
+ continue
64
+ m = matchline.match(s)
65
+ if prefix != "" and s != "" and m != None:
66
+ s = m.group(1).strip("\"")
67
+ space = " " if s != "" else ""
68
+ caret = ""
69
+ if s[0:1] == "^":
70
+ s = s[1:]
71
+ caret = "^"
72
+ s2 = m.group(2) if m.group(2) != None else ""
73
+ s = "\"" + caret + prefix + space + s + "\" " + s2
74
+ m = matchline.match(s)
75
+ if m != None:
76
+ s = _tr(m.group(1).strip("\""), _c[1:4], _m[1:4], "") \
77
+ .replace(_c[_m.find("\\")], "\\\\") # -> \\, ", ;, #
78
+ if m.group(2) != None:
79
+ s2 = m.group(2).strip("\"")
80
+ else:
81
+ s2 = ""
82
+ # \$, \(, \), \|, \[, \] -> \uE004..\uE009
83
+ s2 = _tr(s2, _m[4:], _c[4:], "\\")
84
+ # call inner separator: [ ... $1 ... ] -> $(\uE00A ... \uE00A$1\uE00A ... )
85
+ s2 = re.sub(r"[\[]\$(\d\d?|\([^\)]+\))",
86
+ u"$(\uE00A\uE00A|$\\1\uE00A", s2)
87
+ s2 = re.sub(r"[\[]([^\$[\\]*)\$(\d\d?|\([^\)]+\))",
88
+ u"$(\uE00A\\1\uE00A$\\2\uE00A", s2)
89
+ # add "|" in terminating position
90
+ s2 = re.sub(r"\uE00A]$", "|\uE00A)", s2)
91
+ s2 = re.sub(r"]", ")", s2)
92
+ s2 = re.sub(r"(\$\d|\))\|\$", r"\1||$",
93
+ s2) # $()|$() -> $()||$()
94
+ # \uE000..\uE003-> \, ", ;, #
95
+ s2 = _tr(s2, _c[:4], _m[:4], "")
96
+ # $, (, ), | -> \uE000..\uE003
97
+ s2 = _tr(s2, _m[4:8], _c[:4], "")
98
+ # \uE004..\uE009 -> $, (, ), |, [, ]
99
+ s2 = _tr(s2, _c[4:], _m[4:], "")
100
+ s2 = re.sub(r"\\(\d)", r"\\g<\1>",
101
+ re.sub(r"\uE000(\d)", "\uE000\uE001\\\\g<\\1>\uE002", s2))
102
+ try:
103
+ self.lines = self.lines + [[
104
+ re.compile("^" + s.lstrip("^").rstrip("$") + "$"),
105
+ s2, s[:1] == "^", s[-1:] == "$"]]
106
+ except:
107
+ print("Error in following regex line: " + s, file=sys.stderr)
108
+ raise
109
+
110
+ def run(self, data):
111
+ return self._run(data, True, True)
112
+
113
+ def _run(self, data, begin, end):
114
+ for i in self.lines:
115
+ if not ((begin == False and i[2]) or (end == False and i[3])):
116
+ m = i[0].match(data)
117
+ if m:
118
+ try:
119
+ s = m.expand(i[1])
120
+ except:
121
+ print("Error for the following input: " +
122
+ data, file=sys.stderr)
123
+ raise
124
+ n = _func.search(s)
125
+ while n:
126
+ b = False
127
+ e = False
128
+ if n.group(1)[0:1] == _pipe or n.group()[0:1] == _pipe:
129
+ b = True
130
+ elif n.start() == 0:
131
+ b = begin
132
+ if n.group(1)[-1:] == _pipe or n.group()[-1:] == _pipe:
133
+ e = True
134
+ elif n.end() == len(s):
135
+ e = end
136
+ s = s[:n.start(1)] + self._run(n.group(2),
137
+ b, e) + s[n.end(1):]
138
+ n = _func.search(s)
139
+ return s
140
+ return ""
text/symbols.py ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ """ from https://github.com/keithito/tacotron """
2
+
3
+ '''
4
+ Defines the set of symbols used in text input to the model.
5
+
6
+ The default is a set of ASCII characters that works well for English or text that has been run through Unidecode. For other data, you can modify _characters. See TRAINING_DATA.md for details. '''
7
+ from text import cmudict
8
+
9
+ _pad = '_' # in principle not used in tacotron2
10
+ _punctuation = '\'!,.?…· '
11
+ _letters = 'AÀÁBCÇDEÉÈFGHIÍÏJKLMNOÓÒPQRSTUÜÚVWXYZaàábcçdeéèfghiíïjklmnoóòpqrstuüúvwxyz'
12
+
13
+ # Prepend "@" to ARPAbet symbols to ensure uniqueness (some are the same as uppercase letters):
14
+ _arpabet = ['@' + s for s in cmudict.valid_symbols]
15
+
16
+ # Export all symbols:
17
+ symbols = [_pad] + list(_punctuation) + list(_letters) + _arpabet
text/symbols_en.py ADDED
@@ -0,0 +1,18 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ """ from https://github.com/keithito/tacotron """
2
+
3
+ '''
4
+ Defines the set of symbols used in text input to the model.
5
+
6
+ The default is a set of ASCII characters that works well for English or text that has been run through Unidecode. For other data, you can modify _characters. See TRAINING_DATA.md for details. '''
7
+ from text import cmudict
8
+
9
+ _pad = '_'
10
+ _punctuation = '!\'(),.:;? '
11
+ _special = '-'
12
+ _letters = 'ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz'
13
+
14
+ # Prepend "@" to ARPAbet symbols to ensure uniqueness (some are the same as uppercase letters):
15
+ _arpabet = ['@' + s for s in cmudict.valid_symbols]
16
+
17
+ # Export all symbols:
18
+ symbols = [_pad] + list(_special) + list(_punctuation) + list(_letters) + _arpabet