Spaces:
Runtime error
Runtime error
File size: 9,571 Bytes
e50fe35 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 |
#
# Copyright (c) 2013-present, Anoop Kunchukuttan
# All rights reserved.
#
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
#
import codecs, sys
from indicnlp.script import indic_scripts as si
import re
chillu_char_map= {
'\u0d7a': '\u0d23',
'\u0d7b': '\u0d28',
'\u0d7c': '\u0d30',
'\u0d7d': '\u0d32',
'\u0d7e': '\u0d33',
'\u0d7f': '\u0d15',
}
char_chillu_map= {}
for k,v in chillu_char_map.items():
char_chillu_map[v]=k
def normalize_malayalam(word):
word_mask=re.sub(r'[0-9]','0',word)
# instead of chillu characters, use consonant+halant
for chillu,char in chillu_char_map.items():
word=word.replace(chillu,'{}\u0d4d'.format(char))
word_mask=word_mask.replace(chillu,'41')
word_mask=re.sub(r'[^0-9]','0',word_mask)
return word, word_mask
def denormalize_malayalam(word, word_mask):
word=list(word)
word_mask=list(word_mask)
## pattern 4
idx=0
while idx>=0:
try:
idx=word_mask.index('4',idx)
word[idx:idx+2]=char_chillu_map[word[idx]]
word_mask[idx:idx+2]='0'
start=idx
except ValueError as e:
break
return ''.join(word)
def normalize_punjabi(word):
word_mask=re.sub(r'[0-9]','0',word)
## replace tippi with anusvaar
word=word.replace('\u0a70','\u0a02')
word_mask=word_mask.replace('\u0a70','2')
## replace addak+consonant with consonat+halant+consonant
word=re.sub(r'\u0a71(.)','\\1\u0a4d\\1',word)
word_mask=re.sub(r'\u0a71(.)','311',word_mask)
word_mask=re.sub(r'[^0-9]','0',word_mask)
return word, word_mask
def denormalize_punjabi(word, word_mask):
word=list(word)
word_mask=list(word_mask)
## pattern 2
idx=0
while idx>=0:
try:
idx=word_mask.index('2',idx)
word[idx]='\u0a70'
word_mask[idx]='0'
start=idx
except ValueError as e:
break
## pattern 3
idx=0
while idx>=0:
try:
idx=word_mask.index('3',idx)
word[idx:idx+3]='\u0a71{}'.format(word[idx])
word_mask[idx:idx+3]='00'
start=idx
except ValueError as e:
break
return ''.join(word)
def char_backoff(syllables_list,vocab):
syllables_final=[]
if vocab is None:
syllables_final=syllables_list
else:
for s in syllables_list:
if s in vocab:
syllables_final.append(s)
else:
for x in s:
syllables_final.append(x)
return syllables_final
def orthographic_syllabify_improved(word,lang,vocab=None):
word_mask=['0']*len(word)
if lang=='ml':
word, word_mask = normalize_malayalam(word)
word=word
elif lang=='pa':
word, word_mask = normalize_punjabi(word)
p_vectors=[si.get_phonetic_feature_vector(c,lang) for c in word]
syllables=[]
syllables_mask=[]
for i in range(len(word)):
v=p_vectors[i]
syllables.append(word[i])
syllables_mask.append(word_mask[i])
### simplified syllabification
#if i+1<len(word) and \
# (not si.is_valid(p_vectors[i+1]) or si.is_misc(p_vectors[i+1])):
# syllables.append(u' ')
# syllables_mask.append(u'0')
#elif not si.is_valid(v) or si.is_misc(v) or si.is_vowel(v):
# syllables.append(u' ')
# syllables_mask.append(u'0')
#elif i+1<len(word) and \
# (si.is_consonant(v) or si.is_nukta(v)) and \
# (si.is_consonant(p_vectors[i+1]) or si.is_anusvaar(p_vectors[i+1])):
# syllables.append(u' ')
# syllables_mask.append(u'0')
#### better syllabification
if i+1<len(word) and (not si.is_valid(p_vectors[i+1]) or si.is_misc(p_vectors[i+1])):
syllables.append(' ')
syllables_mask.append('0')
elif not si.is_valid(v) or si.is_misc(v) :
syllables.append(' ')
syllables_mask.append('0')
elif si.is_vowel(v):
anu_nonplos= ( i+2<len(word) and \
si.is_anusvaar(p_vectors[i+1]) and \
not si.is_plosive(p_vectors[i+2])\
)
anu_eow= ( i+2==len(word) and \
si.is_anusvaar(p_vectors[i+1]) )
if not(anu_nonplos or anu_eow):
syllables.append(' ')
syllables_mask.append('0')
elif i+1<len(word) and \
(si.is_consonant(v) or si.is_nukta(v)):
if si.is_consonant(p_vectors[i+1]):
syllables.append(' ')
syllables_mask.append('0')
elif si.is_vowel(p_vectors[i+1]) and \
not si.is_dependent_vowel(p_vectors[i+1]):
syllables.append(' ')
syllables_mask.append('0')
elif si.is_anusvaar(p_vectors[i+1]):
anu_nonplos= ( i+2<len(word) and \
not si.is_plosive(p_vectors[i+2])\
)
anu_eow= i+2==len(word)
if not(anu_nonplos or anu_eow):
syllables.append(' ')
syllables_mask.append('0')
syllables_mask=''.join(syllables_mask)
syllables=''.join(syllables)
#assert len(syllables_mask) == len(syllables)
#assert syllables_mask.find('01') == -1
if syllables_mask.find('01') >= 0:
print('Warning')
if lang=='ml':
syllables = denormalize_malayalam(syllables,syllables_mask)
elif lang=='pa':
syllables = denormalize_punjabi(syllables,syllables_mask)
syllables_list = syllables.strip().split(' ')
return(char_backoff(syllables_list,vocab))
def orthographic_syllabify(word,lang,vocab=None):
p_vectors=[si.get_phonetic_feature_vector(c,lang) for c in word]
syllables=[]
for i in range(len(word)):
v=p_vectors[i]
syllables.append(word[i])
### simplified syllabification
#if i+1<len(word) and \
# (not si.is_valid(p_vectors[i+1]) or si.is_misc(p_vectors[i+1])):
# syllables.append(u' ')
#elif not si.is_valid(v) or si.is_misc(v) or si.is_vowel(v):
# syllables.append(u' ')
#elif i+1<len(word) and \
# (si.is_consonant(v) or si.is_nukta(v)) and \
# (si.is_consonant(p_vectors[i+1]) or si.is_anusvaar(p_vectors[i+1])):
# syllables.append(u' ')
#### better syllabification
if i+1<len(word) and (not si.is_valid(p_vectors[i+1]) or si.is_misc(p_vectors[i+1])):
syllables.append(' ')
elif not si.is_valid(v) or si.is_misc(v) :
syllables.append(' ')
elif si.is_vowel(v):
anu_nonplos= ( i+2<len(word) and \
si.is_anusvaar(p_vectors[i+1]) and \
not si.is_plosive(p_vectors[i+2])\
)
anu_eow= ( i+2==len(word) and \
si.is_anusvaar(p_vectors[i+1]) )
if not(anu_nonplos or anu_eow):
syllables.append(' ')
elif i+1<len(word) and \
(si.is_consonant(v) or si.is_nukta(v)):
if si.is_consonant(p_vectors[i+1]):
syllables.append(' ')
elif si.is_vowel(p_vectors[i+1]) and \
not si.is_dependent_vowel(p_vectors[i+1]):
syllables.append(' ')
elif si.is_anusvaar(p_vectors[i+1]):
anu_nonplos= ( i+2<len(word) and \
not si.is_plosive(p_vectors[i+2])\
)
anu_eow= i+2==len(word)
if not(anu_nonplos or anu_eow):
syllables.append(' ')
syllables_list = ''.join(syllables).strip().split(' ')
return(char_backoff(syllables_list,vocab))
def orthographic_simple_syllabify(word,lang,vocab=None):
p_vectors=[si.get_phonetic_feature_vector(c,lang) for c in word]
syllables=[]
for i in range(len(word)):
v=p_vectors[i]
syllables.append(word[i])
## simplified syllabification
if i+1<len(word) and \
(not si.is_valid(p_vectors[i+1]) or si.is_misc(p_vectors[i+1])):
syllables.append(' ')
elif not si.is_valid(v) or si.is_misc(v) or si.is_vowel(v):
syllables.append(' ')
elif i+1<len(word) and \
(si.is_consonant(v) or si.is_nukta(v)) and \
(si.is_consonant(p_vectors[i+1]) or si.is_anusvaar(p_vectors[i+1])):
syllables.append(' ')
syllables_list = ''.join(syllables).strip().split(' ')
return(char_backoff(syllables_list,vocab))
|