-
Notifications
You must be signed in to change notification settings - Fork 6
/
Copy pathindonesian_stemmer.py
executable file
·301 lines (276 loc) · 11.9 KB
/
indonesian_stemmer.py
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
#!/usr/bin/env python
import os
import random
import pprint
import time
import re
import sys
from collections import OrderedDict
class ILStemmer(object):
def __init__(self):
self.OPTION = {
'SORT_INSTANCE' : False, #sort by number of instances
'NO_NO_MATCH' : False, #hide no match entry
'NO_DIGIT_ONLY' : True, #hide digit only
'STRICT_CONFIX' : False, #use strict disallowed_confixes RULES
'ALLOW_HASHTAGS' : False, #preserve hashtags. cause, you know.. netizen
}
BASEDIR = os.path.dirname(os.path.abspath(__file__))
fp = open(os.path.join(BASEDIR, 'kamus.txt'))
dic = fp.readlines()
fp.close()
self.dicti = {}
for i in dic:
attrib = i.lower().split('\t')
key = attrib[1].replace(' ','').rstrip(' \t\n\r')
self.dicti[key] = { 'class': attrib[0], 'lemma': attrib[1].rstrip(' \t\n\r')}
# Define RULES
self.VOWEL = 'a|i|u|e|o' #vowels
self.CONSONANT = 'b|c|d|f|g|h|j|k|l|m|n|p|q|r|s|t|v|w|x|y|z' #consonants
self.ANY = self.VOWEL + '|' + self.CONSONANT #any characters
self.RULES = {
'affixes': (
(1,('kah','lah', 'tah', 'pun')),
(1,('mu','ku', 'nya')),
(0,('ku','kau')),
(1,('i','kan', 'an'))
),
'prefixes': (
[0, "(di|ke|se)("+self.ANY+")(.+)", ""], # 0
[0, "(ber|ter)("+self.ANY+")(.+)", ""], # 1, 6 normal
[0, "(be|te)(r)("+self.VOWEL+")(.+)", ""], # 1, 6 be-rambut
[0, "(be|te)("+self.CONSONANT+")({"+self.ANY+"}?)(er)(.+)", ""], # 3, 7 te-bersit, te-percaya
[0, "(bel|pel)(ajar|unjur)", ""], # ajar, unjur
[0, "(me|pe)(l|m|n|r|w|y)(.+)", ""], # 10, 20: merawat, pemain
[0, "(mem|pem)(b|f|v)(.+)", ""], # 11 23: membuat, pembuat
[0, "(men|pen)(c|d|j|z)(.+)", ""], # 14 27: mencabut, pencabut
[0, "(meng|peng)(g|h|q|x)(.+)", ""], # 16 29: menggiring, penghasut
[0, "(meng|peng)("+self.VOWEL+")(.+)", ""], # 17 30 meng-anjurkan, peng-anjur
[0, "(mem|pem)("+self.VOWEL+")(.+)", "p"], # 13 26: memerkosa, pemerkosa
[0, "(men|pen)("+self.VOWEL+")(.+)", "t"], # 15 28 menutup, penutup
[0, "(meng|peng)("+self.VOWEL+")(.+)", "k"], # 17 30 mengalikan, pengali
[0, "(meny|peny)("+self.VOWEL+")(.+)", "s"], # 18 31 menyucikan, penyucian
[0, "(mem)(p)("+self.CONSONANT+")(.+)", ""], # memproklamasikan
[0, "(pem)("+self.CONSONANT+")(.+)", "p"], # pemrogram
[0, "(men|pen)(t)("+self.CONSONANT+")(.+)", ""], # mentransmisikan pentransmisian
[0, "(meng|peng)(k)("+self.CONSONANT+")(.+)", ""], # mengkristalkan pengkristalan
[0, "(men|pen)(s)("+self.CONSONANT+")(.+)", ""], # mensyaratkan pensyaratan
[0, "(menge|penge)("+self.CONSONANT+")(.+)", ""], # swarabakti: mengepel
[0, "(mempe)(r)("+self.VOWEL+")(.+)", ""], # 21
[0, "(memper)("+self.ANY+")(.+)", ""], # 21
[0, "(pe)("+self.ANY+")(.+)", ""], # 20
[0, "(per)("+self.ANY+")(.+)", ""], # 21
[0, "(pel)("+self.CONSONANT+")(.+)", ""], # 32 pelbagai, other?
[0, "(mem)(punya)", ""], # Exception: mempunya
[0, "(pen)(yair)", "s"], # Exception: penyair > syair
),
'disallowed_confixes' : (
('ber-', '-i'),
('ke-', '-i'),
('pe-', '-kan'),
('di-', '-an'),
('meng-', '-an'),
('ter-', '-an'),
('ku-', '-an'),
),
'allomorphs' : (
{'be' : ('be-', 'ber-', 'bel-')},
{'te' : ('te-', 'ter-', 'tel-')},
{'pe' : ('pe-', 'per-', 'pel-', 'pen-', 'pem-', 'peng-', 'peny-', 'penge-')},
{'me' : ('me-', 'men-', 'mem-', 'meng-', 'meny-', 'menge-')},
),
}
def cnom(self, pattern, thestring):
# count non overlapping matches
return re.subn(pattern, '', thestring)[1]
def stem(self, query):
words = {}
instance = {}
if self.OPTION['ALLOW_HASHTAGS']:
paw = re.compile('[^A-Za-z0-9#]+')
else:
paw = re.compile('[^A-Za-z0-9]+')
raw = paw.split(query)
for index,r in enumerate(raw):
if self.OPTION['NO_DIGIT_ONLY'] and re.search('^\d',r):
continue
key = r.lower()
words[key]= { 'count': self.cnom(key,query), 'order': index}
words = OrderedDict(sorted(words.iteritems(), key=lambda x: x[1]['order']))
for k,v in words.items():
if k in self.dicti.keys():
words[k]['roots'] = {}
words[k]['roots'][k] = {}
words[k]['roots'][k]['lemma'] = k
else:
words[k]['roots'] = self.stem_word(k)
if len(words[k]['roots']) == 0 and self.OPTION['NO_NO_MATCH']:
del words[k]
continue
instance[k] = { 'count': v['count'], 'order': v['order']}
instance = OrderedDict(sorted(instance.iteritems(), key=lambda x: x[1]['order']))
word_count = len(words)
if self.OPTION['SORT_INSTANCE']:
instance = sorted(instance)
newwords = []
for key in instance:
newwords.append((key, words[key]))
words = newwords
else:
newwords = []
for key in instance.keys():
newwords.append((key, words[key]))
words = newwords
#return words
#presenting the roots
stemmed = ''
for i in range(len(words)):
result = words[i][1]['roots']
for key in result.keys():
if 'prefixes' in result[key].keys():
stemmed += str(result[key]['prefixes'])
stemmed += str(result[key]['lemma'])
if 'suffixes' in result[key].keys():
stemmed += str(result[key]['suffixes'])
stemmed += ' '
if len(result) == 0 and not self.OPTION['NO_NO_MATCH']:
stemmed += words[i][0]
stemmed += ' '
return stemmed.rstrip()
def stem_root(self, query):
words = {}
instance = {}
if self.OPTION['ALLOW_HASHTAGS']:
paw = re.compile('[^A-Za-z0-9#]+')
else:
paw = re.compile('[^A-Za-z0-9]+')
raw = paw.split(query)
for index,r in enumerate(raw):
if self.OPTION['NO_DIGIT_ONLY'] and re.search('^\d',r):
continue
key = r.lower()
words[key]= { 'count': self.cnom(key,query), 'order': index}
words = OrderedDict(sorted(words.iteritems(), key=lambda x: x[1]['order']))
for k,v in words.items():
if k in self.dicti.keys():
words[k]['roots'] = {}
words[k]['roots'][k] = {}
words[k]['roots'][k]['lemma'] = k
else:
words[k]['roots'] = self.stem_word(k)
if len(words[k]['roots']) == 0 and self.OPTION['NO_NO_MATCH']:
del words[k]
continue
instance[k] = { 'count': v['count'], 'order': v['order']}
instance = OrderedDict(sorted(instance.iteritems(), key=lambda x: x[1]['order']))
word_count = len(words)
if self.OPTION['SORT_INSTANCE']:
instance = sorted(instance)
newwords = []
for key in instance:
newwords.append((key, words[key]))
words = newwords
else:
newwords = []
for key in instance.keys():
newwords.append((key, words[key]))
words = newwords
#return words
#presenting the roots
stemmed = ''
for i in range(len(words)):
result = words[i][1]['roots']
for key in result.keys():
stemmed += str(result[key]['lemma'])
stemmed += ' '
if len(result) == 0 and not self.OPTION['NO_NO_MATCH']:
stemmed += words[i][0]
stemmed += ' '
return stemmed.rstrip()
def ksort(self, data):
return [(key,data[key]) for key in sorted(data.keys(), reverse=True)]
def stem_word(self, word):
word = word.strip()
roots = { word : {}}
#check the word in dictionary (kamus.txt)
if word in self.dicti.keys():
roots[word]['affixes'] = {}
else:
roots[word]['affixes'] = ''
#if the word contain dash (-)
dash_parts = word.split('-')
if len(dash_parts) > 1:
for dash_part in dash_parts:
roots[dash_part]['affixes'] = {}
#check affixes rules
for group in self.RULES['affixes']:
is_suffix = group[0]
affixes = group[1]
for affix in affixes:
if is_suffix:
pattern = "(.+)("+affix+")"
else:
pattern = "("+affix+")(.+)"
self.add_root(roots, [is_suffix, pattern, ''])
#check prefixes rules
for i in range(3):
for rule in self.RULES['prefixes']:
self.add_root(roots, rule)
#compiling the output
for lemma, attrib in roots.items():
if lemma not in self.dicti.keys():
del roots[lemma] #delete arrays if it's not in the dictionary
continue
if self.OPTION['STRICT_CONFIX']:
continue;
affixes = attrib['affixes']
for pair in self.RULES['disallowed_confixes']:
prefix = pair[0]
suffix = pair[1]
prefix_key = prefix[:2]
if prefix_key in self.RULES['allomorphs']:
for allomorf in self.RULES['allomorphs'][prefix_key]:
if allomorf in affixes and suffix in affixes:
del roots[lemma]
else:
if prefix in affixes and suffix in affixes:
del roots[lemma]
for lemma,attrib in roots.items():
affixes = attrib['affixes']
attrib['lemma'] = self.dicti[lemma]['lemma']
attrib['class'] = self.dicti[lemma]['class']
for affix in attrib['affixes']:
if affix[:1] == '-':
tipe = 'suffixes'
else:
tipe = 'prefixes'
attrib[tipe] = affix
try:
if len(attrib['suffixes']) > 1:
ksort(attrib['suffixes'])
except:
pass
roots[lemma] = attrib
return roots
def add_root(self, roots, rule):
is_suffix = rule[0]
pattern = '^' + rule[1]
variant = rule[2]
is_array = lambda var: isinstance(var, (list, tuple))
for lemma, attrib in roots.items():
matches = re.findall(pattern,lemma)
if len(matches) > 0:
new_lemma = ''
new_affix = ''
affix_index = 1 if is_suffix else 0
for i in range(len(matches[0])):
if i != affix_index: new_lemma += matches[0][i]
if variant:
new_lemma = variant + new_lemma
new_affix += '-' if is_suffix else ''
new_affix += matches[0][affix_index]+''
new_affix += '' if is_suffix else '-'
new_affix = new_affix.split()
if is_array(attrib['affixes']):
new_affix = attrib['affixes'] + new_affix
roots[new_lemma] = { 'affixes': new_affix}