forked from SergeyShk/Word-to-Number-Russian
-
Notifications
You must be signed in to change notification settings - Fork 2
/
extractor.py
305 lines (241 loc) · 9.08 KB
/
extractor.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
302
303
304
305
from number import NUMBER
from natasha.extractors import Extractor
import re
NON_SUMMABLE_FIRST_NUMBERS = [10]
def is_summable(num1, num2):
"""
If two numbers can be summed in a needed way.
Examples
-----
9000 80 -> True (9080)
20 30 -> False (20 30)
1 2 -> False (1 2)
2 1 -> False (2 1)
20 1 -> True (21)
-----
"""
if num1 == 10 or num2 == 0:
return False
place = len(str(num2))
str1 = str(num1)
if len(str1) > place and str1[-place] == "0":
return True
return False
def can_be_merged(num1, num2):
"""
If two numbers can be merged.
Examples
-----
200 7000 -> True (207000)
250 7000 -> True (257000)
251 7000 -> False (251 7000)
-----
"""
if num2 > num1:
m = get_multiplexer(num2)
num = int(num2 / m)
if is_summable(num1, num):
return True
return False
def get_multiplexer(num):
"""
Get max decimal multiplexer.
"""
m = 1
while num % 10 == 0:
num = num / 10
m *= 10
return m
def squash_spaces(text):
"""
Merge duplicated spaces
"""
return re.sub(" +", " ", text)
class NumberExtractor(Extractor):
def __init__(self):
super(NumberExtractor, self).__init__(NUMBER)
def replace(self, text, apply_regrouping=False):
"""
Replace ungrouped numbers in text
Parameters
----------
text: input text
Returns
-------
new_text: result text
"""
mask, new_text, start = [], "", 0
if text:
# correct span counting
text = f" {text} "
matches = self.parser.findall(text)
for match in matches:
span_text = text[start : match.span.start]
mask = self.update_first_mask(span_text, mask)
if match.fact.multiplier:
num = match.fact.int * match.fact.multiplier
mask.append(2)
else:
num = match.fact.int
mask.append(1)
new_text += text[start : match.span.start] + str(num)
start = match.span.stop
new_text += text[start:]
mask = self.update_first_mask(text[start:], mask)
if start == 0:
if apply_regrouping:
return self.regroup_number_sequences(text.strip(), mask)
return text.strip(), mask
elif apply_regrouping:
return self.regroup_number_sequences(new_text.strip(), mask)
return new_text.strip(), mask
else:
return "", mask
def update_first_mask(self, span_text, mask):
"""
Update first pass mask (mask in 'replace' method)
"""
append_to_mask_count = self.get_words_count(span_text)
if append_to_mask_count < 0:
mask = mask[:-1]
elif append_to_mask_count > 0:
for _ in range(append_to_mask_count):
mask.append(1)
return mask
def get_words_count(self, text):
"""
Calculate words count between two extracted items
"""
text = squash_spaces(text)
if not text:
return 0
space_count = text.count(" ")
if space_count == 0:
return -1
return space_count - 1
def regroup_number_sequences(self, text_orig, first_mask):
"""
Find number sequences and apply regrouping.
"""
text, curr_text = text_orig, ""
if text:
start = 0
handled_matches = set()
retry = True
while True:
start = 0
match_amount = len(re.findall("\d+[ \d]+\d+", text))
if not retry or not match_amount:
break
omit_end = False
for i, m in enumerate(re.finditer("\d+[ \d]+\d+", text)):
if i + 1 == match_amount and not omit_end:
retry = False
if m.span(0) in handled_matches:
continue
elif omit_end:
continue
else:
handled_matches.add(m.span(0))
if m.group(0).strip():
span_start = m.span(0)[0]
span_end = m.span(0)[1]
span_text = text[:span_end]
mask_index = len(span_text.split())
mask_part = first_mask[:mask_index]
curr_text += text[start:span_start]
start = span_end
regrouped, squashed_idxs = self.regroup_after_replace(
m.group(0)
)
curr_text += regrouped
curr_part = self.update_mask(mask_part, squashed_idxs)
first_mask = self.merge_masks(first_mask, mask_part, curr_part)
new_text = self.merge_texts(text, curr_text, span_end)
if new_text != text:
omit_end = True
text = new_text
curr_text = ""
return text, first_mask
def merge_texts(self, old, new, span_end):
"""
Merge partially handled text with full initial one
"""
res = new + old[span_end:]
return res
def merge_masks(self, first_mask, mask_part, curr_part):
"""
Merge partially handled mask with full initial one
"""
res = curr_part + first_mask[len(mask_part) :]
return res
def update_mask(self, mask_part, squashed_idxs):
"""
Calculate mask using squashed words indices
"""
res = []
shift = 0
for count in squashed_idxs[::-1]:
val = 0
for i in range(count):
val += mask_part[-1 - i - shift]
res.insert(0, val)
shift += count
while shift < len(mask_part):
res.insert(0, mask_part[-1 - shift])
shift += 1
return res
def regroup_after_replace(self, text):
"""
Regroup sequence of numbers.
500 20 2000 -> 522000
30 1 2 3 4 -> 31 2 3 4
"""
squashed_idxs = []
if text:
nums = [int(x) for x in text.split()]
new_nums = [nums[0]]
acc = 1
for i, n in enumerate(nums[1:]):
if is_summable(new_nums[-1], n):
new_nums[-1] = new_nums[-1] + n
acc += 1
elif can_be_merged(new_nums[-1], n):
m = get_multiplexer(n)
extracted = int(n / m)
new_nums[-1] = (new_nums[-1] + extracted) * m
acc += 1
else:
new_nums.append(n)
squashed_idxs.append(acc)
acc = 1
squashed_idxs.append(acc)
new_text = " ".join([str(n) for n in new_nums])
assert sum(squashed_idxs) == len(text.split())
return new_text, squashed_idxs
else:
return ""
# fix long numbers
# text = "Госдолг США в тысяча девятьсот пятидесятом году составил двести пятьдесят шесть миллиардов девятьсот миллионов долларов"
# # fix 0
# text = "пять шесть ноль ноль ноль семь двадцать ноль"
# # we need to return word indices
# text = "годы его правления одна тысяча восемьсот тридцать пятый и две тысячи девятьсот пятьдесят четвертый годы"
# text = "семьсот миллиардов один рубль, один, два, три три"
# text = "Выплаты за второго-третьего ребенка выросли на девять сотых процента"
# text = "Я купил сорок пять килограмм картошки и 7 пудов моркови"
# text = "один, ,два"
# text = "Я купил сорок пять килограмм картошки и 7 пудов моркови"
# text = "один, ,два три, четыре, пять"
# text = "один,двадцать два какой то текст"
# text = "один два, тридцать три, пятьдесят пять,шестьдесят шесть сто двадцать четыре, привет как дела"
# text = ""
# text = "здесь тридцать три тысяча два числа"
# replaced, counter_mask = extractor.replace(text)
# print("INPUT:", text)
# print("REPLACED:", replaced)
# print(counter_mask)
extractor = NumberExtractor()
text = "Выплаты за второго-третьего ребенка выросли на девять сотых процента"
text, mask = extractor.replace(text, apply_regrouping=True)
print("RESULT:", text, mask)