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
|
#!/usr/bin/python3
# coding=utf-8
# -*- encoding: utf-8 -*-
from collections import defaultdict
import common
import math
import re
import sys
def safe_float(s):
if not s:
return 0.0, False
try:
f = float(s)
if math.isnan(f):
return 0.0, False
return f, True
except:
return 0.0, False
def strip_tags(s, side, space=False):
idx = 0 if side == 'sl' else 1
ret = s.split('/')[idx]
if not space:
ret = ret.replace(' ', '~')
if ret.count('><') > 0:
ret = ret.split('><')[0] + '>'
return ret
class BiltransCounter:
lu_sep = re.compile(r'\$[^\^]*\^')
tokenizer = 'regex' # or 'biltrans'
line_ids = False
count_ngrams = False
max_ngrams = 3
biltrans_wrap_lus = False
def __init__(self):
self.reading = False
self.am_file = None
self.am_line = None
self.am_row = None
self.am_id = None
self.am_linenum = 0
self.dm_file = None
self.dm_line = None
self.dm_row = None
self.dm_id = None
self.dm_linenum = 0
self.clear_ngrams()
def __del__(self):
if self.am_file:
self.am_file.close()
if self.dm_file:
self.dm_file.close()
def next_am_line(self):
self.am_line = self.am_file.readline()
self.am_linenum += 1
if not self.am_line:
self.am_id, self.am_row = None, []
self.reading = False
return
ls = self.am_line.split('\t')
if self.line_ids:
self.am_id = int(ls[0].strip())
if self.tokenizer == 'regex':
self.am_row = self.lu_sep.split(ls[1].strip()[1:-1])
elif self.tokenizer == 'biltrans':
self.am_row = common.tokenize_biltrans_line(self.am_line)
def next_dm_line(self):
self.dm_linenum += 1
self.dm_line = self.dm_file.readline()
if not self.dm_line:
self.dm_id, self.dm_row = None, []
self.reading = False
return
ls = self.dm_line.split('\t')
if self.line_ids:
self.dm_id = int(self.dm_line.split('.[][')[1].split()[0])
if self.tokenizer == 'regex':
self.dm_row = self.lu_sep.split(ls[1].strip()[1:-1])
elif self.tokenizer == 'biltrans':
self.dm_row = common.tokenize_biltrans_line(self.dm_line)
def clear_ngrams(self):
self.ngrams = defaultdict(
lambda: defaultdict(lambda: defaultdict(lambda: 0.0)))
def check_rows(self):
if len(self.am_row) != len(self.dm_row):
print(
'Mismatch in number of LUs between analysis and training', file=sys.stderr)
print('\t' + self.am_line, file=sys.stderr)
print('\t' + self.dm_line, file=sys.stderr)
print('...skipping', file=sys.stderr)
return False
return True
def read_files_multi_dm(self, am_fname, dm_fname):
self.next_dm_line()
while self.reading:
self.next_am_line()
while self.am_id == self.dm_id and self.reading:
frac_count = 0
if self.dm_line.count('\t') > 1:
frac_count, _ = safe_float(self.dm_line.split('\t')[2])
if self.check_rows():
self.process_row(frac_count)
self.next_dm_line()
if self.am_linenum % 1000 == 0:
print('=> %d SL and %d TL lines read' %
(self.am_linenum, self.dm_linenum), file=sys.stderr)
def read_files(self, am_fname, dm_fname):
self.am_file = open(am_fname)
self.dm_file = open(dm_fname)
self.reading = True
if self.line_ids:
self.read_files_multi_dm(am_fname, dm_fname)
return
while self.reading:
self.next_am_line()
self.next_dm_line()
if self.reading and self.check_rows():
self.process_row()
if self.am_linenum % 1000 == 0:
print('=> %d lines read' % self.am_linenum, file=sys.stderr)
def process_row(self, frac_count=0):
if self.tokenizer == 'regex':
cur_sl_row = [strip_tags(s, 'sl', space=True) for s in self.am_row]
for i in range(len(self.am_row)):
if self.am_row[i].count('/') > 1:
sl = strip_tags(self.am_row[i], 'sl')
tl = strip_tags(self.dm_row[i], 'tl')
self.process_lu_internal(sl, tl, i, cur_sl_row, frac_count)
elif self.tokenizer == 'biltrans':
cur_sl_row = [x['sl'] for x in self.am_row]
for i in range(len(self.am_row)):
if len(self.am_row[i]['tls']) > 1:
sl = self.am_row[i]['sl']
tl = self.dm_row[i]['tls'][0]
if self.biltrans_wrap_lus:
sl = common.wrap(sl)
tl = common.wrap(tl)
self.process_lu_internal(sl, tl, i, cur_sl_row, frac_count)
def process_lu_internal(self, sl, tl, idx, cur_sl_row, frac_count=0):
if self.count_ngrams:
for j in range(1, self.max_ngrams):
pregram = ' '.join(map(common.wrap, cur_sl_row[idx-j:idx+1]))
postgram = ' '.join(map(common.wrap, cur_sl_row[idx:idx+j+1]))
roundgram = ' '.join(
map(common.wrap, cur_sl_row[idx-j:idx+j+1]))
self.ngrams[sl][pregram][tl] += frac_count
self.ngrams[sl][postgram][tl] += frac_count
self.ngrams[sl][roundgram][tl] += frac_count
self.process_lu(sl, tl, idx, cur_sl_row, frac_count)
def process_lu(self, sl, tl, idx, cur_sl_row, frac_count=0):
pass
def features_and_outline(ngrams, sl, tl, sl_tl, features, indexes,
frac_count=None):
if not ngrams[sl]:
return
meevents = []
for ni in ngrams[sl]:
if ni not in features:
feature_counter = len(features) + 1
features[ni] = feature_counter
meevents.append(features[ni])
if sl not in sl_tl or len(sl_tl[sl]) < 2:
return
outline = str(indexes[(sl, tl)])
if frac_count != None:
outline += ' $ ' + str(int(frac_count * 10000)) + ' '
outline += ' # '
for j in range(len(sl_tl[sl])):
for feature in meevents:
outline += '%s:%s ' % (feature, j)
outline += ' # '
print('%s\t%s\t%s' % (sl, len(sl_tl[sl]), outline))
def read_frequencies(fname):
with open(fname) as fin:
sl_tl = {}
sl_tl_defaults = {}
indexes = {}
trad_counter = defaultdict(lambda: 0)
for line_ in fin.readlines():
line = line_.strip()
if not line:
continue
row = common.tokenize_tagger_line(line)
sl = row[0]
tl = row[1]
fr = float(line.split(' ')[0])
indexes[(sl, tl)] = trad_counter[sl]
trad_counter[sl] += 1
if '@' in line:
sl_tl_defaults[sl] = tl
if fr == 0.0:
print(
'!!! Prolly something went wrong here, the default has freq of 0.0', file=sys.stderr)
else:
print(' %s => %s = %.10f' %
(sl, tl, fr), file=sys.stderr)
else:
sl_tl[sl] = tl
return sl_tl, sl_tl_defaults, indexes
|
