|
1
2
3
4
5
6
7
|
'''
Created on Nov 7, 2013
@author: mlenart
'''
import codecs
|
|
8
|
import logging
|
|
9
10
11
|
class Lemma(object):
|
|
12
|
def __init__(self, cutLength, suffixToAdd, casePattern):
|
|
13
14
|
self.cutLength = cutLength
self.suffixToAdd = suffixToAdd
|
|
15
|
self.casePattern = casePattern
|
|
16
17
18
|
class Interpretation(object):
|
|
19
|
def __init__(self, orth, base, tagnum, namenum, typenum):
|
|
20
21
22
23
|
assert type(orth) == unicode
assert type(base) == unicode
root = u''
for o, b in zip(orth, base):
|
|
24
|
if o.lower() == b.lower():
|
|
25
|
root += b
|
|
26
27
|
else:
break
|
|
28
|
cutLength = len(orth) - len(root)
|
|
29
30
|
self.lemma = Lemma(
cutLength=cutLength,
|
|
31
32
|
suffixToAdd=base[len(root):],
casePattern = [c == c.upper() for c in root])
|
|
33
34
|
self.tagnum = tagnum
self.namenum = namenum
|
|
35
|
self.typenum = typenum
|
|
36
37
|
def getSortKey(self):
|
|
38
39
40
41
42
43
|
return (
self.lemma.cutLength,
tuple(self.lemma.suffixToAdd),
tuple(self.lemma.casePattern),
self.tagnum,
self.namenum)
|
|
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
|
def __eq__(self, other):
if isinstance(other, Interpretation):
return self.getSortKey() == other.getSortKey()
else:
return False
def __hash__(self):
return hash(self.getSortKey())
class Tagset(object):
TAGS = 1
NAMES = 2
SEP = '\t'
def __init__(self, filename, encoding='utf8'):
self.tag2tagnum = {}
self.name2namenum = {}
self._doInit(filename, encoding)
print self.tag2tagnum
print self.name2namenum
def _doInit(self, filename, encoding):
addingTo = None
with codecs.open(filename, 'r', encoding) as f:
for line in f:
line = line.strip('\n')
if line == u'[TAGS]':
addingTo = Tagset.TAGS
elif line == u'[NAMES]':
addingTo = Tagset.NAMES
elif line and not line.startswith(u'#'):
assert addingTo in [Tagset.TAGS, Tagset.NAMES]
res = {Tagset.TAGS: self.tag2tagnum,
Tagset.NAMES: self.name2namenum}[addingTo]
tagNum = line.split(Tagset.SEP)[0]
tag = line.split(Tagset.SEP)[1]
assert tag not in res
res[tag] = int(tagNum)
|