-
Notifications
You must be signed in to change notification settings - Fork 0
/
checker.py
200 lines (188 loc) · 10.2 KB
/
checker.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
import numpy as np
import pandas as pd
from gd_tools.acainn import Morphology
class Checker():
# for simple matches
def _list_to_df(self, tokens, code):
return pd.DataFrame({
"token": tokens,
"code": np.resize([code], len(tokens))
})
def _make_df(self, tagged_tokens):
tokens = [t[0] for t in tagged_tokens]
postags = [t[1] for t in tagged_tokens]
t_1 = ['<START>'] + tokens[:-1]
p_1 = ['<START>'] + postags[:-1]
t_2 = ['<START>'] + t_1[:-1]
p_2 = ['<START>'] + p_1[:-1]
t1 = tokens[1:] + ['<END>']
t2 = t1[1:] + ['<END>']
p1 = postags[1:] + ['<END>']
p2 = p1[1:] + ['<END>']
return pd.DataFrame(
{ 'token': tokens, 'pos': postags, '_t_1': t_1, '_p_1': p_1,
'_t_2': t_2, '_p_2': p_2, '_t1': t1, '_p1': p1, '_t2': t2, '_p2': p2 }
)
def _feats(self, df):
return df.assign(
_lenited = lambda x: self.l.lenited_pd(x.token.str),
_chalenited = lambda x: self.l.chalenited_pd(x.token.str),
_nondentallenited = lambda x: self.l.ndlenited_pd(x.token.str),
_genitive = lambda x: x.pos.str.match('N.*g'),
_sing = lambda x: x.pos.str.match('N.s'),
_genitivesing = lambda x: x.pos.str.match('N.s.g'),
_pl = lambda x: x.pos.str.match('N.p'),
_c0 = lambda x: x.token.str[0],
_coarsepos = lambda x: x.pos.str[0],
_acute = lambda x: self._acutes(x.token.str),
code = lambda x: '')
def __init__(self):
self.l = Morphology()
hyphen_series = ["a màireach", "a nis", "a nochd", "a raoir", "a rithist", "am bliadhna",
"an ceartuair", "an dè", "an diugh", "an dràsta", "an earar", "an-uiridh",
"a bhàn", "a bhos", "an àird", "a nall", "a nìos", "a nuas", "a null",
"a chaoidh", "a cheana", "am feast", "a mhàin", "a riamh", "a mach",
"a muigh", "a staigh", "a steach"]
personal_number_series = ["dithis", "tri", "ceathrar", "cignear",
"sianar", "seachdnar", "ochdnar", "naoinear",
"deichnear"]
self.personal_number_adjs = pd.DataFrame({
"_t_1": personal_number_series,
"_coarsepos": np.resize(["A"], len(personal_number_series)),
"_lenited": [False, False, True, True, True, True, True, True, True],
"code": np.resize(["144ii"], len(personal_number_series))
})
self.personal_number_nouns = pd.DataFrame({
"_t_1": personal_number_series,
"_coarsepos": np.resize(["N"], len(personal_number_series)),
"_lenited": np.resize([False], len(personal_number_series)),
"code": np.resize(["144iii"], len(personal_number_series))
})
self.acutes = pd.DataFrame({"_acute": [True], "code":["GOC-ACUTE"]})
self.hyphens = self._list_to_df(hyphen_series, "GOC-HYPHEN")
self.apos = self._list_to_df(["Sann", "Se", "sann", "se"], "GOC-APOS")
self.noapos = self._list_to_df(
["de'n", "do'n", "fo'n", "mu'n", "ro'n", "tro'n"],
"GOC-NOAPOS")
self.micheart = pd.DataFrame({
"token": ['radh', 'cearr'],
"code": ["spelling", "spelling"],
'message': ['should be ràdh', 'should be ceàrr' ]})
lenite_Ar_series = ["deagh", "droch"]
self.lenite_Ar = pd.DataFrame({
"_t_1": lenite_Ar_series,
"code": np.resize(["LENITE"], len(lenite_Ar_series)),
"_lenited": np.resize([False], len(lenite_Ar_series))
})
self.messages = {
"45iaepsilon": "Nouns lenite after mo, do, a (masculine). Cox §45iaε",
"45iaepsilon-": "Nouns do not lenite after a (feminine), ar, ur, an. Cox §45iaε",
"44iiia": "Independent forms in the past and mixed tenses lenite: Cox §44iiia,§44iiib/§246ii",
"44iiia-":"Dependent forms in the past and mixed tenses do not lenite.",
# thinking about how to implement Cox §44iiic/§45iaα.
"45iabeta": "Masculine names in the genitive lenite: Cox §45iaβ",
"45iabeta-": "Feminine names in the genitive do not usually lenite: Cox §45iaβ",
# Cox §45iaδ is complicated to do on the basis of tagging.
# consider doing based on the surface form
"45iazeta": "Nouns lenite after aon, dà, dhà. Cox §45iaζ",
# Cox distinguishes air, air^s, and air^n. How to deal with this?
"45iaiota": "nouns lenite after prepositions bho, o, de, do, eadar, fo, gun, mu, ro, thar tre and tro. Cox §45iaι",
"45iealpha": "verbs lenite after ma and relative-form verbs lenite after a. Cox §45ieα",
"45iebeta": "past-tense verbs lenite after do. Cox §45ieβ",
"45iegamma": "verbs immediately after cha lenite. Cox §45ieγ",
"45iia": "verbs beginning with f after a question particle lenite. Cox §45iia",
"170": "barrachd takes the genitive singular: Cox §170",
"173": "iomadh is followed by a singular noun: Cox §173",
"176":"tuilleadh takes the genitive singular: Cox §176"
}
self.lenitesp_1 = pd.DataFrame({
"_p_1": ["Dp1s", "Dp2s", "Dp3sm"],
"_lenited": [False, False, False],
"code": ["45iaepsilon","45iaepsilon","45iaepsilon"]})
self.nolenitesp_1 = pd.DataFrame({
"_p_1": ["Dp3sf", "Dp1p", "Dp2p", "Dp3p"],
"_lenited": [True, True, True, True],
"code": ["45iaepsilon-","45iaepsilon-","45iaepsilon-","45iaepsilon-"]})
self.lenitepos = pd.DataFrame({
"pos": ["V-h", "V-s", "V-h1s", "Nn-mg","V-h--d","V-s--d", "Nn-fg"],
"_lenited": [False, False, False, False,True,True,True],
"code": ["44iiia","44iiia", "44iiia", "45iabeta","44iiia-","44iiia-","45iabeta-"]})
self.nd_lenite_t_1 = pd.DataFrame({
"_t_1":["aon"],
"_nondentallenited": [False],
"code": ["45iazeta"]})
self.lenite_t_1 = pd.DataFrame({
"_t_1": ["dà", "dhà"],
"_lenited":[False,False],
"code": ["45iazeta","45iazeta"]})
self.lenite_p_1_t_1 = pd.DataFrame({
"_p_1": np.resize(["Sp"],12),
"_lenited": np.resize([False],12),
"_t_1": ["bho", "o","de","do","eadar","fo","gun","mu","ro","thar","tre","tro"],
"code": np.resize(["45iaiota"],12)})
# check for verb here?
self.lenite_p_1_t_1a = pd.DataFrame({
"_p_1": ["Cs", "Q-r", "Q--s"],
"_t_1": ["ma", "a", "do"],
"_lenited":[False,False,False],
"code":["45iealpha","45iealpha","45iebeta"]})
self.chalenite_p_1_t_1 = pd.DataFrame({
"_p_1": ["Qn"],
"_t_1": ["cha"],
"_chalenited": [False],
"code":["45iegamma"]})
self.fh = pd.DataFrame({
"_p_1": ["Qq"],
"_c0": ["f"],
"_lenited": [False],
"code": ["45iia"]})
self.t_1_case = pd.DataFrame({
"_t_1": ["barrachd","tuilleadh"],
"_genitivesing": [False,False],
"code": ["170","176"]})
self.t_1_number = pd.DataFrame({
"_t_1": ["iomadh"],
"_pl": [True],
"code": ["173"]})
self.lantest = pd.DataFrame({
"_t_1":["làn"], "_genitive": False, "code":"174"})
self.genitest = pd.DataFrame({
"_t_1":["chum"], "_p_1":"Sp", "_genitive":False, "code":"344"})
def _checkdf(self, df):
result = (df.
merge(self.lenitesp_1, on=("_p_1","_lenited"), how="left", suffixes = ("-a","-b")).
merge(self.nolenitesp_1, on =("_p_1","_lenited"), how="left").
merge(self.lenitepos, on=("pos","_lenited"), how="left", suffixes = ("-c","-d")).
merge(self.lenite_t_1, on=("_t_1","_lenited"), how="left").
merge(self.nd_lenite_t_1, on=("_t_1","_nondentallenited"),how="left", suffixes = ("-e","-f")).
merge(self.lenite_p_1_t_1, on=("_p_1","_t_1","_lenited"), how="left").
merge(self.lenite_p_1_t_1a, on=("_p_1","_t_1","_lenited"), how="left", suffixes = ("-g","-h")).
merge(self.chalenite_p_1_t_1, on=("_p_1","_t_1","_chalenited"), how="left").
merge(self.fh, on=("_p_1","_c0", "_lenited"), how="left", suffixes = ("-i", "-j")).
merge(self.t_1_case, on=("_t_1","_genitivesing"), how="left").
merge(self.t_1_number, on=("_t_1","_pl"), how="left", suffixes = ("-k", "-l")).
merge(self.lantest, on=("_t_1","_genitive"), how="left").
merge(self.genitest, on=("_t_1","_p_1","_genitive"), how="left", suffixes = ("-m", "-n")).
merge(self.micheart, on="token", how="left").
merge(self.hyphens, on="token", how="left", suffixes = ("-o", "-p")).
merge(self.apos, on="token", how="left").
merge(self.lenite_Ar, on=("_t_1","_lenited"), how="left", suffixes = ("-q","-r")).
merge(self.acutes, on=("_acute"), how="left").
merge(self.personal_number_adjs, on=("_t_1", "_coarsepos", "_lenited"), how="left", suffixes = ("-s","-t")).
merge(self.personal_number_nouns, on=("_t_1", "_coarsepos", "_lenited"), how="left").
merge(self.personal_number_nouns, on=("_t_1", "_coarsepos", "_lenited"), how="left", suffixes = ("-u","-v"))
)
result = result.fillna('')
result['code'] = result.filter(regex="code-").agg(lambda x:",".join(x),axis="columns")
return result.drop(columns = result.filter(regex="[_-]"))
def _acutes(self, form):
return form.contains(r"[úóíéá]")
def check(self, text):
"""Tokenises and tags using Edinburgh code. Consider udpipe?"""
tokens = self.t.tokenise(text)
tagged_tokens = self.p.tagfile_default(tokens)
return self.check_tokens(tagged_tokens)
def check_tokens(self, tagged_tokens):
"""Expects a list of 2-tuples of form and xpos."""
df = self._feats(self._make_df(tagged_tokens))
return self._checkdf(df)