Commit
·
84d2513
1
Parent(s):
9ab6f28
commit files to HF hub
Browse files- mecab_tokenizer.py +28 -15
mecab_tokenizer.py
CHANGED
|
@@ -14,9 +14,9 @@ class MeCabResult(NamedTuple):
|
|
| 14 |
hinshi_saibunrui_3: str
|
| 15 |
katsuyokei_1: str
|
| 16 |
katsuyokei_2: str
|
| 17 |
-
genkei: str
|
| 18 |
-
yomi: str
|
| 19 |
-
hatsuon: str
|
| 20 |
|
| 21 |
|
| 22 |
class MeCabTokenizer(PreTrainedTokenizer):
|
|
@@ -34,9 +34,9 @@ class MeCabTokenizer(PreTrainedTokenizer):
|
|
| 34 |
|
| 35 |
self.target_hinshi = hinshi
|
| 36 |
if mecab_dicdir is not None:
|
| 37 |
-
self.mecab = MeCab.Tagger(f"-d {mecab_dicdir}")
|
| 38 |
else:
|
| 39 |
-
self.mecab = MeCab.Tagger()
|
| 40 |
|
| 41 |
super().__init__(**kwargs)
|
| 42 |
|
|
@@ -76,17 +76,30 @@ class MeCabTokenizer(PreTrainedTokenizer):
|
|
| 76 |
Returns:
|
| 77 |
list[MeCabResult]: MeCabの解析結果
|
| 78 |
"""
|
| 79 |
-
|
| 80 |
#形態素1つ1つを処理
|
| 81 |
out = []
|
| 82 |
-
|
| 83 |
-
args =
|
| 84 |
-
args
|
| 85 |
-
|
| 86 |
-
|
| 87 |
-
|
| 88 |
-
|
| 89 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 90 |
out.append(mecab_result)
|
| 91 |
-
node = node.next # 最後のEOSを省く
|
| 92 |
return out
|
|
|
|
| 14 |
hinshi_saibunrui_3: str
|
| 15 |
katsuyokei_1: str
|
| 16 |
katsuyokei_2: str
|
| 17 |
+
genkei: str = ""
|
| 18 |
+
yomi: str = ""
|
| 19 |
+
hatsuon: str = ""
|
| 20 |
|
| 21 |
|
| 22 |
class MeCabTokenizer(PreTrainedTokenizer):
|
|
|
|
| 34 |
|
| 35 |
self.target_hinshi = hinshi
|
| 36 |
if mecab_dicdir is not None:
|
| 37 |
+
self.mecab = MeCab.Tagger(f"-d {mecab_dicdir} -O '' -F '%m,%H\n'")
|
| 38 |
else:
|
| 39 |
+
self.mecab = MeCab.Tagger("-O '' -F '%m,%H\n'")
|
| 40 |
|
| 41 |
super().__init__(**kwargs)
|
| 42 |
|
|
|
|
| 76 |
Returns:
|
| 77 |
list[MeCabResult]: MeCabの解析結果
|
| 78 |
"""
|
| 79 |
+
nodes = self.mecab.parse(text).split("\n")
|
| 80 |
#形態素1つ1つを処理
|
| 81 |
out = []
|
| 82 |
+
for node in nodes:
|
| 83 |
+
args = node.split(",")
|
| 84 |
+
if args[0] in ["EOS", ""]:
|
| 85 |
+
continue
|
| 86 |
+
# 辞書によって異なる
|
| 87 |
+
if len(args) == 10:
|
| 88 |
+
mecab_result = MeCabResult(args[0], args[1], args[2], args[3],
|
| 89 |
+
args[4], args[5], args[6], args[7],
|
| 90 |
+
args[8], args[9])
|
| 91 |
+
elif len(args) == 7:
|
| 92 |
+
# 英語
|
| 93 |
+
mecab_result = MeCabResult(args[0], args[1], args[2], args[3],
|
| 94 |
+
args[4], args[5], args[6], "*", "*",
|
| 95 |
+
"*")
|
| 96 |
+
elif len(args) == 27:
|
| 97 |
+
# 補助記号
|
| 98 |
+
mecab_result = MeCabResult(args[0], args[1], args[2], args[3],
|
| 99 |
+
args[4], args[5], args[6], "*", "*",
|
| 100 |
+
"*")
|
| 101 |
+
else:
|
| 102 |
+
raise NotImplementedError(
|
| 103 |
+
f"unsupposed parse args_length: {len(args)}, args: {args}")
|
| 104 |
out.append(mecab_result)
|
|
|
|
| 105 |
return out
|