pushed new nfkc_cf.tsv
This commit is contained in:
Родитель
20469aa882
Коммит
465a419250
|
@ -1,7 +1,3 @@
|
||||||
9 20 # =>
|
|
||||||
A 20 # =>
|
|
||||||
C 20 # =>
|
|
||||||
D 20 # =>
|
|
||||||
3C 338 226E # ≮ => ≮
|
3C 338 226E # ≮ => ≮
|
||||||
3D 338 2260 # ≠ => ≠
|
3D 338 2260 # ≠ => ≠
|
||||||
3E 338 226F # ≯ => ≯
|
3E 338 226F # ≯ => ≯
|
||||||
|
@ -57024,7 +57020,6 @@ FB9 F90 FB5 # ྐྵ => ྐྵ
|
||||||
13FB 13F3 # ᏻ => Ᏻ
|
13FB 13F3 # ᏻ => Ᏻ
|
||||||
13FC 13F4 # ᏼ => Ᏼ
|
13FC 13F4 # ᏼ => Ᏼ
|
||||||
13FD 13F5 # ᏽ => Ᏽ
|
13FD 13F5 # ᏽ => Ᏽ
|
||||||
1680 20 # =>
|
|
||||||
1B05 1B35 1B06 # ᬆ => ᬆ
|
1B05 1B35 1B06 # ᬆ => ᬆ
|
||||||
1B07 1B35 1B08 # ᬈ => ᬈ
|
1B07 1B35 1B08 # ᬈ => ᬈ
|
||||||
1B09 1B35 1B0A # ᬊ => ᬊ
|
1B09 1B35 1B0A # ᬊ => ᬊ
|
||||||
|
@ -57948,16 +57943,11 @@ FB9 F90 FB5 # ྐྵ => ྐྵ
|
||||||
2008 20 # =>
|
2008 20 # =>
|
||||||
2009 20 # =>
|
2009 20 # =>
|
||||||
200A 20 # =>
|
200A 20 # =>
|
||||||
200B 20 # =>
|
|
||||||
200E 20 # =>
|
|
||||||
200F 20 # =>
|
|
||||||
2011 2010 # ‑ => ‐
|
2011 2010 # ‑ => ‐
|
||||||
2017 20 333 # ‗ => ̳
|
2017 20 333 # ‗ => ̳
|
||||||
2024 2E # ․ => .
|
2024 2E # ․ => .
|
||||||
2025 2E 2E # ‥ => ..
|
2025 2E 2E # ‥ => ..
|
||||||
2026 2E 2E 2E # … => ...
|
2026 2E 2E 2E # … => ...
|
||||||
2028 20 #
=>
|
|
||||||
2029 20 #
=>
|
|
||||||
202F 20 # =>
|
202F 20 # =>
|
||||||
2033 2032 2032 # ″ => ′′
|
2033 2032 2032 # ″ => ′′
|
||||||
2034 2032 2032 2032 # ‴ => ′′′
|
2034 2032 2032 2032 # ‴ => ′′′
|
||||||
|
@ -59598,7 +59588,6 @@ FB9 F90 FB5 # ྐྵ => ྐྵ
|
||||||
24E9 331 1E95 # ⓩ̱ => ẕ
|
24E9 331 1E95 # ⓩ̱ => ẕ
|
||||||
24E9 341 17A # ⓩ́ => ź
|
24E9 341 17A # ⓩ́ => ź
|
||||||
24EA 30 # ⓪ => 0
|
24EA 30 # ⓪ => 0
|
||||||
2581 20 # ▁ =>
|
|
||||||
2A0C 222B 222B 222B 222B # ⨌ => ∫∫∫∫
|
2A0C 222B 222B 222B 222B # ⨌ => ∫∫∫∫
|
||||||
2A74 3A 3A 3D # ⩴ => ::=
|
2A74 3A 3A 3D # ⩴ => ::=
|
||||||
2A75 3D 3D # ⩵ => ==
|
2A75 3D 3D # ⩵ => ==
|
||||||
|
@ -157901,7 +157890,6 @@ FEF9 644 625 # ﻹ => لإ
|
||||||
FEFA 644 625 # ﻺ => لإ
|
FEFA 644 625 # ﻺ => لإ
|
||||||
FEFB 644 627 # ﻻ => لا
|
FEFB 644 627 # ﻻ => لا
|
||||||
FEFC 644 627 # ﻼ => لا
|
FEFC 644 627 # ﻼ => لا
|
||||||
FEFF 20 # =>
|
|
||||||
FF01 21 # ! => !
|
FF01 21 # ! => !
|
||||||
FF02 22 # " => "
|
FF02 22 # " => "
|
||||||
FF03 23 # # => #
|
FF03 23 # # => #
|
||||||
|
@ -158584,6 +158572,7 @@ FF5A 341 17A # ź => ź
|
||||||
FF5B 7B # { => {
|
FF5B 7B # { => {
|
||||||
FF5C 7C # | => |
|
FF5C 7C # | => |
|
||||||
FF5D 7D # } => }
|
FF5D 7D # } => }
|
||||||
|
FF5E 7E # ~ => ~
|
||||||
FF5F 2985 # ⦅ => ⦅
|
FF5F 2985 # ⦅ => ⦅
|
||||||
FF60 2986 # ⦆ => ⦆
|
FF60 2986 # ⦆ => ⦆
|
||||||
FF61 3002 # 。 => 。
|
FF61 3002 # 。 => 。
|
||||||
|
@ -213835,7 +213824,6 @@ FFEB 338 219B # ↛ => ↛
|
||||||
FFEC 2193 # ↓ => ↓
|
FFEC 2193 # ↓ => ↓
|
||||||
FFED 25A0 # ■ => ■
|
FFED 25A0 # ■ => ■
|
||||||
FFEE 25CB # ○ => ○
|
FFEE 25CB # ○ => ○
|
||||||
FFFD 20 # <20> =>
|
|
||||||
10400 10428 # 𐐀 => 𐐨
|
10400 10428 # 𐐀 => 𐐨
|
||||||
10401 10429 # 𐐁 => 𐐩
|
10401 10429 # 𐐁 => 𐐩
|
||||||
10402 1042A # 𐐂 => 𐐪
|
10402 1042A # 𐐂 => 𐐪
|
||||||
|
|
Не удается отобразить этот файл, потому что он слишком большой.
|
|
@ -160,7 +160,7 @@ int main(int argc, char **argv) {
|
||||||
std::pair<std::string, std::function<Status(Builder::CharsMap *)>>>
|
std::pair<std::string, std::function<Status(Builder::CharsMap *)>>>
|
||||||
kRuleList = {{"nfkc", Builder::BuildNFKCMap},
|
kRuleList = {{"nfkc", Builder::BuildNFKCMap},
|
||||||
{"nmt_nfkc", Builder::BuildNmtNFKCMap},
|
{"nmt_nfkc", Builder::BuildNmtNFKCMap},
|
||||||
{"nfkc_cf", Builder::BuildNmtNFKC_CFMap},
|
{"nfkc_cf", Builder::BuildNFKC_CFMap},
|
||||||
{"nmt_nfkc_cf", Builder::BuildNmtNFKC_CFMap}};
|
{"nmt_nfkc_cf", Builder::BuildNmtNFKC_CFMap}};
|
||||||
|
|
||||||
std::vector<std::pair<std::string, std::string>> data;
|
std::vector<std::pair<std::string, std::string>> data;
|
||||||
|
|
|
@ -16,6 +16,7 @@
|
||||||
|
|
||||||
from setuptools import setup
|
from setuptools import setup
|
||||||
from setuptools import find_packages
|
from setuptools import find_packages
|
||||||
|
import tensorflow as tf
|
||||||
import codecs
|
import codecs
|
||||||
import string
|
import string
|
||||||
import sys
|
import sys
|
||||||
|
@ -26,6 +27,12 @@ sys.path.append(os.path.join('.', 'test'))
|
||||||
with codecs.open(os.path.join('..', 'VERSION'), 'r', 'utf-8') as f:
|
with codecs.open(os.path.join('..', 'VERSION'), 'r', 'utf-8') as f:
|
||||||
version = f.read()
|
version = f.read()
|
||||||
|
|
||||||
|
tf_version = tf.__version__
|
||||||
|
if tf_version[0] == '.':
|
||||||
|
tf_version = tf_version[1:]
|
||||||
|
|
||||||
|
version = tf_version + '.' + version
|
||||||
|
|
||||||
setup(name = 'tf_sentencepiece',
|
setup(name = 'tf_sentencepiece',
|
||||||
author = 'Taku Kudo',
|
author = 'Taku Kudo',
|
||||||
author_email='taku@google.com',
|
author_email='taku@google.com',
|
||||||
|
|
Загрузка…
Ссылка в новой задаче