mirror of
https://github.com/mii443/akaza.git
synced 2025-08-30 10:39:27 +00:00
39 lines
1.1 KiB
Python
39 lines
1.1 KiB
Python
from tempfile import NamedTemporaryFile
|
|
|
|
import pytest
|
|
|
|
from comb.engine import Comb
|
|
from comb.system_dict import SystemDict
|
|
from comb.user_language_model import UserLanguageModel
|
|
|
|
tmpfile = NamedTemporaryFile(delete=False)
|
|
user_language_model = UserLanguageModel(tmpfile.name)
|
|
system_dict = SystemDict()
|
|
|
|
comb = Comb(user_language_model=user_language_model, system_dict=system_dict)
|
|
|
|
|
|
@pytest.mark.parametrize('src, expected', [
|
|
# Wnn で有名なフレーズ。
|
|
('わたしのなまえはなかのです', '私の名前は中野です'),
|
|
# カタカナ語の処理が出来ていること。
|
|
('わーど', 'ワード'),
|
|
('にほん', '日本'),
|
|
('それなwww', 'それなwww'),
|
|
('siinn', '子音'),
|
|
('zh', '←'),
|
|
('IME', 'IME'),
|
|
])
|
|
def test_wnn(src, expected):
|
|
clauses = comb.convert2(src)
|
|
got = ''.join([clause[0].word for clause in clauses])
|
|
assert got == expected
|
|
|
|
|
|
def test_wnn2():
|
|
clauses = comb.convert2("わたし")
|
|
hiragana_len = len([True for node in clauses[0] if node.word == 'わたし'])
|
|
for node in clauses[0]:
|
|
print(node)
|
|
assert hiragana_len == 1
|