From 66180f65f320a0f8aab39bf5f138d65ebfe52457 Mon Sep 17 00:00:00 2001 From: pengzhendong <275331498@qq.com> Date: Mon, 15 Jun 2026 14:49:30 +0800 Subject: [PATCH] fix: move OOV test cases out of shared test data to fix C++ unit tests The C++ runtime test loads test data from normalizer.txt but uses FSTs built without tag_oov=True. OOV test cases (Korean/Japanese characters) that require wrapping are moved to a separate file only loaded by the Python test (which uses tag_oov=True). Fixes #368 --- tn/chinese/test/data/normalizer.txt | 2 -- tn/chinese/test/data/normalizer_tag_oov.txt | 2 ++ tn/chinese/test/normalizer_test.py | 1 + 3 files changed, 3 insertions(+), 2 deletions(-) create mode 100644 tn/chinese/test/data/normalizer_tag_oov.txt diff --git a/tn/chinese/test/data/normalizer.txt b/tn/chinese/test/data/normalizer.txt index 0ebce9db..64f1cc4c 100644 --- a/tn/chinese/test/data/normalizer.txt +++ b/tn/chinese/test/data/normalizer.txt @@ -35,8 +35,6 @@ 我儿子喜欢这地儿 => 我儿子喜欢这地 O2O => O to O B2B => B to B -我们안녕 => 我们안녕 -雪の花 => 雪の花 给12315打个电话 => 给幺二三幺五打个电话 人均200以内 => 人均两百以内 当场票数≥100万 => 当场票数大于等于一百万 diff --git a/tn/chinese/test/data/normalizer_tag_oov.txt b/tn/chinese/test/data/normalizer_tag_oov.txt new file mode 100644 index 00000000..57af0deb --- /dev/null +++ b/tn/chinese/test/data/normalizer_tag_oov.txt @@ -0,0 +1,2 @@ +我们안녕 => 我们안녕 +雪の花 => 雪の花 \ No newline at end of file diff --git a/tn/chinese/test/normalizer_test.py b/tn/chinese/test/normalizer_test.py index 89def576..40d92117 100644 --- a/tn/chinese/test/normalizer_test.py +++ b/tn/chinese/test/normalizer_test.py @@ -34,6 +34,7 @@ class TestNormalizer: parse_test_case("data/time.txt"), parse_test_case("data/whitelist.txt"), parse_test_case("data/normalizer.txt"), + parse_test_case("data/normalizer_tag_oov.txt"), ) @pytest.mark.parametrize("written, spoken", normalizer_cases)