Skip to content

Open-source speech recognition toolkit for training, inference, streaming ASR, VAD, punctuation, speaker diarization pipelines, and OpenAI-compatible/MCP serving.

fix: include runtime package data (#3445)

zhifu gao135915179459
diff --git a/setup.py b/setup.py
index 066137e3a..7875d77e3 100644
--- a/setup.py
+++ b/setup.py
@@ -140,7 +140,16 @@ setup(
long_description_content_type="text/markdown",
license="The MIT License",
packages=find_packages(include=["funasr*"]),
- package_data={"funasr": ["version.txt"]},
+ package_data={
+ "funasr": [
+ "version.txt",
+ "models/sense_voice/whisper_lib/normalizers/english.json",
+ "models/rwkv_bat/cuda_encoder/*.cpp",
+ "models/rwkv_bat/cuda_encoder/*.cu",
+ "models/rwkv_bat/cuda_decoder/*.cpp",
+ "models/rwkv_bat/cuda_decoder/*.cu",
+ ]
+ },
install_requires=install_requires,
setup_requires=setup_requires,
tests_require=tests_require,
diff --git a/tests/test_release_package_data.py b/tests/test_release_package_data.py
new file mode 100644
index 000000000..f849b2b70
--- /dev/null
+++ b/tests/test_release_package_data.py
@@ -0,0 +1,34 @@
+from pathlib import Path
+import subprocess
+import sys
+from zipfile import ZipFile
+
+
+ROOT = Path(__file__).resolve().parents[1]
+RUNTIME_PACKAGE_FILES = {
+ "funasr/models/sense_voice/whisper_lib/normalizers/english.json",
+ "funasr/models/rwkv_bat/cuda_encoder/wkv_cuda.cu",
+ "funasr/models/rwkv_bat/cuda_encoder/wkv_op.cpp",
+ "funasr/models/rwkv_bat/cuda_decoder/wkv_cuda.cu",
+ "funasr/models/rwkv_bat/cuda_decoder/wkv_op.cpp",
+}
+
+
+def test_wheel_includes_runtime_package_data(tmp_path):
+ subprocess.run(
+ [
+ sys.executable,
+ "-c",
+ "import sys; from setuptools import build_meta; "
+ "build_meta.build_wheel(sys.argv[1])",
+ str(tmp_path),
+ ],
+ check=True,
+ cwd=ROOT,
+ )
+ [wheel_path] = tmp_path.glob("*.whl")
+
+ with ZipFile(wheel_path) as wheel:
+ wheel_files = set(wheel.namelist())
+
+ assert RUNTIME_PACKAGE_FILES <= wheel_files