✨ add f5 test cases #176

tests/pipeline/test_chat_tts_pipe.py

@@ -7,6 +7,7 @@
 from modules.core.pipeline.factory import PipelineFactory
 from modules.core.spk.TTSSpeaker import TTSSpeaker
 from tests.pipeline.misc import load_audio, load_audio_wav, save_audio
+from tests.pipeline.voice_clone_pipe import run_voice_clone_pipeline_test
 
 
 @pytest.mark.pipe_chat_tts
@@ -40,30 +41,9 @@ def test_chat_tts_voice_clone_pipe():
     voice_target_path = "./tests/test_inputs/chattts_out1.wav"
     out_audio_path = "./tests/test_outputs/pipe_chat_tts_voice_clone_out1.wav"
 
-    voice_target = load_audio_wav(voice_target_path)
-    voice_spk = TTSSpeaker.from_ref_wav_bytes(
-        ref_wav=voice_target, text="这是一个测试文本。"
+    run_voice_clone_pipeline_test(
+        pipeline_func=PipelineFactory.create_cosyvoice_pipeline,
+        voice_target_path=voice_target_path,
+        voice_target_text="这是一个测试文本。",
+        out_audio_path=out_audio_path,
     )
-
-    pipe0 = PipelineFactory.create_chattts_pipeline(
-        ctx=TTSPipelineContext(
-            text="你好，这里是音色克隆测试~",
-            tts_config=TTSConfig(
-                mid="chat-tts",
-            ),
-            spk=voice_spk,
-            infer_config=InferConfig(eos=" ", sync_gen=True),
-        ),
-    )
-
-    audio_sr, audio_data = pipe0.generate()
-    assert audio_data.dtype == np.float32
-    assert audio_data.size != 0
-    save_audio(
-        #
-        file_path=out_audio_path,
-        audio_data=audio_data,
-        sample_rate=audio_sr,
-    )
-    # 检查文件不为空
-    assert load_audio(out_audio_path)[1].size != 0

tests/pipeline/test_cosyvoice_pipe.py

@@ -7,37 +7,17 @@
 from modules.core.pipeline.factory import PipelineFactory
 from modules.core.spk.TTSSpeaker import TTSSpeaker
 from tests.pipeline.misc import load_audio, load_audio_wav, save_audio
+from tests.pipeline.voice_clone_pipe import run_voice_clone_pipeline_test
 
 
 @pytest.mark.pipe_cosyvoice
 def test_cosy_voice_clone_pipe():
     voice_target_path = "./tests/test_inputs/chattts_out1.wav"
     out_audio_path = "./tests/test_outputs/pipe_cosyvoice_voice_clone_out1.wav"
 
-    voice_target = load_audio_wav(voice_target_path)
-    voice_spk = TTSSpeaker.from_ref_wav_bytes(
-        ref_wav=voice_target, text="这是一个测试文本。"
+    run_voice_clone_pipeline_test(
+        pipeline_func=PipelineFactory.create_cosyvoice_pipeline,
+        voice_target_path=voice_target_path,
+        voice_target_text="这是一个测试文本。",
+        out_audio_path=out_audio_path,
     )
-
-    pipe0 = PipelineFactory.create_cosyvoice_pipeline(
-        ctx=TTSPipelineContext(
-            text="你好，这里是音色克隆测试~",
-            tts_config=TTSConfig(
-                mid="cosy-voice",
-            ),
-            infer_config=InferConfig(eos=" ", sync_gen=True),
-            spk=voice_spk,
-        ),
-    )
-
-    audio_sr, audio_data = pipe0.generate()
-    assert audio_data.dtype == np.float32
-    assert audio_data.size != 0
-    save_audio(
-        #
-        file_path=out_audio_path,
-        audio_data=audio_data,
-        sample_rate=audio_sr,
-    )
-    # 检查文件不为空
-    assert load_audio(out_audio_path)[1].size != 0

tests/pipeline/test_f5_tts_pipe.py

@@ -0,0 +1,17 @@
+import pytest
+
+from modules.core.pipeline.factory import PipelineFactory
+from tests.pipeline.voice_clone_pipe import run_voice_clone_pipeline_test
+
+
+@pytest.mark.pipe_f5_tts
+def test_f5_voice_clone_pipe():
+    voice_target_path = "./tests/test_inputs/chattts_out1.wav"
+    out_audio_path = "./tests/test_outputs/pipe_f5_voice_clone_out1.wav"
+
+    run_voice_clone_pipeline_test(
+        pipeline_func=PipelineFactory.create_f5_tts_pipeline,
+        voice_target_path=voice_target_path,
+        voice_target_text="这是一个测试文本。",
+        out_audio_path=out_audio_path,
+    )

tests/pipeline/test_fire_red_tts_pipe.py

@@ -1,43 +1,17 @@
-import numpy as np
 import pytest
 
-from modules.core.handler.datacls.audio_model import AdjustConfig
-from modules.core.handler.datacls.tts_model import InferConfig, TTSConfig
-from modules.core.pipeline.dcls import TTSPipelineContext
 from modules.core.pipeline.factory import PipelineFactory
-from modules.core.spk.TTSSpeaker import TTSSpeaker
-from tests.pipeline.misc import load_audio, load_audio_wav, save_audio
+from tests.pipeline.voice_clone_pipe import run_voice_clone_pipeline_test
 
 
 @pytest.mark.pipe_fire_red_tts
 def test_fire_red_tts_clone_pipe():
     voice_target_path = "./tests/test_inputs/chattts_out1.wav"
     out_audio_path = "./tests/test_outputs/pipe_fire_red_tts_voice_clone_out1.wav"
 
-    voice_target = load_audio_wav(voice_target_path)
-    voice_spk = TTSSpeaker.from_ref_wav_bytes(
-        ref_wav=voice_target, text="这是一个测试文本。"
+    run_voice_clone_pipeline_test(
+        pipeline_func=PipelineFactory.create_fire_red_tts_pipeline,
+        voice_target_path=voice_target_path,
+        voice_target_text="这是一个测试文本。",
+        out_audio_path=out_audio_path,
     )
-
-    pipe0 = PipelineFactory.create_fire_red_tts_pipeline(
-        ctx=TTSPipelineContext(
-            text="你好，这里是音色克隆测试~",
-            tts_config=TTSConfig(
-                mid="fire-red-tts",
-            ),
-            infer_config=InferConfig(eos=" ", sync_gen=True),
-            spk=voice_spk,
-        ),
-    )
-
-    audio_sr, audio_data = pipe0.generate()
-    assert audio_data.dtype == np.float32
-    assert audio_data.size != 0
-    save_audio(
-        #
-        file_path=out_audio_path,
-        audio_data=audio_data,
-        sample_rate=audio_sr,
-    )
-    # 检查文件不为空
-    assert load_audio(out_audio_path)[1].size != 0

tests/pipeline/voice_clone_pipe.py

@@ -0,0 +1,56 @@
+from typing import Callable
+import numpy as np
+import pytest
+
+from modules.core.handler.datacls.audio_model import AdjustConfig
+from modules.core.handler.datacls.tts_model import InferConfig, TTSConfig
+from modules.core.pipeline.dcls import TTSPipelineContext
+from modules.core.pipeline.factory import PipelineFactory
+from modules.core.pipeline.pipeline import AudioPipeline
+from modules.core.spk.TTSSpeaker import TTSSpeaker
+from tests.pipeline.misc import load_audio, load_audio_wav, save_audio
+
+
+# Pipeline function type, which returns a pipeline with a generate() method
+PipelineFunc = Callable[
+    [TTSPipelineContext], "AudioPipeline"
+]  # Type hint for pipeline function
+
+
+def run_voice_clone_pipeline_test(
+    pipeline_func: PipelineFunc,
+    voice_target_path: str,
+    voice_target_text: str,
+    out_audio_path: str,
+    text: str = "你好，这里是音色克隆测试~",
+) -> None:
+    """辅助函数：执行音色克隆测试并保存结果音频。
+
+    参数：
+    - pipeline_func: 用于创建 TTS 流水线的函数。
+    - voice_target_path: 目标语音的文件路径。
+    - voice_target_text: 目标语音的文本。
+    - out_audio_path: 输出音频的文件路径。
+    - text: 生成音频时使用的文本。
+    """
+    voice_target: bytes = load_audio_wav(voice_target_path)
+    voice_spk: TTSSpeaker = TTSSpeaker.from_ref_wav_bytes(
+        ref_wav=voice_target, text=voice_target_text
+    )
+
+    pipe: AudioPipeline = pipeline_func(
+        ctx=TTSPipelineContext(
+            text=text,
+            tts_config=TTSConfig(mid="cosy-voice"),
+            infer_config=InferConfig(eos=" ", sync_gen=True),
+            spk=voice_spk,
+        ),
+    )
+
+    audio_sr, audio_data = pipe.generate()  # Tuple[int, np.ndarray]
+    assert audio_data.dtype == np.float32
+    assert audio_data.size != 0
+
+    save_audio(file_path=out_audio_path, audio_data=audio_data, sample_rate=audio_sr)
+    # 检查文件不为空
+    assert load_audio(out_audio_path)[1].size != 0