Add test script

tests/test_config.py

@@ -0,0 +1,9 @@
+import os
+
+TEST_FILE_DOWNLOAD_URL = "http://www.voiptroubleshooter.com/open_speech/american/OSR_us_000_0010_8k.wav"
+TEST_YOUTUBE_URL = "https://www.youtube.com/watch?v=4WEQtgnBu0I&ab_channel=AndriaFitzer"
+TEST_WHISPER_MODEL = "tiny"
+TEST_UVR_MODEL = "UVR-MDX-NET-Inst_HQ_4"
+TEST_NLLB_MODEL = "facebook/nllb-200-1.3B"
+TEST_SUBTITLE_PATH = os.path.join("translation_test.srt")
+

tests/transcription.py

@@ -0,0 +1,47 @@
+from modules.whisper.whisper_factory import WhisperFactory
+from modules.whisper.whisper_parameter import WhisperValues
+from test_config import *
+
+import pytest
+import gradio as gr
+import os
+
+
+@pytest.mark.parametrize("whisper_type", ["whisper", "faster-whisper", "insanely_fast_whisper"])
+def test_transcribe(whisper_type: str):
+    audio_path = os.path.join("test.wav")
+    if not os.path.exists(audio_path):
+        download_file(TEST_FILE_DOWNLOAD_URL, audio_path)
+
+    whisper_inferencer = WhisperFactory.create_whisper_inference(
+        whisper_type=whisper_type,
+    )
+
+    print("Device : ", whisper_inferencer.device)
+
+    hparams = WhisperValues(
+        model_size=TEST_WHISPER_MODEL,
+    ).as_list()
+
+    whisper_inferencer.transcribe_file(
+        files=[audio_path],
+        progress=gr.Progress(),
+        *hparams,
+    )
+
+    whisper_inferencer.transcribe_youtube(
+        youtube_link=TEST_YOUTUBE_URL,
+        progress=gr.Progress(),
+        *hparams,
+    )
+
+    whisper_inferencer.transcribe_mic(
+        mic_audio=audio_path,
+        progress=gr.Progress(),
+        *hparams,
+    )
+
+
+def download_file(url: str, path: str):
+    if not os.path.exists(path):
+        os.system(f"wget {url} -O {path}")

tests/translation_test.srt

@@ -0,0 +1,7 @@
+1
+00:00:00,000 --> 00:00:02,240
+You've got
+
+2
+00:00:02,240 --> 00:00:04,160
+a friend in me.