feat: add transcriber module with faster-whisper integration

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

tests/test_transcriber.py

@@ -0,0 +1,57 @@
+from unittest.mock import MagicMock, patch
+
+import numpy as np
+import pytest
+
+from whisper_local.transcriber import Transcriber
+
+
+class TestTranscriber:
+    @patch("whisper_local.transcriber.WhisperModel")
+    def test_init_loads_model(self, mock_model_class):
+        t = Transcriber(model_name="small", compute_type="int8", language="de")
+        mock_model_class.assert_called_once_with("small", compute_type="int8")
+        assert t.language == "de"
+
+    @patch("whisper_local.transcriber.WhisperModel")
+    def test_transcribe_returns_text(self, mock_model_class):
+        mock_model = MagicMock()
+        mock_segment = MagicMock()
+        mock_segment.text = " Hallo Welt "
+        mock_model.transcribe.return_value = ([mock_segment], None)
+        mock_model_class.return_value = mock_model
+
+        t = Transcriber(model_name="small", compute_type="int8", language="de")
+        audio = np.zeros(16000, dtype=np.float32)
+        result = t.transcribe(audio)
+
+        assert result == "Hallo Welt"
+        mock_model.transcribe.assert_called_once_with(audio, language="de")
+
+    @patch("whisper_local.transcriber.WhisperModel")
+    def test_transcribe_empty_segments(self, mock_model_class):
+        mock_model = MagicMock()
+        mock_model.transcribe.return_value = ([], None)
+        mock_model_class.return_value = mock_model
+
+        t = Transcriber(model_name="small", compute_type="int8", language="de")
+        audio = np.zeros(16000, dtype=np.float32)
+        result = t.transcribe(audio)
+
+        assert result == ""
+
+    @patch("whisper_local.transcriber.WhisperModel")
+    def test_transcribe_multiple_segments(self, mock_model_class):
+        mock_model = MagicMock()
+        seg1 = MagicMock()
+        seg1.text = " Hallo "
+        seg2 = MagicMock()
+        seg2.text = " Welt "
+        mock_model.transcribe.return_value = ([seg1, seg2], None)
+        mock_model_class.return_value = mock_model
+
+        t = Transcriber(model_name="small", compute_type="int8", language="de")
+        audio = np.zeros(16000, dtype=np.float32)
+        result = t.transcribe(audio)
+
+        assert result == "Hallo Welt"