feat: add progress callback to transcription with Rich progress bar in CLI

Adds an optional on_progress(ratio: float) callback to WhisperTranscriber.transcribe()
and _run_whisper(), and wires it to a Rich live progress bar in main.py so users
see incremental feedback instead of silence during long transcriptions.

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

Author: 2 people

tests/test_transcriber.py

@@ -92,3 +92,36 @@ def test_whisper_model_cached_across_instances():
 
     assert mock_wm.call_count == 1
     assert t1._model is t2._model
+
+
+def test_transcribe_calls_progress_callback(tmp_path):
+    """on_progress callback should be called with values between 0.0 and 1.0."""
+    import soundfile as sf
+    from unittest.mock import patch, MagicMock
+    import tinysteno.transcriber as mod
+
+    audio = np.zeros(32000, dtype=np.float32)  # 2 seconds at 16kHz
+    wav_path = tmp_path / "t.wav"
+    sf.write(str(wav_path), audio, 16000)
+
+    seg1 = MagicMock()
+    seg1.text = " hello"
+    seg1.start = 0.5
+
+    seg2 = MagicMock()
+    seg2.text = " world"
+    seg2.start = 1.5
+
+    progress_values = []
+
+    mod._MODEL_CACHE.clear()
+    with patch("tinysteno.transcriber.WhisperModel") as mock_wm:
+        mock_model = MagicMock()
+        mock_model.transcribe.return_value = (iter([seg1, seg2]), MagicMock(language="en"))
+        mock_wm.return_value = mock_model
+        t = mod.WhisperTranscriber()
+
+    t.transcribe(str(wav_path), on_progress=lambda r: progress_values.append(r))
+
+    assert len(progress_values) >= 1
+    assert all(0.0 <= v <= 1.0 for v in progress_values)

tinysteno/main.py

@@ -106,11 +106,24 @@ def _process_audio(  # pylint: disable=too-many-arguments,too-many-positional-ar
     timestamp: datetime,
 ) -> None:
     """Shared pipeline: transcribe → summarize → export."""
-    print("Transcribing...")
+    from rich.progress import Progress, SpinnerColumn, BarColumn, TextColumn
+
     transcriber = WhisperTranscriber(
         model_size=config.get("whisper_model", "small")
     )
-    result = transcriber.transcribe(wav_path, diarize=config.get("diarization", False))
+    with Progress(
+        SpinnerColumn(),
+        TextColumn("[progress.description]{task.description}"),
+        BarColumn(),
+        TextColumn("{task.percentage:>3.0f}%"),
+        transient=True,
+    ) as progress:
+        task = progress.add_task("Transcribing...", total=100)
+        result = transcriber.transcribe(
+            wav_path,
+            diarize=config.get("diarization", False),
+            on_progress=lambda r: progress.update(task, completed=int(r * 100)),
+        )
     logger.debug(f"Detected language: {result['detected_language']}")
     logger.debug(f"Duration: {result['duration_seconds']:.0f}s")
 

tinysteno/transcriber.py

@@ -1,7 +1,7 @@
 """Whisper transcription module for TinySteno."""
 
 from pathlib import Path
-from typing import Optional
+from typing import Optional, Callable
 import numpy as np
 import soundfile as sf
 from faster_whisper import WhisperModel
@@ -22,7 +22,12 @@ def __init__(self, model_size: str = "small"):
             _MODEL_CACHE[cache_key] = WhisperModel(model_size, device="cpu", compute_type="int8")
         self._model = _MODEL_CACHE[cache_key]
 
-    def transcribe(self, audio_path: str, diarize: bool = False) -> dict:
+    def transcribe(
+        self,
+        audio_path: str,
+        diarize: bool = False,
+        on_progress: Optional[Callable[[float], None]] = None,
+    ) -> dict:
         """Transcribe an audio file and return results.
 
         Returns:
@@ -38,7 +43,7 @@ def transcribe(self, audio_path: str, diarize: bool = False) -> dict:
         is_stereo = data.ndim == 2 and data.shape[1] >= 2
 
         audio_16k = self._convert_to_16khz_array(data, sr)
-        text, language = self._run_whisper(audio_16k)
+        text, language = self._run_whisper(audio_16k, on_progress=on_progress)
         duration = len(audio_16k) / 16000.0
 
         diarised_text = None
@@ -52,11 +57,22 @@ def transcribe(self, audio_path: str, diarize: bool = False) -> dict:
             "detected_language": language,
         }
 
-    def _run_whisper(self, audio: np.ndarray) -> tuple[str, str]:
+    def _run_whisper(
+        self,
+        audio: np.ndarray,
+        on_progress: Optional[Callable[[float], None]] = None,
+    ) -> tuple[str, str]:
         """Run Whisper on a 16kHz mono float32 numpy array."""
+        duration = len(audio) / 16000.0
         segments, info = self._model.transcribe(audio, beam_size=5)
-        text = "".join(segment.text for segment in segments)
-        return text, info.language
+        parts = []
+        for seg in segments:
+            parts.append(seg.text)
+            if on_progress is not None and duration > 0:
+                on_progress(min(seg.start / duration, 1.0))
+        if on_progress is not None:
+            on_progress(1.0)
+        return "".join(parts), info.language
 
     def _run_whisper_segments(self, audio: np.ndarray) -> list[tuple[float, str]]:
         """Run Whisper and return (start_seconds, text) tuples."""