first commit

backend/app/voice/audio_utils.py

@@ -0,0 +1,40 @@
+"""PCM helpers for Bonvoice media (from test_bonvoice/audio_utils.py)."""
+
+import audioop
+import wave
+from pathlib import Path
+
+TARGET_RATE = 8000
+CHUNK_SIZE = 320
+
+
+def wav_to_pcm8000_mono(path: str | Path) -> bytes:
+    path = Path(path)
+    with wave.open(str(path), "rb") as wf:
+        channels = wf.getnchannels()
+        sample_width = wf.getsampwidth()
+        frame_rate = wf.getframerate()
+        pcm = wf.readframes(wf.getnframes())
+
+    if sample_width != 2:
+        raise ValueError(f"Expected 16-bit audio, got {sample_width * 8}-bit")
+
+    if channels == 2:
+        pcm = audioop.tomono(pcm, 2, 0.5, 0.5)
+    elif channels != 1:
+        raise ValueError(f"Unsupported channel count: {channels}")
+
+    if frame_rate != TARGET_RATE:
+        pcm, _ = audioop.ratecv(pcm, 2, 1, frame_rate, TARGET_RATE, None)
+
+    return pcm
+
+
+def chunk_pcm(pcm: bytes, chunk_size: int = CHUNK_SIZE) -> list[bytes]:
+    chunks: list[bytes] = []
+    for i in range(0, len(pcm), chunk_size):
+        chunk = pcm[i : i + chunk_size]
+        if len(chunk) < chunk_size:
+            chunk = chunk + b"\x00" * (chunk_size - len(chunk))
+        chunks.append(chunk)
+    return chunks