feat(dictate): use local socket-activated Parakeet

services/dictate/src/pi_dictate_service/audio.py

@@ -0,0 +1,22 @@
+"""In-memory PCM validation and conversion."""
+
+from __future__ import annotations
+
+from array import array
+import struct
+
+MAX_PCM_BYTES = 5 * 60 * 16_000 * 2
+
+
+class AudioError(ValueError):
+    pass
+
+
+def pcm_s16le_to_floats(pcm: bytes, *, max_bytes: int = MAX_PCM_BYTES) -> array[float]:
+    """Convert mono little-endian signed-16 PCM to a bounded float32 array."""
+
+    if len(pcm) > max_bytes:
+        raise AudioError("PCM exceeds the five-minute duration limit")
+    if len(pcm) % 2:
+        raise AudioError("PCM byte length must be even")
+    return array("f", (sample[0] / 32768.0 for sample in struct.iter_unpack("<h", pcm)))