feat: establish voice assistant V1 baseline

src/voice_assistant/audio/models.py

@@ -0,0 +1,63 @@
+"""
+Canonical audio models for the Voice Assistant.
+
+Purpose:
+    Define the provider-independent audio format used internally by
+    the Voice Core.
+
+Responsibilities:
+    - Describe an audio format in a typed structure.
+    - Define the canonical audio format used by the Voice Core.
+    - Provide a shared audio contract for transports and providers.
+
+Non-responsibilities:
+    - Encoding or decoding audio.
+    - Resampling audio.
+    - Converting provider-specific audio formats.
+    - Managing WebSocket audio frames.
+    - Performing STT or TTS operations.
+
+Architecture:
+    Client / Transport
+            |
+            v
+      Canonical Audio
+      PCM S16LE / 16 kHz / mono
+            |
+            v
+        Voice Core
+            |
+      +-----+-----+
+      |           |
+      v           v
+     STT          TTS
+   Adapter      Adapter
+
+    Provider-specific format conversion, when required, belongs
+    inside the corresponding provider adapter.
+"""
+
+from typing import Literal
+
+from pydantic import BaseModel
+
+
+class AudioFormat(BaseModel):
+    """
+    Describe one PCM audio format.
+
+    This model is independent from WebSocket, STT, TTS, or any
+    concrete provider.
+    """
+
+    encoding: Literal["pcm_s16le"]
+    sample_rate: int
+    channels: int
+
+
+CANONICAL_AUDIO_FORMAT = AudioFormat(
+    encoding="pcm_s16le",
+    sample_rate=16000,
+    channels=1,
+)
+