From eca3944f6454f0dd44fde0e130ed7df7fa3a1617 Mon Sep 17 00:00:00 2001
From: Hayden Housen <hayden@haydenhousen.com>
Date: Thu, 25 May 2023 15:28:06 -0400
Subject: [PATCH] Add `FileInputDevice` (#151)

* Add FileInputDevice
---
 .../input_device/file_input_device.py         | 52 +++++++++++++++++++
 1 file changed, 52 insertions(+)
 create mode 100644 vocode/streaming/input_device/file_input_device.py

diff --git a/vocode/streaming/input_device/file_input_device.py b/vocode/streaming/input_device/file_input_device.py
new file mode 100644
index 000000000..6b4a7328f
--- /dev/null
+++ b/vocode/streaming/input_device/file_input_device.py
@@ -0,0 +1,52 @@
+from vocode.streaming.models.audio_encoding import AudioEncoding
+import janus
+from vocode.streaming.input_device.base_input_device import BaseInputDevice
+import wave
+import struct
+import numpy as np
+
+
+class FileInputDevice(BaseInputDevice):
+    DEFAULT_SAMPLING_RATE = 44100
+    DEFAULT_CHUNK_SIZE = 2048
+
+    def __init__(
+        self,
+        file_path: str,
+        sampling_rate: int = DEFAULT_SAMPLING_RATE,
+        chunk_size: int = DEFAULT_CHUNK_SIZE,
+        silent_duration: float = 3.0,
+    ):
+        super().__init__(sampling_rate, AudioEncoding.LINEAR16, chunk_size)
+
+        self.queue: janus.Queue[bytes] = janus.Queue()
+
+        with wave.open(file_path, "rb") as wave_file:
+            n_channels = wave_file.getnchannels()
+            frame_rate = wave_file.getframerate()
+            n_frames = wave_file.getnframes()
+            self.total_chunks = n_frames // chunk_size
+            if n_channels != 1:
+                raise ValueError("Only mono audio is supported")
+            if frame_rate != sampling_rate:
+                raise ValueError(
+                    f"Sampling rate of file ({frame_rate}) does not match "
+                    + f"sampling rate of input device ({sampling_rate}). Only "
+                    + f"{self.DEFAULT_SAMPLING_RATE} is supported."
+                )
+
+            for _ in range(self.total_chunks):
+                chunk_data = wave_file.readframes(chunk_size)
+                self.queue.sync_q.put_nowait(chunk_data)
+
+            silent_chunk_data = self.generate_silent_chunk(silent_duration)
+            self.queue.sync_q.put_nowait(silent_chunk_data)
+
+    def generate_silent_chunk(self, duration: float) -> bytes:
+        num_samples = int(self.sampling_rate * duration)
+        samples = np.zeros(num_samples, dtype=np.int16)
+        silent_wave = struct.pack("<" + "h" * len(samples), *samples)
+        return silent_wave
+
+    async def get_audio(self) -> bytes:
+        return await self.queue.async_q.get()