From bb8da14e8a25d638e85e828837824d7b7cdb85b4 Mon Sep 17 00:00:00 2001
From: Shiven Mian <shivenmian97@gmail.com>
Date: Sun, 4 Feb 2024 00:40:58 -0800
Subject: [PATCH] feat: tts

---
 OS/01/assistant/tts.py | 40 ++++++++++++++++++++++++++++++++--------
 1 file changed, 32 insertions(+), 8 deletions(-)

diff --git a/OS/01/assistant/tts.py b/OS/01/assistant/tts.py
index 8d58619..e5254da 100644
--- a/OS/01/assistant/tts.py
+++ b/OS/01/assistant/tts.py
@@ -3,16 +3,40 @@ Defines a function which takes text and returns a path to an audio file.
 """
 
 from openai import OpenAI
+import pydub
+import pydub.playback
+import tempfile
+import os
+from datetime import datetime
+from io import BytesIO
 
 client = OpenAI()
+chunk_size = 1024
+read_chunk_size = 4096
 
-def tts(text, file_path):
+def tts(text):
 
-    response = client.with_streaming_response.audio.speech.create(
-        model="tts-1",
-        voice="alloy",
-        input=text,
-    )
+    temp_dir = tempfile.gettempdir()
+    output_path = os.path.join(temp_dir, f"output_{datetime.now().strftime('%Y%m%d%H%M%S%f')}.mp3")
 
-    response.stream_to_file(file_path)
-    
+    try:
+        with (
+            client.with_streaming_response.audio.speech.create(
+            model="tts-1",
+            voice="alloy",
+            input=text,
+            response_format='mp3',
+            speed=1.2)
+        ) as response:
+            with open(output_path, 'wb') as f:
+                for chunk in response.iter_bytes(chunk_size):                    
+                    f.write(chunk)
+        
+        with open(output_path, 'rb') as f:
+            byte_chunk = f.read(read_chunk_size)
+            yield byte_chunk
+
+        seg = pydub.AudioSegment.from_mp3(output_path)
+        pydub.playback.play(seg)
+    finally:
+        os.remove(output_path)
\ No newline at end of file