videosdk-plugins-openai 0.0.20__tar.gz → 0.0.21__tar.gz

{videosdk_plugins_openai-0.0.20 → videosdk_plugins_openai-0.0.21}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: videosdk-plugins-openai
-Version: 0.0.20
+Version: 0.0.21
 Summary: VideoSDK Agent Framework plugin for OpenAI services
 Author: videosdk
 License-Expression: Apache-2.0
@@ -13,7 +13,7 @@ Classifier: Topic :: Multimedia :: Video
 Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
 Requires-Python: >=3.11
 Requires-Dist: openai[realtime]>=1.68.2
-Requires-Dist: videosdk-agents>=0.0.20
+Requires-Dist: videosdk-agents>=0.0.21
 Description-Content-Type: text/markdown
 
 # VideoSDK OpenAI Plugin

{videosdk_plugins_openai-0.0.20 → videosdk_plugins_openai-0.0.21}/pyproject.toml

@@ -21,7 +21,7 @@ classifiers = [
     "Topic :: Scientific/Engineering :: Artificial Intelligence",
 ]
 dependencies = [
-    "videosdk-agents>=0.0.20",
+    "videosdk-agents>=0.0.21",
     "openai[realtime]>=1.68.2",
 ]
 

{videosdk_plugins_openai-0.0.20 → videosdk_plugins_openai-0.0.21}/videosdk/plugins/openai/tts.py

@@ -4,6 +4,7 @@ from typing import Any, AsyncIterator, Literal, Optional, Union
 import httpx
 import os
 import openai
+import asyncio
 
 from videosdk.agents import TTS
 
@@ -81,6 +82,7 @@ class OpenAITTS(TTS):
                 self.emit("error", "Audio track or event loop not set")
                 return
 
+            audio_data = b""
             async with self._client.audio.speech.with_streaming_response.create(
                 model=self.model,
                 voice=voice_id or self.voice,
@@ -91,13 +93,32 @@ class OpenAITTS(TTS):
             ) as response:
                 async for chunk in response.iter_bytes():
                     if chunk:
-                        self.loop.create_task(self.audio_track.add_new_bytes(chunk))
+                        audio_data += chunk
+
+
+            if audio_data:
+                await self._stream_audio_chunks(audio_data)
 
         except openai.APIError as e:
             self.emit("error", str(e))
         except Exception as e:
             self.emit("error", f"TTS synthesis failed: {str(e)}")
 
+    async def _stream_audio_chunks(self, audio_bytes: bytes) -> None:
+        """Stream audio data in chunks for smooth playback"""
+        chunk_size = int(OPENAI_TTS_SAMPLE_RATE * OPENAI_TTS_CHANNELS * 2 * 20 / 1000) 
+        
+        for i in range(0, len(audio_bytes), chunk_size):
+            chunk = audio_bytes[i:i + chunk_size]
+            
+            if len(chunk) < chunk_size and len(chunk) > 0:
+                padding_needed = chunk_size - len(chunk)
+                chunk += b'\x00' * padding_needed
+            
+            if len(chunk) == chunk_size:
+                self.loop.create_task(self.audio_track.add_new_bytes(chunk))
+                await asyncio.sleep(0.001)
+
     async def aclose(self) -> None:
         """Cleanup resources"""
         await self._client.close()

videosdk_plugins_openai-0.0.21/videosdk/plugins/openai/version.py

@@ -0,0 +1 @@
+__version__ = "0.0.21"

videosdk_plugins_openai-0.0.20/videosdk/plugins/openai/version.py

@@ -1 +0,0 @@
-__version__ = "0.0.20"