Fix BUGs of streaming non-wav format audio; improve robustness of releasing audio container

Refactor StreamingAudioWriter to improve audio encoding reliability

- Restructure audio encoding logic for better error handling
- Create a new method `_create_container()` to manage container creation
- Improve handling of different audio formats and encoding scenarios
- Add error logging for audio chunk encoding failures
- Simplify container and stream management in write_chunk method

api/src/services/streaming_audio_writer.py

@@ -2,7 +2,7 @@
 
 import struct
 from io import BytesIO
-from typing import Optional
+from typing import Optional, Dict
 
 import numpy as np
 import soundfile as sf
@@ -18,18 +18,35 @@ class StreamingAudioWriter:
         self.sample_rate = sample_rate
         self.channels = channels
         self.bytes_written = 0
-        self.pts=0
-
-        codec_map = {"wav":"pcm_s16le","mp3":"mp3","opus":"libopus","flac":"flac", "aac":"aac"}
+        self.pts = 0
+        
         # Format-specific setup
-        if self.format in ["wav", "opus","flac","mp3","aac","pcm"]:
-            if self.format != "pcm":
-                self.output_buffer = BytesIO()
-                self.container = av.open(self.output_buffer, mode="w", format=self.format)
-                self.stream = self.container.add_stream(codec_map[self.format],sample_rate=self.sample_rate,layout='mono' if self.channels == 1 else 'stereo')
-                self.stream.bit_rate = 96000
-        else:
+        if self.format not in ["wav", "opus", "flac", "mp3", "aac", "pcm"]:
             raise ValueError(f"Unsupported format: {format}")
+            
+        # Codec mapping
+        self.codec_map = {
+            "wav": "pcm_s16le",
+            "mp3": "mp3",
+            "opus": "libopus",
+            "flac": "flac", 
+            "aac": "aac"
+        }
+
+    def _create_container(self):
+        """Create a new container for each write operation"""
+        if self.format == "pcm":
+            return None, None
+            
+        buffer = BytesIO()
+        container = av.open(buffer, mode="w", format=self.format)
+        stream = container.add_stream(
+            self.codec_map[self.format],
+            sample_rate=self.sample_rate,
+            layout='mono' if self.channels == 1 else 'stereo'
+        )
+        stream.bit_rate = 96000
+        return container, buffer
 
     def write_chunk(
         self, audio_data: Optional[np.ndarray] = None, finalize: bool = False
@@ -40,38 +57,50 @@ class StreamingAudioWriter:
             audio_data: Audio data to write, or None if finalizing
             finalize: Whether this is the final write to close the stream
         """
-
-        if finalize:
-            if self.format != "pcm":
-                # Flush encoder buffers
-                for packet in self.stream.encode(None):
-                    self.container.mux(packet)
-                self.container.close()
-                data = self.output_buffer.getvalue()
-                self.output_buffer.seek(0)
-                self.output_buffer.truncate(0)
-                return data
-            return b""
-
-        if audio_data is None or len(audio_data) == 0:
-            return b""
-
+        # Handle PCM format separately as it doesn't use PyAV
         if self.format == "pcm":
+            if finalize or audio_data is None or len(audio_data) == 0:
+                return b""
             return audio_data.tobytes()
-        else:
-            frame = av.AudioFrame.from_ndarray(audio_data.reshape(1, -1), format='s16', layout='mono' if self.channels == 1 else 'stereo')
+            
+        # Handle empty input
+        if not finalize and (audio_data is None or len(audio_data) == 0):
+            return b""
+            
+        try:
+            # Create a new container for this operation
+            container, buffer = self._create_container()
+            stream = container.streams[0]
+            
+            if finalize:
+                # Just return empty bytes for finalize in the new design
+                return b""
+                
+            # Create audio frame
+            frame = av.AudioFrame.from_ndarray(
+                audio_data.reshape(1, -1), 
+                format='s16', 
+                layout='mono' if self.channels == 1 else 'stereo'
+            )
             frame.sample_rate = self.sample_rate
             frame.pts = self.pts
             self.pts += frame.samples
             
-            encoded_data = b""
-            for packet in self.stream.encode(frame):
-                self.container.mux(packet)
-                # Get the encoded data from the buffer
-                encoded_data = self.output_buffer.getvalue()
-                # Clear the buffer for next write
-                self.output_buffer.seek(0)
-                self.output_buffer.truncate(0)
+            # Encode the frame
+            for packet in stream.encode(frame):
+                container.mux(packet)
+                
+            # Flush any remaining packets
+            for packet in stream.encode(None):
+                container.mux(packet)
+                
+            # Close the container and get the data
+            container.close()
+            data = buffer.getvalue()
             
-            return encoded_data
+            return data
+            
+        except Exception as e:
+            logger.error(f"Error encoding audio chunk: {e}")
+            return b""