replace base64 with raw data

src/om1_speech/audio/audio_input_stream.py

@@ -5,9 +5,10 @@
 import json
 import logging
 import queue
+import struct
 import threading
 import time
-from typing import Any, Callable, Dict, Generator, List, Optional, Tuple, Union
+from typing import Any, Callable, Generator, List, Optional, Tuple, Union
 
 import pyaudio
 import zenoh
@@ -416,7 +417,7 @@ def fill_buffer_remote(self, data: str):
                 self._rate = rate
                 self._language_code = language_code
 
-    def generator(self) -> Generator[Dict[str, Union[bytes, int]], None, None]:
+    def generator(self) -> Generator[bytes, None, None]:
         """
         Generates a stream of audio data chunk.
 
@@ -437,17 +438,17 @@ def generator(self) -> Generator[Dict[str, Union[bytes, int]], None, None]:
                 if self._is_tts_active:
                     continue
 
-            data = [chunk]
-
-            response = {
-                "audio": base64.b64encode(b"".join(data)).decode("utf-8"),
+            config = {
                 "rate": self._rate,
                 "language_code": self._language_code,
                 "alternative_language_codes": self._alternative_language_codes,
-                "timestamp": int(time.time()),
             }
+
+            config["timestamp"] = int(time.time() * 1000)
+            header = json.dumps(config).encode("utf-8")
+            response = struct.pack(">I", len(header)) + header + chunk
             for audio_callback in self._audio_data_callbacks:
-                audio_callback(json.dumps(response))
+                audio_callback(response)
 
             yield response
 

src/om1_speech/audio/audio_rtsp_input_stream.py

@@ -2,10 +2,11 @@
 import json
 import logging
 import multiprocessing as mp
+import struct
 import threading
 import time
 from queue import Empty, Full
-from typing import Callable, Dict, Generator, List, Optional, Union
+from typing import Callable, Generator, List, Optional, Union
 
 import av
 import numpy as np
@@ -414,7 +415,7 @@ def fill_buffer_remote(self, data: str):
                 self._rate = rate
                 self._language_code = language_code
 
-    def generator(self) -> Generator[Dict[str, Union[bytes, int]], None, None]:
+    def generator(self) -> Generator[bytes, None, None]:
         """
         Generates a stream of audio data chunk.
 
@@ -435,17 +436,17 @@ def generator(self) -> Generator[Dict[str, Union[bytes, int]], None, None]:
                 if self._is_tts_active:
                     continue
 
-            data = [chunk]
-
-            response = {
-                "audio": base64.b64encode(b"".join(data)).decode("utf-8"),
+            config = {
                 "rate": self._rate,
                 "language_code": self._language_code,
                 "alternative_language_codes": self._alternative_language_codes,
-                "timestamp": int(time.time()),
             }
+
+            config["timestamp"] = int(time.time() * 1000)
+            header = json.dumps(config).encode("utf-8")
+            response = struct.pack(">I", len(header)) + header + chunk
             for audio_callback in self._audio_data_callbacks:
-                audio_callback(json.dumps(response))
+                audio_callback(response)
 
             yield response
 

tests/om1_speech/audio/test_audio_input_stream.py

@@ -1,6 +1,7 @@
 import base64
 import json
 import queue
+import struct
 import threading
 from unittest.mock import MagicMock, Mock, patch
 
@@ -146,7 +147,7 @@ def collect_data():
     audio_stream._buff.put(None)
 
     assert len(collected_chunks) > 0
-    assert all(isinstance(data, dict) for data in collected_chunks)
+    assert all(isinstance(data, bytes) for data in collected_chunks)
 
 
 def test_stop(audio_stream, mock_pyaudio):
@@ -182,16 +183,17 @@ def test_callback(data):
         # Process one chunk through generator
         next(stream.generator())
 
-    # Verify callback was called with correct data
-    assert callback_data == json.dumps(
-        {
-            "audio": base64.b64encode(test_data).decode("utf-8"),
-            "rate": 16000,
-            "language_code": "en-US",
-            "alternative_language_codes": [],
-            "timestamp": fixed_time,
-        }
-    )
+    # Verify callback was called with binary header format
+    assert isinstance(callback_data, bytes)
+    header_len = struct.unpack(">I", callback_data[:4])[0]
+    header = json.loads(callback_data[4 : 4 + header_len])
+    audio = callback_data[4 + header_len :]
+
+    assert header["rate"] == 16000
+    assert header["language_code"] == "en-US"
+    assert header["alternative_language_codes"] == []
+    assert header["timestamp"] == int(fixed_time * 1000)
+    assert audio == test_data
 
     stream.stop()
 
@@ -258,10 +260,15 @@ def test_generator_with_alternative_languages(mock_pyaudio):
 
     generated = next(stream.generator())
 
-    assert generated["language_code"] == "en-GB"
-    assert generated["alternative_language_codes"] == alt_langs
-    assert generated["audio"] == base64.b64encode(test_data).decode("utf-8")
-    assert generated["rate"] == 16000
+    assert isinstance(generated, bytes)
+    header_len = struct.unpack(">I", generated[:4])[0]
+    header = json.loads(generated[4 : 4 + header_len])
+    audio = generated[4 + header_len :]
+
+    assert header["language_code"] == "en-GB"
+    assert header["alternative_language_codes"] == alt_langs
+    assert header["rate"] == 16000
+    assert audio == test_data
 
     stream.stop()
 
@@ -318,15 +325,15 @@ def test_callback(data):
 
         next(stream.generator())
 
-    expected_data = json.dumps(
-        {
-            "audio": base64.b64encode(test_data).decode("utf-8"),
-            "rate": 16000,
-            "language_code": "es-MX",
-            "alternative_language_codes": alt_langs,
-            "timestamp": fixed_time,
-        }
-    )
-    assert callback_data == expected_data
+    assert isinstance(callback_data, bytes)
+    header_len = struct.unpack(">I", callback_data[:4])[0]
+    header = json.loads(callback_data[4 : 4 + header_len])
+    audio = callback_data[4 + header_len :]
+
+    assert header["rate"] == 16000
+    assert header["language_code"] == "es-MX"
+    assert header["alternative_language_codes"] == alt_langs
+    assert header["timestamp"] == int(fixed_time * 1000)
+    assert audio == test_data
 
     stream.stop()