Enhance Discord voice activity detection configurability

Makes the `silence_threshold` and `min_speech_duration` parameters configurable during `VoiceReceiver` initialization, allowing for dynamic tuning of voice activity detection. The default `min_speech_duration` is also increased from 0.25s to 0.75s to help filter out shorter, potentially spurious speech segments.

Additionally, the log level for NaCl decryption failures is reduced from `WARNING` to `DEBUG` to minimize log noise.

This commit also updates the project version to 0.6.5.

Author: DataBassGit

setup.py

@@ -1,6 +1,6 @@
 from setuptools import setup, find_packages
 
-__version__ = "0.6.4"
+__version__ = "0.6.5"
 LICENSE = "GNU General Public License v3 or later (GPLv3+)"
 
 
@@ -11,7 +11,7 @@ def get_long_description():
 
 setup(
     name="agentforge",
-    version="0.6.4",
+    version="0.6.5",
     description="AI-driven task automation system",
     author="John Smith, Ansel Anselmi",
     author_email="contact@agentforge.net",

src/agentforge/utils/discord/discord_voice.py

@@ -39,17 +39,19 @@ class VoiceReceiver:
     Handles NaCl transport decryption, RTP padding removal, DAVE E2EE decryption,
     and Opus decoding natively. (Adapted from Hermes-Agent example.py)
     """
-    SILENCE_THRESHOLD = 1.5
-    MIN_SPEECH_DURATION = 0.25
     SAMPLE_RATE = 48000
     CHANNELS = 2
 
-    def __init__(self, voice_client, logger: Logger):
+    def __init__(self, voice_client, logger: Logger, silence_threshold=1.5, min_speech_duration=0.75):
         self._vc = voice_client
         self.logger = logger
         self._running = False
         self._paused = False
 
+        # Dynamic attributes for audio filtering
+        self.silence_threshold = silence_threshold
+        self.min_speech_duration = min_speech_duration
+
         self._secret_key = None
         self._bot_ssrc = 0
         self._dave_session = None
@@ -163,7 +165,7 @@ def _on_packet(self, data: bytes):
             box = nacl.secret.Aead(self._secret_key)
             decrypted = box.decrypt(encrypted, header, bytes(nonce))
         except Exception as e:
-            self.logger.warning(f"NaCl decrypt failed: {e} (hdr={header_size}, enc={len(encrypted)})")
+            self.logger.debug(f"NaCl decrypt failed: {e} (hdr={header_size}, enc={len(encrypted)})")
             return
 
         # Skip encrypted extension data
@@ -226,7 +228,7 @@ def check_silence(self) -> list:
                 buf = self._buffers[ssrc]
                 buf_duration = len(buf) / (self.SAMPLE_RATE * self.CHANNELS * 2)
 
-                if silence_duration >= self.SILENCE_THRESHOLD and buf_duration >= self.MIN_SPEECH_DURATION:
+                if silence_duration >= self.silence_threshold and buf_duration >= self.min_speech_duration:
                     user_id = self._ssrc_to_user.get(ssrc, 0)
                     if not user_id:
                         user_id = self._infer_user_for_ssrc(ssrc)
@@ -237,7 +239,7 @@ def check_silence(self) -> list:
                     self._buffers[ssrc] = bytearray()
                     self._last_packet_time.pop(ssrc, None)
 
-                elif silence_duration >= self.SILENCE_THRESHOLD * 2:
+                elif silence_duration >= self.silence_threshold * 2:
                     self._buffers.pop(ssrc, None)
                     self._last_packet_time.pop(ssrc, None)