Integrate eou and vad

This commit is contained in:
Xin Wang
2026-01-29 13:57:12 +08:00
parent 4cb267a288
commit cd90b4fb37
25 changed files with 2592 additions and 297 deletions

81
app/config.py Normal file
View File

@@ -0,0 +1,81 @@
"""Configuration management using Pydantic settings."""
from typing import List, Optional
from pydantic import Field
from pydantic_settings import BaseSettings, SettingsConfigDict
import json
class Settings(BaseSettings):
"""Application settings loaded from environment variables."""
model_config = SettingsConfigDict(
env_file=".env",
env_file_encoding="utf-8",
case_sensitive=False,
extra="ignore"
)
# Server Configuration
host: str = Field(default="0.0.0.0", description="Server host address")
port: int = Field(default=8000, description="Server port")
external_ip: Optional[str] = Field(default=None, description="External IP for NAT traversal")
# Audio Configuration
sample_rate: int = Field(default=16000, description="Audio sample rate in Hz")
chunk_size_ms: int = Field(default=20, description="Audio chunk duration in milliseconds")
default_codec: str = Field(default="pcm", description="Default audio codec")
# VAD Configuration
vad_type: str = Field(default="silero", description="VAD algorithm type")
vad_model_path: str = Field(default="data/vad/silero_vad.onnx", description="Path to VAD model")
vad_threshold: float = Field(default=0.5, description="VAD detection threshold")
vad_min_speech_duration_ms: int = Field(default=250, description="Minimum speech duration in milliseconds")
vad_eou_threshold_ms: int = Field(default=400, description="End of utterance (silence) threshold in milliseconds")
# Logging
log_level: str = Field(default="INFO", description="Logging level")
log_format: str = Field(default="json", description="Log format (json or text)")
# CORS
cors_origins: str = Field(
default='["http://localhost:3000", "http://localhost:8080"]',
description="CORS allowed origins"
)
# ICE Servers (WebRTC)
ice_servers: str = Field(
default='[{"urls": "stun:stun.l.google.com:19302"}]',
description="ICE servers configuration"
)
@property
def chunk_size_bytes(self) -> int:
"""Calculate chunk size in bytes based on sample rate and duration."""
# 16-bit (2 bytes) per sample, mono channel
return int(self.sample_rate * 2 * (self.chunk_size_ms / 1000.0))
@property
def cors_origins_list(self) -> List[str]:
"""Parse CORS origins from JSON string."""
try:
return json.loads(self.cors_origins)
except json.JSONDecodeError:
return ["http://localhost:3000", "http://localhost:8080"]
@property
def ice_servers_list(self) -> List[dict]:
"""Parse ICE servers from JSON string."""
try:
return json.loads(self.ice_servers)
except json.JSONDecodeError:
return [{"urls": "stun:stun.l.google.com:19302"}]
# Global settings instance
settings = Settings()
def get_settings() -> Settings:
"""Get application settings instance."""
return settings