111 lines
4.7 KiB
Python
111 lines
4.7 KiB
Python
"""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")
|
|
|
|
# OpenAI / LLM Configuration
|
|
openai_api_key: Optional[str] = Field(default=None, description="OpenAI API key")
|
|
openai_api_url: Optional[str] = Field(default=None, description="OpenAI API base URL (for Azure/compatible)")
|
|
llm_model: str = Field(default="gpt-4o-mini", description="LLM model name")
|
|
llm_temperature: float = Field(default=0.7, description="LLM temperature for response generation")
|
|
|
|
# TTS Configuration
|
|
tts_provider: str = Field(default="siliconflow", description="TTS provider (edge, siliconflow)")
|
|
tts_voice: str = Field(default="anna", description="TTS voice name")
|
|
tts_speed: float = Field(default=1.0, description="TTS speech speed multiplier")
|
|
|
|
# SiliconFlow Configuration
|
|
siliconflow_api_key: Optional[str] = Field(default=None, description="SiliconFlow API key")
|
|
siliconflow_tts_model: str = Field(default="FunAudioLLM/CosyVoice2-0.5B", description="SiliconFlow TTS model")
|
|
|
|
# ASR Configuration
|
|
asr_provider: str = Field(default="siliconflow", description="ASR provider (siliconflow, buffered)")
|
|
siliconflow_asr_model: str = Field(default="FunAudioLLM/SenseVoiceSmall", description="SiliconFlow ASR model")
|
|
asr_interim_interval_ms: int = Field(default=500, description="Interval for interim ASR results in ms")
|
|
asr_min_audio_ms: int = Field(default=300, description="Minimum audio duration before first ASR result")
|
|
|
|
# Duplex Pipeline Configuration
|
|
duplex_enabled: bool = Field(default=True, description="Enable duplex voice pipeline")
|
|
duplex_greeting: Optional[str] = Field(default=None, description="Optional greeting message")
|
|
duplex_system_prompt: Optional[str] = Field(
|
|
default="You are a helpful, friendly voice assistant. Keep your responses concise and conversational.",
|
|
description="System prompt for LLM"
|
|
)
|
|
|
|
# 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
|