Skip to content

README.md

Latest commit

 

History

History
186 lines (147 loc) · 4.39 KB

README.md

File metadata and controls

186 lines (147 loc) · 4.39 KB
Raw

Warning

This service project is only for developer quick experience and demonstration purposes. Do not use in production environment. Production environment services need to be developed by developers.

IoT Conversational AI Server 🚀

Python 3.7+ License

This server provides APIs for managing IoT devices and conversational AI agents, built with Python and supporting real-time communication via Agora RTC.

Features

  • 🎙️ Real-time voice communication
  • 🤖 Conversational AI integration
  • 🔐 Secure token generation
  • 📦 Easy deployment options
  • 📊 Comprehensive logging

Requirements

  • Python 3.7+
  • Required packages: 
    requests
flask
pyjwt

Installation

  1. Clone this repository
  2. Install dependencies:
pip install -r requirements.txt

Configuration

Configuration File

Create a config.json file with the following structure and explanations:

{
  // Agora related configuration
  "app_id": "YOUR_AGORA_APP_ID",  // Agora App ID
  "app_certificate": "YOUR_AGORA_APP_CERTIFICATE",  // Agora App Certificate
  
  // Customer authentication information
  "customer_key": "YOUR_CUSTOMER_KEY",  // Customer Key
  "customer_secret": "YOUR_CUSTOMER_SECRET",  // Customer Secret
  
  // Automatic Speech Recognition (ASR) configuration
  "asr": {
    "language": "zh-CN"  // Recognition language, default Chinese (supports Chinese-English mixed)
  },
  
  // System parameters configuration
  "parameters": {
    "output_audio_codec": "G722"  // RTC streaming audio codec format, supported formats: "PCMU" "PCMA" "G722" "OPUS" "OPUSFB"
    "transcript": {               // Subtitle feature parameter configuration
      "enable": false             // Disable subtitle feature
    }
  },
  
  // Text-to-Speech (TTS) configuration
  "tts": {
    "vendor": "YOUR_TTS_VENDOR",  // TTS service provider
    "params": {
      
    }
  },
  
  // Session timeout configuration
  "idle_timeout": 30,  // Session timeout (seconds)
  
  // Large Language Model (LLM) configuration
  "llm": {
    "url": "YOUR_LLM_API_URL",  // LLM service URL
    "params": {
      "model": "YOUR_LLM_MODEL"  // Model used
    },
    "api_key": "YOUR_LLM_API_KEY",  // LLM service API key
    "system_messages": [  // System preset messages
      {
        "role": "system",
        "content": "You are a helpful chatbot."
      }
    ],
    "max_history": 10,  // Maximum history records
    "greeting_message": "Hello, I'm your AI assistant. How can I help you?",  // Greeting message
    "failure_message": "Sorry, I am unable to answer your question..."  // Failure message
  }
}

For more detailed parameter configuration, see: https://doc.shengwang.cn/doc/convoai/restful/convoai/operations/start-agent

Running the Server

Start the server with:

python3 main.py

The server will run on port 5001 by default.

API Documentation

Base URL

https://your-domain.com/api/v1

Authentication

All requests require an Authorization header:

Authorization: Bearer <access_token>

API Endpoints

POST /device

Register a new device and generate RTC token

Request body:

{
  "channel_name": "DEVICE_ID",
  "uid": DEVICE_USER_ID
}

POST /agent/start

Start a conversational AI agent

Request body:

{
  "channel_name": "DEVICE_ID",
  "uid": DEVICE_USER_ID,
  "agent_uid": AGENT_USER_ID
}

POST /agent/stop

Stop a conversational AI agent

Request body:

{
  "agent_id": "AGENT_ID"
}

Logging

Logs are written to stdout with the following format:

[timestamp] [level] - [message]

Log levels:

  • DEBUG: Detailed debug information
  • INFO: General operational messages
  • WARNING: Indicates potential issues
  • ERROR: Errors that need attention
  • CRITICAL: Critical system failures

Security Considerations

  • Always keep your Agora credentials secure
  • Use HTTPS in production environments
  • Regularly rotate your access tokens
  • Implement rate limiting for API endpoints

License

This project is licensed under the MIT License - see the LICENSE file for details.

Example Usage

import requests

# Generate RTC token
response = requests.post(
    "https://your-domain.com/device",
    json={"channel_name": "12345", "uid": 1}
)
print(response.json())