Chat-Backend

This project simplifies the development of AI Agents by managing chat sessions. Chat-Backend, along with its peer services, provides chat summaries, file handling, and retrieval of specific numbers of chats from previous sessions to AI-Agents. It automatically manages chat responses, allowing developers to focus on implementing AI Agents rather than managing the infrastructure.

System Flow Diagram

Services

• Chat-AI

• Chat-UI

• Sync-Backend

  we need all of these services to make system work

Getting Started

1. Clone the repository:

   git clone https://github.com/AI-at-Work/Chat-Backend
   cd Chat-Backend

2. Copy the .env.sample to .env and configure the environment variables:

   cp .env.sample .env

   Edit the .env file to set your specific configurations and add openai api key.

3. Start the service:

   make && docker compose up -d --build

4. Create a user in the database and note down the user_id for further use:

   insert into user_data(username, models, balance) VALUES ('Test User', array[4,5,8], 1);

   4 and 5 represent that user have access to GPT4Turbo, GPT4Turbo09 and llama3.1:8b models. balance is in dollar.

5. Start Sync-Backend and Chat-AI service by following steps mention in respective repositories.

6. Restart the service:

   docker-compose down && docker-compose up -d

7. Start Chat-UI service for the existing UI by following steps mention in respective repository.

Configuration

Key configuration options in the .env file:

• SERVER_HOST and SERVER_PORT: Host and port for the Chat-Backend server
• DB_*: PostgreSQL database configurations
• REDIS_*: Redis configurations
• AI_SERVER_HOST and AI_SERVER_PORT: AI service gRPC server details
• MAX_FILE_SIZE: Maximum allowed file upload size in MB
• MAX_CHAT_HISTORY_CONTEXT: Number of previous chat messages to include in context

Refer to the .env.sample file for a complete list of configuration options.

API Documentation

Model Id Mapping

GPTTurbo125      = 0
GPTTurbo         = 1
GPTTurbo1106     = 2
GPTTurboInstruct = 3
GPT4Turbo        = 4
GPT4Turbo09      = 5
GPT4             = 6
GPT40613         = 7
llama3.1:8b      = 8

gRPC Service

The Chat-Backend uses gRPC to communicate with the AI Agent. The protocol is defined in proto/ai_service.proto.

Key message types:

• Request: Contains user chat information, including user ID, session ID, chat message, model name, etc.
• Response: Contains the AI's response text and timestamp.

WebSocket API

This documentation provides an overview of the WebSocket request handlers defined in the provided code. Each function generates a request to be sent via WebSocket for various operations related to user details, sessions, and chat messages. Below is the detailed explanation of each function and the corresponding message types.

Table of Contents

• Message Types
• Functions
  • getUserDetails
  • getUserSessions
  • getUserChatsBySessionId
  • getUserChatsResponse
  • deleteUserSession
  • modelList

Message Types

The following constants represent the different message types used in the WebSocket requests:

• MessageCodeUserDetails: 0
• MessageCodeListSessions: 1
• MessageCodeChatsBySessionId: 2
• MessageCodeChatMessage: 3
• MessageCodeSessionDelete: 4
• MessageCodeGetAIModels: 5

Functions

getUserDetails

Generates a request to fetch user details.

Parameters

• user_id (String): The ID of the user.

{
    type: MessageCodeUserDetails,
    data: {
        user_id: (String),
    },
}

Returns

{
  "user_id": "String",
  "username": "String"
}

getUserSessions

Generates a request to fetch the list of user sessions.

Parameters

• user_id (String): The ID of the user.

{
    type: MessageCodeListSessions,
    data: {
        user_id: (String),
    },
}

Returns

{
  "user_id": "String", 
  "session_info": [{
    "session_id": "String",
    "session_name": "String"
  }]
}

getUserChatsBySessionId

Generates a request to fetch chats by session ID.

Parameters

• user_id (String): The ID of the user.
• session_id (String): The ID of the session.

{
    type: MessageCodeChatsBySessionId,
    data: {
        user_id: (String),
        session_id: (String),
    },
}

Returns

{
  "user_id": "String",
  "session_id": "String",
  "chat": "String"
}

getUserChatsResponse

Generates a request to send a chat message in a specific session.

Parameters

• user_id (String): The ID of the user.
• session_id (String): The ID of the session.
• model_id (Int): The ID of the AI model.
• message (String): The chat message.
• session_prompt (String): The session prompt.
• file_name (String): The name of the file (optional).

{
    type: MessageCodeChatMessage,
    data: {
        user_id: (String),
        session_id: (String),
        model_id: (Int),
        message: (String),
        session_prompt: (String),
        file_name: (String),
    },
}

Returns

{
  "user_id": "String",
  "session_id": "String",
  "session_name": "String",
  "message": "String"
}

deleteUserSession

Generates a request to delete a user session.

Parameters

• user_id (String): The ID of the user.
• session_id (String): The ID of the session.

{
    type: MessageCodeSessionDelete,
    data: {
        user_id: userId,
        session_id: sessionId,
    },
}

Returns

{
    "user_id": "String"
}

modelList

Generates a request to fetch the list of AI models available.

Parameters

• user_id (String): The ID of the user.

{
    type: MessageCodeGetAIModels,
    data: {
        user_id: userId,
    },
}

Returns

{
    "models": ["String"]
}

Contributing

We welcome contributions to the Chat-Backend project! Here's how you can contribute:

1. Fork the repository
2. Create a new branch (git checkout -b feature/AmazingFeature)
3. Make your changes
4. Commit your changes (git commit -m 'Add some AmazingFeature')
5. Push to the branch (git push origin feature/AmazingFeature)
6. Open a Pull Request

License

This project is licensed under the GNU GENERAL PUBLIC LICENSE V3.

Contact

For any questions or suggestions, please open an issue in the GitHub repository.