Skip to content

Latest commit

 

History

History
225 lines (176 loc) · 5.4 KB

local-machine-deployment.md

File metadata and controls

225 lines (176 loc) · 5.4 KB
Raw

Local Machine Deployment Guide

This guide provides instructions to set up and run the following locally:

  1. ConnectedAI - Web App - Backend Server
  2. ConnectedAI - Web App - Frontend Server
  3. ConnectedAI - AI - Backend Server

1. ConnectedAI - Web App - Backend Server

Overview

The backend server handles:

  • User authentication (with JWT tokens).
  • Guest token generation.
  • Chatbot message processing.

Features

  1. User login with JWT authentication.
  2. Guest mode support with token issuance.
  3. API to process chat messages.

Requirements

  • Node.js (v16 or higher)
  • NPM or Yarn
  • .env file with the following variables: 
    PORT=5001
SECRET_KEY=your_secret_key

How to setup the NoSql Google Cloud Firestore collection

  • Make sure you have the Service Account private key downloaded as JSON file serviceAccount.json

    To use the Firebase Admin SDK on your own server (or any other Node.js environment), use a service account. Go to IAM & admin > Service accounts in the Google Cloud console. Generate a new private key and save the JSON file.

  • The csv file with the Users data to be loaded into the Collection

  1. Go to the db-load folder

    cd 8_code/apps/api-server/db-load

  2. install dependencies

    npm install

  3. Run the upload

    node index.js

  4. The records will be uploaded to the Firestore users collection.

Installation

  1. Clone the repository:

    git clone <repository_url>
cd 8_code/apps/api-server

  2. Install dependencies:

    npm install

  3. Start the server:

    npm start

By default, the server uses PORT=5001 and the specified SECRET_KEY.

2. ConnectedAI - Web App - Frontend

Overview

The React-based frontend allows:

  • User login with a username and password.
  • Guest mode for interacting with the chatbot without an account.
  • A modern and responsive chat interface built with @chatscope/chat-ui-kit-react.
  • JWT token-based authentication for secure API requests.

Features

  1. Login with username and password.
  2. Guest mode for non-registered users.
  3. Responsive chat UI.
  4. Token-based authentication for requests to the backend.

Prerequisites

  • Node.js (v16 or higher)
  • NPM or Yarn

Installation

  1. Clone the repository:

    git clone <repository_url>
cd 8_code/apps/app-ui

  2. Install dependencies:

    npm install

  3. Configure the backend API base URL:

    • Open src/api/apiClient.js.
    • Update the BASE_URL constant to match your backend server's URL: 
      const BASE_URL = 'http://localhost:5001';

  4. Start the development server:

    npm start

  5. To build the code for deployment:

    npm run build

Notes

  • Ensure both backend and frontend servers are running to enable full functionality.
  • Configure the backend and frontend to communicate using consistent environment variables or configuration files.

3. ConnectedAI - AI - Backend Server

Initialization

To install the necessary dependencies, run:

make init

Running the Service

To start the service, use:

make run

Calling the API

To call the API, execute:

make post

Switch service

Comment the lines in start_service.py to switch between the agents.

from core.server.main_gemini import app
# from core.server.main_mock import app

main_gemini service is using Gemini model, either the free Google Gemini model or VertexAI.

main_mock service is using the mock model, which is a simple model that returns a fixed response.

You will need to create local/mock_data.py contains the mock output data. i.e.

outputs=[
    {"user":"Hello","AI":"How can I help you!"},
    {"user":"Hello","AI":"How can I help you 2!"},
    {"user":"Hello","AI":"How can I help you 3!"},
]

You will also need a local/local_config.py, use the one in example folder as a template.

AI Backend Endpoints

API Endpoints

  • /chat
  • /chat_mock
  • /chat_free
  • /chat_graph

Streaming (WebSocket) Endpoints

  • /ws/chat
  • /ws/chat_mock
  • /ws/chat_free
  • /ws/chat_graph

Example API Request

{
  "chatMessage": "Hello",
  "sessionId": "test_session",
  "isAuthenticated": true,
  "customerId": 123
}

or

{"chatMessage":"Hello","sessionId":"test_session","isAuthenticated":false}

example api response:

{"chatMessage":"Hello","sessionId":"test_session","isAuthenticated":true,"customerId":123,"response":"\n","isCompleted":true}

example websocket request:

{"chatMessage":"Hello","sessionId":"test_session","isAuthenticated":true,"customerId":123,"response":"\n"}

example websocket response as sequence:

{"response":"","isCompleted":false}
{"response":"","isCompleted":false}
{"response":"Hi","isCompleted":false}
{"response":" there! How can I help you today?  I can assist with questions about","isCompleted":false}
{"response":" your orders or products.\n","isCompleted":false}
{"chatMessage":"Hello","sessionId":"test_session","isAuthenticated":true,"customerId":123,"response":"\n","isCompleted":true}

Note that JWT tokens are now used to encode and decode all messages. Please refer to convert_jwt_to_chat_item function in ai-backend/core/server/main_server.py for detail.