Infrastructure-Chatbot

Table of Contents

• Overview
• Local Development Setup
• Production Setup

Overview

The Infrastructure-Chatbot repository contains the necessary files to run the infrastructure services for the ACCESS AI project. The infrastructure services include the following components:

• Keycloak: An open-source identity and access management solution.
• PostgreSQL: An open-source relational database.
• Vector Store: A service for storing and querying vectors.
• Context Service: A service for storing and querying context information.

Local Development Setup

Prerequisites

• Download the necessary repositories:

  git clone https://github.com/ACCESS-AI/Frontend-Re.git
  git clone https://github.com/ACCESS-AI/Infrastructure-RE.git
  git clone https://github.com/ACCESS-AI/Context-Service.git
  git clone https://github.com/ACCESS-AI/Backend-Re.git
  git clone https://github.com/ACCESS-AI/Chatbot.git

1. Run the Infrastructure

1. Insert your Mistral AI API key in the MISTRAL_API_KEY variable of the .env file in the Infrastructure-Chatbot directory.
2. Choose a working directory for saving the course repositories and insert it in the WORKING_DIR variable of the .env file in the Infrastructure-Chatbot directory.
3. If you wish to change the default values of the environment variables, you can do so by editing the .env file in the Infrastructure-Chatbot directory.
4. Run the following command to start the infrastructure services:

cd Infrastructure-Chatbot
docker compose up -d

2. Run the Context Service (optional: if you want to develop the context service)

1. Stop the context service container which is automatically started by the Infrastructure Chatbot docker compose up -d command.
2. Follow the steps in the README.md file of the Context Service repository.

3. Build and Publish the Chatbot Library

1. If you want to change the default publishsing method, refer to the README.md file of the Chatbot repository.
2. Run the following command to build and publish the Chatbot library:

   cd Chatbot
   ./gradlew publish

4. Run the Backend

1. Load the Backend-Re-Chatbot project in IntelliJ.
2. In the run configuration for AccessApplicationKt, add the following environment variables:
   • WORKING_DIR should be an existing empty directory that ACCESS will use to clone course repos.
   • API_KEY is a mock key used for the ACCESS API.
   • MISTRAL_API_KEY is the API key for the Mistral API.
   • MISTRAL_EMBEDDING_MODEL is the URL of the Mistral embedding model.
   • MISTRAL_EMBEDDING_HOST is the host of the Mistral embedding model.
   • MISTRAL_LLM_MODEL is the URL of the Mistral LLM model.
   • MISTRAL_LLM_HOST is the host of the Mistral LLM model.
   • VECTOR_STORE_HOST is the host of the vector store.
   • SIMILARITY_SCORE_THRESHOLD is the similarity score threshold for the vector store.
   • TOP_K_RESULTS is the number of top results to return from the vector store.
   • CHATBOT_DB_URL is the URL of the chatbot database.
   • CHATBOT_DB_USER is the username of the chatbot database.
   • CHATBOT_DB_PASSWORD is the password of the chatbot database.
   • CHATBOT_CONTEXT_SERVICE_URL is the URL of the context service.
3. Replace all occurrences of 0.0.0.0 in the backend with localhost.
4. Run the backend.

5. Run the Frontend

1. Run the following command to install the frontend dependencies:

   cd Frontend-Re-Chatbot
   npm install

2. Run the following command to start the frontend:

   npm start

Production Setup

Prerequisites

• Make sure the Backend-Re has the correct gradle.properties file based on the chosen publishing method. For more information, refer to the README.md file of the Chatbot repository.

1. Run all components frm the Infrastructure

1. Add the correct values to the empty environment variables in the .env_prod file in the Infrastructure directory.
2. If you wish to change the default values of the environment variables, you can do so by editing the .env_prod file in the Infrastructure directory.
3. Run the following command to start the infrastructure services:

cd Infrastructure
docker compose -f docker-compose-prod.yml --env-file .env_prod up -d