gen-ai-knowledgebase

Project Setup Guide

Welcome to the RAG Knowledgebase project! This guide will walk you through the steps to get the application up and running on your local machine.

---

Prerequisites

Before you begin, ensure you have the following installed:

• Python 3.8+: The project is built with Python.
• pip: Python's package installer.
• Git: For cloning the repository.
• Docker (optional): For running the Milvus vector database locally.

---

1. Clone the Repository

First, clone the project from its Git repository to your local machine:

git clone <repository_url>
cd rag_knowledgebase

---

2. Set Up the Environment

This project uses a .env file to manage environment-specific variables, including API keys and configuration settings.

A. Create the .env file

A template .env.template is provided to show you the required variables.

1. Copy the template file to create your local .env file:

   cp .env.template .env

2. Open the newly created .env file in a text editor.

B. Configure Environment Variables

Fill in the values for each variable in your .env file. Do not share this file with anyone or commit it to your Git repository.

• SECRET_KEY: A unique, random string used by Flask for session management. You can generate one using os.urandom(24).
• FLASK_ENV: For local development, keep this as development.
• GEMINI_KEY: If you plan to use Google's Gemini API, provide your API key here.
• MILVUS_URI and MILVUS_TOKEN: If you are using Zilliz Cloud, find these connection details in your Zilliz Cloud dashboard.
• SUPABASE_URL and SUPABASE_KEY: If you are using Supabase for storage, get these from your Supabase project settings.
• SUPABASE_BUCKET_NAME: The name of your storage bucket in Supabase.
• DEFAULT_MODEL: The name of the AI model you want to use for generating responses (e.g., gemini-pro).
• ENABLE_OCR, EXTRACT_TABLES, EXTRACT_IMAGES: Set these to true or false based on your document processing needs.

Example .env file with values:

# Flask Configuration
SECRET_KEY=a_very_secret_and_random_string
FLASK_ENV=development

# OpenAI API Key (optional, for response generation)
GEMINI_KEY=YOUR_GEMINI_API_KEY

# Vector Database (Milvus/Zilliz Configuration)
MILVUS_HOST=localhost
MILVUS_PORT=19530
COLLECTION_NAME=document_embeddings

# Connection details from your Zilliz Cloud dashboard
MILVUS_URI=YOUR_ZILLIZ_URI
MILVUS_TOKEN=YOUR_ZILLIZ_TOKEN

# Supabase Configuration
SUPABASE_URL=YOUR_SUPABASE_URL
SUPABASE_KEY=YOUR_SUPABASE_KEY
SUPABASE_BUCKET_NAME=rag_files

# Storage Configuration
USE_SUPABASE_STORAGE=true
STORE_ORIGINAL_FILES=true

# AI/LLM Configuration
DEFAULT_MODEL=gemini-pro

# DocLing Processing Configuration
ENABLE_OCR=true
EXTRACT_TABLES=true
EXTRACT_IMAGES=true

---

Create config.py file in the app.py root directory Example config.py file with values:

import os
from dotenv import load_dotenv

load_dotenv()

class Config:
    # Flask/FastAPI settings
    SECRET_KEY = os.environ.get('SECRET_KEY') or 'a-default-secret-key'
    MAX_CONTENT_LENGTH = 50 * 1024 * 1024  # 50MB max file size

    # Milvus/Zilliz settings
    MILVUS_HOST = os.environ.get('MILVUS_HOST', 'localhost')
    MILVUS_PORT = os.environ.get('MILVUS_PORT', '19530')
    MILVUS_URI = os.environ.get('MILVUS_URI')
    MILVUS_TOKEN = os.environ.get('MILVUS_TOKEN')
    COLLECTION_NAME = os.environ.get('COLLECTION_NAME', 'document_embeddings')
    DIMENSION = 384

    # Supabase settings
    SUPABASE_URL = os.environ.get('SUPABASE_URL')
    SUPABASE_KEY = os.environ.get('SUPABASE_KEY')
    SUPABASE_BUCKET_NAME = os.environ.get('SUPABASE_BUCKET_NAME', 'uploads')

    # Storage settings
    USE_SUPABASE_STORAGE = os.environ.get('USE_SUPABASE_STORAGE', 'False').lower() == 'true'
    STORE_ORIGINAL_FILES = os.environ.get('STORE_ORIGINAL_FILES', 'True').lower() == 'true'
    UPLOAD_FOLDER = 'uploads'
    ALLOWED_EXTENSIONS = {'pdf', 'docx', 'doc', 'pptx', 'ppt', 'txt', 'md', 'html', 'rtf'}

    # Text processing settings
    CHUNK_SIZE = 1000
    CHUNK_OVERLAP = 200

    # AI/LLM settings
    GEMINI_KEY = os.environ.get('GEMINI_KEY')
    DEFAULT_MODEL = os.environ.get('DEFAULT_MODEL', 'gpt-3.5-turbo')

    # Embedding model
    EMBEDDING_MODEL = 'sentence-transformers/all-MiniLM-L6-v2'

    # DocLing settings
    ENABLE_OCR = os.environ.get('ENABLE_OCR', 'True').lower() == 'true'
    EXTRACT_TABLES = os.environ.get('EXTRACT_TABLES', 'True').lower() == 'true'
    EXTRACT_IMAGES = os.environ.get('EXTRACT_IMAGES', 'True').lower() == 'true'

---

3. Install Dependencies

Create a virtual environment and install the required Python packages:

python -m venv .venv
source .venv/bin/activate  # On Windows, use `.venv\Scripts\activate`
pip install -r requirements.txt

---

4. Run the Application

Now you can start the Flask application:

flask run

The application should now be running on http://127.0.0.1:8000. You can access the API endpoints or the front-end interface from there.