GraphRAG

GraphRag is a standalone application that uses Flask as backend and ReactJS as frontend, langchain , pinecone DB to store vector embeddings and SqLite to store chat history.

Table of Contents

• General Info
• Technologies Used
• Features
• Screenshots
• Setup
• Usage

General Information

GraphRAG uses the power of Generative AI and NLP to centralize all brand information in a user-friendly, model-agnostic platform. This innovative approach aims to: Reduce brand guideline misuse

• Accelerate access to brand knowledge for creatives
• Enable precise retrieval of brand information
• Alleviate the stress on creatives by simplifying prompt engineering

Technologies Used

• React JS
• Python
• Flask
• SqlLite
• Pinecone DB
• Langchain
• Open AI API

Features

• API endpoints
• Login and Signup
• Home Page
• Dashboard to view and upload new PDF
• Chat screen
• History dropdown
• RAG Pipeline on Vector Database and Orchestration tool like LangChain
• Standalone application that one can use any frontend to integrate APIS'

ScreenShots

Setup

Since this is a standalone application that uses APIS', one can ignore backend folder and jump to frontend folder

cd frontend
npm install // to install the dependencies
npm run dev // to run the local application

Note - If you want to start server locally then follow the steps:

cd backend

## Using Venv 

# Create the venv virtual environment
python -m venv .venv

# On MacOS, WSL, Linux
source .venv/bin/activate

# On Windows
.\.venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt


# Running the app 

This server is built using Flask with different API endpoints. So you just need to start the server and can use Frontend as a standalone application.

Commands to start are listed below. If you need to stop them, select the terminal window the process is running in and press Control-C

### To run the Python server

Open a new terminal window and create a new virtual environment:


# On MacOS, WSL, Linux
source .venv/bin/activate

# On Windows
.\.venv\Scripts\activate

Then:

inv dev

There are some environment variables that needed for correct functioning of local server. in backend folder, cd backend, create .env inside backend folder and paste the followings

SECRET_KEY=123
SQLALCHEMY_DATABASE_URI=
UPLOAD_URL=upload-url
OPENAI_API_KEY=openAI API key
PINECONE_API_KEY=pinecone API key
PINECONE_ENV_NAME=us-east-1
PINECONE_INDEX_NAME=pdf-docs

• Open a new terminal cd /path/to/backend and activate the .venv, use the below command to create new user

flask --app app.web add-user user@example.com supersecret

in frontend folder, cd frontend Then in .env inside frontend folder, paste

VITE_API_BASE_URL=http://localhost:5173

Usage

• Signup using email and password
• Once logged in, it will open a dashboard and upload button.
• User can upload new document by clicking on the upload button at top right
• User can click on view and will sent to the chat interface
• In Chat interface, user can chat and ask questions regarding the documents and brand guidelines. -There is a history dropdown where user can switch to the past conversations as well.
• Clicking on new chat button will open up the new conversation.

Note - In chat interface on left side is also a pdf viewer where user can view the uploaded docs while having a conversation. I have commented that code because it was giving performance issue if the uploaded file is large.
You can go to [FILE](frontend/src/page/document-view.tsx) and remove the comment.```