Chat PDF Application

A Django application that allows users to chat with PDF documents using vector embeddings and PostgreSQL with pgvector.

Prerequisites

• Docker
• Docker Compose

Tech Stack

• Python 3.12
• Django
• PostgreSQL with pgvector
• ChromaDB
• UV package manager (instead of pip)
• pgAdmin

Getting Started

Clone the repository

git clone <repository-url>
cd chat-pdf

Running with Docker Compose

1. Build and start the containers:

docker compose up --build

This will start the following services:

• Django application (http://localhost:8000)
• PostgreSQL with pgvector
• pgAdmin (http://localhost:5050)
• ChromaDB (http://localhost:8080)

Running Locally (Without Docker)

1. Set up a Python virtual environment (Python 3.12 recommended):

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

2. Install UV package manager here:

3. Install dependencies:

uv sync

4. Install and configure PostgreSQL:

• Install PostgreSQL 15 or later
• Enable pgvector extension:

CREATE EXTENSION IF NOT EXISTS vector;

• Create a database named 'pdf_chat'

5. Set up environment variables in a .env file:

POSTGRES_DB=your_database_name
POSTGRES_USER=your_username
POSTGRES_PASSWORD=your_password
POSTGRES_HOST=localhost
POSTGRES_PORT=5432

6. Run migrations:

uv run manage.py migrate

7. Start the development server:

uv run manage.py runserver

The application will be available at http://localhost:8000

Note

When running locally, make sure you have:

• PostgreSQL 15+ installed with pgvector extension
• Python 3.12+ installed
• System dependencies for pdf processing (python-magic requirements)
  • On Ubuntu/Debian: sudo apt-get install libmagic1
  • On MacOS: brew install libmagic
  • On Windows: No additional steps needed