Ad Composer

Ad Composer is a full-stack application for personalizing landing pages using AI. It uses Django with REST Framework for both the web interface and API, integrated with OpenAI and Temporal for workflow orchestration.

Demo

Generic Peronsalization

Deep Personalization

System Architecture

• Web & API: Django application with DRF API endpoints and OpenAI integration
• Database: PostgreSQL
• Workflow Engine: Temporal for asynchronous workflow orchestration
• Development/Production: Docker containerization

Project Structure

AD-COMPOSER/
├── assets/           # Asset files
├── db/               # Database files
│   └── init/         # Database initialization
├── web/              # Django web application
    ├── ad_composer/  # Main Django app
    ├── config/       # Django project configuration
    ├── myenv/        # Virtual environment
    └── staticfiles/  # Static files

API Endpoints

Get Account Names

• GET /api/account-names
  • Returns a list of available account names from the database
  • Response: Array of account names

Personalize Text

• POST /api/personalize
  • Personalizes text content for a specific client using AI
  • Request Body:

    {
      "client": "client_name",
      "texts": ["text1", "text2", ...]
    }

  • Response:

    {
      "client": "client_name",
      "originalTexts": ["text1", "text2", ...],
      "personalizedContent": ["personalized1", "personalized2", ...]
    }

Fetch URL

• GET /fetch-url/?url=https://example.com
  • Fetches a web page and returns its HTML content
  • Response: HTML content of the requested page

Prerequisites

• Docker and Docker Compose
• Python 3.9+ (for local development)
• PostgreSQL 16+ (for local development)

Quick Start with Docker

1. Clone the repository:

git clone https://github.com/yourusername/ad-composer.git
cd ad-composer

2. Set up environment variables:

cp .env.example .env
# Edit .env with your configuration including OpenAI API key

3. Start the services:

docker-compose up --build

The application will be available at:

• Web Application: http://localhost:8000
• Temporal UI: http://localhost:8080

4. Test Temporal connection:

docker exec -it ad-composer-web-1 python test_temporal.py

• http://localhost:8000

Local Development Setup

1. Create and activate virtual environment:

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

2. Install dependencies:

pip install -r requirements.txt

3. Run migrations:

python manage.py migrate

4. Start the development server:

python manage.py runserver

Database Setup

The database is automatically initialized with Docker. For local development:

1. Create PostgreSQL database:

createdb addb

2. Run initialization scripts:

psql -d addb -f db/create-db.sql

Environment Variables

# Database Configuration
DB_USER=ad_user
DB_HOST=db
DB_DATABASE=addb
DB_PASSWORD=your_secure_password
DB_PORT=5432

# Django Configuration
SECRET_KEY=your_secure_django_secret
DEBUG=False
ALLOWED_HOSTS=localhost,127.0.0.1

# API Keys
OPENAI_API_KEY=your_openai_api_key

# Temporal Configuration
TEMPORAL_HOST=temporal:7233

Temporal Workflow Engine

This application uses Temporal as a workflow engine to orchestrate asynchronous ad generation processes. Temporal provides:

• Durable execution of workflows
• Automatic retries and error handling
• Scalable processing of multiple targets in parallel
• Visibility and monitoring through the Temporal UI

Temporal UI

The Temporal UI is available at http://localhost:8080 when running with Docker. Use it to:

• Monitor workflow executions
• View workflow history and details
• Debug failed workflows
• Manually trigger workflows

Testing Temporal Connection

To verify that Temporal is properly connected:

# When running with Docker
docker exec -it ad-composer-web-1 python test_temporal.py

# When running locally
python web/test_temporal.py

Production Deployment

The application is configured for deployment using Docker:

docker-compose up --build

License

This project is licensed under the MIT License - see the LICENSE file for details.