AI Application Generator

Project Summary

This project is a tool for automatically generating a complete application with a Domain-Driven Design (DDD) model and application services from requirements and commands. It leverages Large Language Models (LLMs) to interpret natural language requirements and generate clean, well-structured code that follows DDD and onion architecture principles.

Key Features

• Model Generation: Convert natural language requirements into a complete domain model with entities, value objects, aggregates, and repositories
• Application Services Generation: Create application services based on commands that interact with the domain model
• Infrastructure Code Generation: Generate the infrastructure layer based on the domain model and 'infrastructure requirements'
• Interface Generation: Generate an interface for the application (currently restricted to a CLI)
• Project Files Generation: Generate the project files, e.g. README, third-party dependencies, etc.
• Command-Line Interface: Easy-to-use CLI for all operations

How It Works

1. Requirements Collection: Users define model requirements, commands and infrastructure requirements in natural language
2. App Generation: The system uses LLMs to interpret requirements and generate a the different application layers (onion architecture)
3. Code Output: Generated code is written to the file system in the proper structure

Milestones

• The ai-app-generator can generate itself from natural language requirements (CLI only)
• The generated application can be built incrementally by adding and modifying requirements
• Go beyond CLI applications, i.e. serverless
• Build a SaaS for generating applications from requirements
• Build a comprehensive requirements engineering SaaS that provides the base for generating applications using AI

Usage Guide

Installation

# Clone the repository
git clone [email protected]:Flugtiger/ai-app-generator.git
cd ai-app-generator

# Install dependencies
pip install -e .

Creating Requirements

Create a model requirement:

python -m src.interface.cli.main model-requirement create --text "A model requirement"

Create a command:

python -m src.interface.cli.main command create --name MyFirstCommand --description "A command description"

Create an infrastructure requirement:

python -m src.interface.cli.main infrastructure-requirement create --text "An infra requirement"

Generating an Application

You can generate the layers of an application one by one (in the correct order: application and infrastructure depend on the model; the interface depends on application, model and infrastructure; the project files depend on the interface and infrastructure layers):

python -m src.interface.cli.main generate model
python -m src.interface.cli.main generate application
python -m src.interface.cli.main generate infrastructure
python -m src.interface.cli.main generate interface
python -m src.interface.cli.main generate project

You can also generate the complete application in one go:

python -m src.interface.cli.main generate all

Generate the ai-app-generator recursively

Warning

Generating the ai-app-generator itself currently costs about 0,60$ (using claude-3-7-sonnet-latest)

The ai-app-generator can generate itself from its own requirements. To do so, use the following workflow:

./ingest_model_reqs.sh
./ingest_commands.sh
./ingest_infra_reqs.sh
python -m src.interface.cli.main generate all

Note: There is currently an incompatibility between the libraries 'tree-sitter' and 'tree-sitter-languages' that are used by Claude to implement the CodeCompressor service. To fix this, replace 'tree-sitter-languages' with 'tree-sitter-language-pack' in the generated app. Sadly, Claude doesn't know about this new library yet.

Building and Development

Prerequisites

• Python 3.8 or higher
• Access to an LLM API (currently configured for Anthropic Claude)

Development Setup

1. Clone the repository:

   git clone [email protected]:Flugtiger/ai-app-generator.git
   cd ai-app-generator

2. Create a virtual environment:

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

3. Install in development mode:

   pip install -e .

4. Set up your API key:

   export ANTHROPIC_API_KEY=your_api_key_here
   # On Windows: set ANTHROPIC_API_KEY=your_api_key_here

Running Tests

pytest

Project Structure

• src/application: Contains application services
• src/model: Contains the domain model
• src/infrastructure: Contains infrastructure implementations
• src/interface: Command-line interface
• requirements: Requirements for the ai-app-generator itself

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

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