Synthetic Data

An automated flowchart generation system that creates and validates Mermaid diagrams using Large Language Models (LLMs). The project generates synthetic flowcharts across multiple topics and difficulty levels, with built-in quality validation and multi-format output support.

Results

The project generates flowcharts and converts them to different formats. Here are some example outputs:

Matplotlib Generated Flowchart

Mermaid Generated Flowchart

Package Manager

This project uses uv as the Python package manager for faster dependency resolution and installation.

Installation

First, install uv:

# On macOS and Linux
curl -LsSf https://astral.sh/uv/install.sh | sh

# On Windows
powershell -c "irm https://astral.sh/uv/install.ps1 | iex"

# Or with pip
pip install uv

If using Linux is probably that you have to add uv to your path, we can add the follow line to the shell config (like ~/.bashr).

export PATH="$HOME/.cargo/bin:$PATH"

Update your development workflow sections:

Development

Setup

# Clone the repository
git clone https://github.com/jorgemunozl/Synthetic-Data.git
cd Synthetic-Data

Install dependencies

Use

uv sync

Activate the virtual environment (if needed)

source .venv/bin/activate  # Linux/macOS
.venv\Scripts\activate     # Windows

Add Ids memory to the llm Detect flag true false ifexist checkpoint.sqlite trivial