Metta Atomspace Graph Demo

This project demonstrates how to create, query, and visualize a small graph using MeTTa and AtomSpace via the Hyperon Python bindings.

What the Script Does

• Graph Construction:
  The script builds a simple directed graph representing relationships (e.g., "Alice knows Bob") using MeTTa code. It inserts nodes and links into AtomSpace through the Hyperon Python API.

• AtomSpace Querying:
  It attempts to read back the graph structure (OrderedLink atoms) from AtomSpace using several fallback methods, adapting to different Hyperon/MeTTa API versions.

• Visualization:
  The recovered graph triples are converted into a DOT graph and rendered as a PNG image using the graphviz Python package.

• Export:
  The script saves the graph visualization as graph.png and the recovered triples as triples.json.

Example Output

Below is an example of the generated graph:

Requirements

• Python 3.8+
• hyperon Python package (pip install hyperon)
• graphviz Python package (pip install graphviz)
• System graphviz installed for PNG rendering (sudo apt install graphviz on Ubuntu)

Usage

1. Install dependencies:

   pip install hyperon graphviz
   sudo apt install graphviz
   

2. Run the script:

   python test.py
   

3. View the output files:
   • graph.png: Visual representation of the graph
   • triples.json: List of recovered triples

Notes

• The script is robust to changes in the Hyperon/MeTTa API and will fall back to local data if AtomSpace querying fails.
• You can modify the graph structure by editing the triples list in the script.