ProvSyn

This project provides a provenance graph synthesis framework including graph structure synthesis, textual attribute synthesis, and multi-dimensional fidelity evaluation.

Before running the code, please make sure all dependencies are installed:

pip install -r requirements.txt

1. Graph Structure Synthesis

cd structure
python main.py                 # Train the graph generation model
python evaluate.py             # Generate graphs and compute MMD, Novelty, and Uniqueness

---

2. Text Attribute Synthesis

cd text
python lora_data.py            # Prepare training data
# Refer to training.ipynb for model training
# Refer to inference.ipynb for attribute synthesis

3. Fidelity Evaluation

cd fidelity

Dimension	Script
Text Attribute	python text.py
Temporal	python lcs.py, python dtw.py
Embedding	python netlsd.py, python graph2vec.py
Semantic	python semantic.py