How p4zaa/graphprepkit works
Project type and purpose
• Project type: a Python library (open-source toolkit) for graph data preprocessing.
• Purpose: provide utilities to transform raw tabular graph data into edge lists, index mappings, and visualizations for homogeneous and heterogeneous graphs.
File structure
• Top-level package “GraphPrepKit”
• construct.py – functions to build connection tables (edge lists)
• homo.py – functions for homogeneous graph mapping
• hetero.py – functions for heterogeneous graph mapping
• dataset/ – example data (CSV files) for demonstration or tests
• setup.py – packaging and dependency declarations
• README.md – usage instructions and high-level overview
README for additional insights
• Key operations:
• build connection dataframes from raw source→target lists
• Unique IDs to integer indices for model ingestion
• support for both “homo” and “hetero” graph types
• reliance on pandas for data manipulation and networkx for plotting