Reproducible Artifacts for Skill-Aware Explainable Framework for Reliable Code Smell Detection in Human–AI Collaborative Programming

This repository provides the official computational framework, datasets, and reproducibility artifacts for the paper: "Skill-Aware Explainable Framework for Reliable Code Smell Detection in Human–AI Collaborative Programming". The project is designed to support transparent, reproducible, and stage-by-stage empirical analysis of human-written and AI-generated Python code, structured into three core research stages: 1. Stage 1 (Data Source): Processing raw human and AI code datasets.2. Stage 2 (Structural Ground Truth): Metric extraction (Pylint, Radon, JSCPD) and smell consensus modeling.3. Stage 3 (Human-AI Evaluation Framework): Simulating pairing scenarios, analyzing decision robustness, and generating XAI (e.g., faithfulness, stability) explanation artifacts. ### Key Included Artifacts:- Core datasets: 'human_feature_label_matrix.csv', 'ai_feature_label_matrix.csv', and 'final_features_with_skill_and_correctness.csv'.- Evaluation logs and decision mapping files: 'simulated_pair_selection.csv', 'simulated_pair_features.csv', 'mapping_table.csv', and 'explanation_artifacts.csv'.- Source code scripts structured for pipeline replication under Python 3.12.- Pre-rendered visualization plots supporting the publication's empirical findings. For environment setup and execution order instructions, please refer to the main README.md file inside the archive.