Replication Package for Synthesizing Static Rules for Code Smell Detection

A 2026 replication package by Zijie Huang for research on code smell detection. It includes the MLCQ benchmark with 14,739 annotations from 522 repositories, 1,840 developer evaluations, and 40 qualitative interview transcripts. The package contains datasets, source code for a Java subsystem, model implementations, and analysis scripts.

Use Cases

• Benchmarking code smell detection models based on the MLCQ dataset with 14,739 annotations
• Evaluating developer perceptions of detected smells based on 1,840 human evaluations
• Training or testing machine learning models like CodeBERT-Fusion or DCTO for software metric analysis
• Conducting qualitative analysis on code quality based on 40 developer interview transcripts

Strengths

• Includes a large benchmark dataset (MLCQ) with 14,739 annotations across 522 software repositories
• Contains human evaluation data from 1,840 developer assessments for validation
• Provides replication scripts and pre-computed results for reproducibility
• Includes source code for a real-world Java subsystem (~5.3 KLOC) for case study

Limitations

• Column-level documentation is absent; field semantics must be inferred after download
• Row count for individual datasets is unknown, which may limit suitability assessment
• Data may reflect bias inherent to the specific repositories and developers sampled

Provenance

Source

figshare

Collection Method

Research replication package containing benchmark data, human study results, and source code.

Freshness

Last updated 2026-04-16 08:56:57