Tools & Datasets

Tool choice depends on language and research question. The list below is intentionally modular.

Parsing and structural analysis

• Eclipse JDT (Java AST and bindings)
• JavaParser / Spoon (Java source transformation)
• srcML (multi-language structural representation)
• tree-sitter (incremental parsing across many languages)
• ANTLR (custom grammars and language tooling)

Static analysis and smells

• PMD
• SonarQube / Sonar analyzers
• Designite (language-specific editions)
• Understand or other commercial code-analysis platforms where licensed

Refactoring / transformation

• IDE refactoring engines (IntelliJ, Eclipse, Roslyn)
• Spoon / OpenRewrite for Java transformations
• Roslyn analyzers/code fixes for .NET
• LibCST/Bowler-style tools for Python, where appropriate

Search and experimentation

• DEAP / pymoo / jMetal-family frameworks
• NetworkX for dependency graphs
• pandas / SciPy / scikit-learn for analysis and baselines
• Docker/Podman for environment capture

Subject systems

Prefer mature open-source projects with:

• reproducible tags/commits;
• buildable tests;
• clear license;
• non-trivial history;
• enough size/variation for the research question.

Avoid constructing the entire evaluation from toy pattern examples unless the project specifically studies synthetic benchmarks.