PSALM

This repository contains the experimental code for the paper:

PSALM: applying Proportional SAmpLing strategy in Metamorphic Testing

The code is research-oriented and intended to support the experiments reported in the paper.

---

What this repository is for

• Implementing PSALM (a proportional sampling strategy) and several baselines (Random, ART, MT-ART).
• Running metamorphic testing experiments on multiple Java and Python subject programs.
• Comparing PSALM with baseline strategies under metamorphic testing.
• Producing experimental results, tables and figures for RQ1–RQ3 in the paper.

---

High-level structure

• java/
  Java subject programs, metamorphic relations, generators, mutant integration, and experiment drivers.

• python/
  Python subject programs with the same experimental structure as the Java side.

• RQ1/, RQ2/, RQ3/
  Analysis scripts that aggregate raw results and generate tables/figures for the paper.

---

Subjects

• Java subjects
  jackson_project, jfreeChart_project, lang_project, math1_project, math2_project

• Python subjects
  geometricSum, incomeTax, mortgageRate

Each subject follows a similar layout:

• generation/phase1, generation/phase2 – test or MG sampling strategies
• metamorphicRelations/ – subject-specific MRs
• model/ – test case and MR abstractions
• mutants/ – externally generated mutants
• raw_results/ – experiment drivers and JSON outputs
• mutants_analysis/ – optional mutant classification and subsumption analysis

---

Experimental workflow

1. Prepare mutants externally
   Generate mutants using external tools (mutmut and Major) and place them into the expected directory structure.

2. Run per-subject experiments
   Use subject-specific raw result scripts (e.g., get_raw_results_*.java or get_raw_results_*.py) to run PSALM and baseline strategies.

3. Aggregate results
   Run scripts under RQ1/, RQ2/, and RQ3/ to compute statistics and generate tables/figures.

---

Quick start

• Java experiments: run java/<subject>/raw_results/get_raw_results_*.java.
• Python experiments: run python/<subject>/raw_results/get_raw_results_*.py.
• Mutants must be generated externally before running experiments.

---

Notes on usage

• This is research code meant to accompany the paper.
• Reproducing results requires:
  • Reading the paper for experimental settings.
  • Preparing mutants in advance.
  • Adjusting paths and configurations to your environment.
• Code-level understanding is expected.
• The mutants files are not included in this repository due to their size. Please generate them using the external tools.

---