SBST-DPG

Defect prediction guided search-based software testing (SBST-DPG) combines defect prediction and search-based software testing (SBST) to focus the test generation towards the defective classes in a project. SBST-DPG consists of three modules: i) Defect Predictor, ii) Budget Allocation Based on Defect Scores (BADS) and iii) SBST.

By default, SBST-DPG uses Schwa as the defect predictor, which outputs the likelihoods of classes being defective (also called defect scores). Any defect predictor that outputs defect scores for the classes in the project would also be suitable for SBST-DPG.

BADS takes defect scores from the defect predictor and allocates time budgets to classes based on their likelihood of being defective.

We use EvoSuite with DynaMOSA as SBST and it runs test generation for all the classes with the allocated time budgets by BADS.

See our paper, Defect Prediction Guided Search-Based Software Testing for more details.

How to use SBST-DPG?

Requirements

• JDK 1.8 (as required by EvoSuite)

Parameter Settings

All the parameters of SBST-DPG are defined in configs.yml file.

Parameter	Description	Default Value
Defect Predictor
name	Defect predictor name	Schwa
output_file	Output file name	schwa_results.csv
BADS
mode	BADS mode	exp
e_a	ea + eb . exp(ec . x)	0.02393705
e_b	ea + eb . exp(ec . x)	0.9731946
e_c	ea + eb . exp(ec . x)	-10.47408
tier_threshold	Threshold used in 2-Tier approach	0.5
grouping	Small group size1	5p
t_total	Available time budget per class (seconds)	60
t_total_first_tier	Available time budget per class in the first tier (seconds)	100
t_total_second_tier	Available time budget per class in the second tier (seconds)	20
t_min_first_tier	Minimum time budget for a class in the first tier (seconds)	60
t_min_second_tier	Minimum time budget for a class in the second tier (seconds)	10
SBST
src_path	Filepath to the source files of the project under test	src/main/java
sbst	SBST tool	EvoSuite
project_classpath	Classpath of the project under test	target/classes
Log
primary_logger_type	Primary logger type	console
log_level	Log level	info

¹All the classes in a small group get the same time budget. If you need only one class in a small group, then use 1 for grouping. 5p means small group size is 0.05 (=5%) * # classes in the project.

Steps to Run SBST-DPG

Currently, SBST-DPG can be run by executing the source code.

1. Clone SBST-DPG:

   • git clone https://github.com/SBST-DPG/sbst-dpg.git <sbst-dpg dir>

2. Setup Schwa:

   • Currently, Schwa needs to be setup manually. We are thinking of automating this process in the future.
   • git clone https://github.com/SBST-DPG/schwa.git <schwa dir>
   • cd <schwa dir>
   • git checkout csv-output
   • python3 setup.py bdist_wheel
   • cd dist
   • pip3 install Schwa-0.1.dev27-py3-none-any.whl

3. Create workspace directory for sbst-dpg in the root of the project under test. Configure the parameters listed above and locate the configs.yml file in the workspace directory.

4. Run SBST-DPG

   • cd <sbst-dpg dir>
   • python3 -u sbst_dpg/dpg_sbst.py <path to project under test>

Publications

If you use the SBST-DPG tool, please cite:

@inproceedings{perera2020defect,
  title={Defect prediction guided search-based software testing},
  author={Perera, Anjana and Aleti, Aldeida and B{\"o}hme, Marcel and Turhan, Burak},
  booktitle={2020 35th IEEE/ACM International Conference on Automated Software Engineering (ASE)},
  pages={448--460},
  year={2020},
  organization={IEEE}
}

More Information

• Contact Anjana Perera