A machine-learning based ensemble method to aggregate various anti-patterns detection approaches on the basis of their internal detection rules and, thus, improve detection performances.
SMAD is implemented for the detection of God Class and Feature Envy.
This repository is organized as follows:
- approaches: Source code and data related to the approaches implemented in this work, including the replication of HIST, InCode, Vote, ASCI, as well as the Tensorflow code of the NN used by our approach SMAD.
- data: Contains necessary data to run the experiments.
- antipatterns: The oracle, i.e., manually-tagged occurrences of God Class and Feature Envy in eight software systems.
- entities: Names of the classes and methods considered for each of the studied system.
- data_construction: Code used to generate the data.
- oracle_feature_envy: To create the data contained in ~/data/antipatterns/feature_envy/ from the answers collected via our survey.
- repository_miner: To mine systems' repository and create the rest of the data.
- experiments: The source code of our experiments: training, tuning, and performance comparison.
- java: Jars and src of the Java code implemented in this work. These jars are used in ~/data_construction/repository_miner/repository_miner.py to extract the data related to HIST, InCode, DECOR and JDeodorant.
- utils: Some utility modules.
This code has been tested under Python 2.7.16, and the following steps assume you have already installed Python 2.7 on your machine.
- Create a Python 2.7 virtual environment: To be able to run the code you should preferably setup this repository within a virtual environment. The command below uses virtualenvwrapper to do so.
mkvirtualenv --python=/usr/bin/python2.7 smad
workon smad
- Add the root of the repository to your PYTHONPATH:
export PYTHONPATH="$PYTHONPATH:PATH_TO_SMAD_DIRECTORY"
- Install requirements:
pip install -r requirements.txt
The paper associated to this repository has been published in Journal of Systems and Software.