This repository is a companion page for the submission "Know Your Neighbor: Fast Static Prediction of Test Flakiness".
It contains all the material required for replicating the experiments, including: the algorithm implementation, the datasets and their ground truth, and the scripts for the experiments replication.
In order to replicate the experiment follow these steps:
-
Clone the repository:
git clone https://github.com/FlakyFAST/FLAST
-
If you do not have python3 installed you can get the appropriate version for your OS here.
-
Install the additional python packages required:
python -m pip install -r requirements.txt
Decompress the dataset:
tar zxvf dataset.tgz
Execute the research questions scripts.
python params-k.py
(varying k)python params-dist.py
(varying distance)python params-eps.py
(varying epsilon)python params-sigma.py
(varying sigma)
python training-size.py
python single-projects.py
(RQ3 & RQ4, effectiveness and running time)python storage.py
(RQ4, storage overhead)python random-classifier.py
(comparison with random classifier)
The pseudocode of FLAST is available here.
This is the root directory of the repository. The directory is structured as follows:
FLAST
. Scripts with FLAST implementation and scripts to run experiments.
|
|--- dataset/ Dataset folder, automatically generated after the decompression of `dataset.tgz`.
|
|--- manual-inspection/ Tests considered in the manual inspection
|
|--- pseudocode/ The pseudocode of FLAST.