Near-duplicate build failure detection analysis files and Oodikone dataset

Near-duplicate build failure detection analysis files and Oodikone dataset

All of the outputs, intermediate output files, as well as the scripts used for data analysis are included in this package.

Before running any of the scripts, please create a virtual environment using Python 3.11 and install the python libraries in requirements.txt

Here's an example using conda.

Create a new conda environment with Python 3.11

conda create -n oodikone python=3.11

Activate the newly created environment

conda activate oodikone

Navigate to the directory containing the replication package

cd [Directory of the replication package]

Install the required dependencies from requirements.txt

pip install -r requirements.txt

The folder paths for the current implementation assume the replication package directory as the root folder.  If you run into errors, most likely they are related to incorrect folder paths, so please check that the folder paths are correct.

The original raw dataset of log files are in the folder "logs_all_original_input".

The OOVD filtered log files are in the folder "logs_all_OOVD_input". 

If you wish to perform OOVD filtering separately, please navigate to the folder "oovd_filtering" and run "create_oovd_filtering.ipynb".

Note that this is unnecessary, as all of the OOVD outputs are already in the folder "logs_all_OOVD_input". 

For Top-K results, please run config_analyzer_top_K.ipynb

For Top-R results, please run config_analyzer_top_R.ipynb