This repository provides the replication package for the paper titled Investigating the survival of open-source libraries
The repository contains all the data required to generate the plots shown in the paper and, thus, come to the conclusions presented in the paper.
A virtual environment is recommended but not necessary. Follow this simple guide to setting up a virtual environment on Linux or Windows: https://www.geeksforgeeks.org/creating-python-virtual-environment-windows-linux/ For Mac: https://www.geeksforgeeks.org/how-to-install-virtual-environment-in-python-on-macos/
The requirements.txt file should list all Python libraries that this project depend on, and they will be installed using:
pip install -r requirements.txt
- Navigate to the project directory replication-package-survival-analysis
- Install the dependencies using above command
- Run the project with
python src/main.py
The main.py is the only main file that needs for the results in output folder created after completing execution.
The merge_data.py is the initial step in getting the results, it combines the two data files pairs into one.
The clean_data.py is the second step, it removes unncessary fields not required for the analysis.
The regression_analysis.py performs logistic regression classification on the clean_data.csv with target label status field.
Other files stats_analysis.py and summary_stats.py are supporting files for the analysis.
This folder includes the data in CSV files from previous related research by Ait. et. al. https://github.com/SOM-Research/survivalAnalysisGitHub/tree/main/data.
There are 2 types of CSV files, each divided by ecosystem (see suffixes -laravel, -npm, -r and -wp for each file).
These are the files included in source folder:
-
metadata-ECOSYSTEM.csv, which, for each repository, has the information of whether it is a repository owned by a user or organization account (repoType), the status at the end of the previous study (used to answer RQ1 and RQ2) (status) -
oracle-ECOSYSTEM.csv, which complementsmetadata.csvproviding the number of users in the community (authors), and the number of project resources created along its lifespan (Commits,Pulls,Issues,CommentsandReviews).