tosem2021-personality-rep-package 
Replication package for the manuscript "Using Personality Detection Tools for Software Engineering Research: How Far Can We Go?" under review at ACM TOSEM.
Requirements
Ensure that your box fulfills the following requirements:
- Python 3.8.3+
- R 4.0.4+
- Java 1.8+
- python3-venv
- pandoc
- LaTeX
- homebrew (macOS-only)
On macOS, you can install LaTeX and pandoc by executing brew install mactex pandoc.
On Ubuntu, run sudo apt install texlive-latex-extra pandoc.
Finally, after installing Java, ensure that the environment variable $JAVA_HOME is properly set.
Cloning
The repository uses git submodules. To clone the code and its submodules, run git clone --recursive <repo-url.git>
Setup
Setup instructions are contained in the file setup.sh and must be run only once. The script has been tested on macOS Big Sur and Ubuntu 20.04 LTS. To complete the setup, simply run bash setup.sh. Please, note that the installation of the R packages will prompt you to enter the your password via sudo.
If you have anaconda installed, ensure that the conda base environment is not active during setup.
Reproducible pipeline
To reproduce the pipeline, run bash repro.sh:
Usage: repro.sh [-h] [-v] -s all|phase1|phase2 -d test|full
Available options:
-h, --help Print this help and exit
-v, --verbose Print script debug info
-s, --stage Pipeline stage. Accepted values: all, phase1, phase2
-d, --dataset Dataset size. Accepted values: test, full
You can choose to reproduce the full pipeline by passing the argument -s|--stage all. Otherwise, you can reproduce individually the two phases of the experiment by passing instead the arguments phase1 or phase2. Also, to reduce the computational time, you can chose between running the script on the full dataset or on a random test subsample.
Notes
-
Phase Two is quite time-consuming---it can take hours, in not days, depending on your box specs. For test purposes, you can supply the argument
testto work with a very small, random subsample of the original dataset and keep the computational time within minutes. -
The script
ph1_0-goldstandard_creation.shis intentionally not part of the reproducible pipeline. It is meant to be executed only once, to create the anonymized gold standard in which email addresses have been replaced with hashed ids and all the sensitive content from emails (e.g., names, urls) have been scrubbed. This is intended to prevent others from tracking down the participants by searching for matching text into the public email archives of the Apache Software Foundation.
3 Aug 29, 2022
148 Nov 30, 2022
0 Jan 27, 2022
11 Dec 13, 2022
1.2k Dec 21, 2022
3 Nov 24, 2022
38 Dec 17, 2022
315 Dec 31, 2022
0 Dec 23, 2021
189 Dec 20, 2022
2 Dec 28, 2021
137 Dec 31, 2022
1.6k Jan 06, 2023
321 Dec 29, 2022
3.5k Jan 08, 2023
91 Dec 23, 2022
276 Jan 07, 2023
172 Dec 14, 2022
23 Oct 21, 2022
774 Dec 31, 2022