andreas-bauer / tax_se

Replication package for our work on "Taxing Collaborative Software Engineering"

Taxing Collaborative Software Engineering

Replication package for our work on "Taxing Collaborative Software Engineering"

Requirements

This replication package requires Python 3.10 or higher. Install the dependencies via:

python3 -m pip install -r requirements.txt

For a faster loading, we recommend to optionally install orjson via pip:

python3 -m pip install orjson

How to run

Step 1: Crawl

First, we collect all timelines from all pull requests at a GitHub instance. crawler.py requires an <api_token> for your GitHub instance and an <out_dir> where the results are stored into:

python3 crawl.py <api_token> <out_dir>

crawler.py also provides the following optional command line arguments:

• --api_url for the GitHub instance URL (default: https://api.github.com)
• --disable_cache for disable caching (for larger instances not recommended)
• --num_workers for parallel processes (default: 1)
• --organization for limiting to one organization (helpful for organizations hosted on github.com)

To list all options in detail, run:

python3 crawl.py --h

Step 2: Model pull requests as cross-border communication channels

For this step, you will need:

1. The directory of the previously collected data; and,
2. A mapping of users and countries. This can be either a dict for a static mapping (does not capture changes in the users' location over time) or a dataframe for time-dependent mapping as data frame monthly sampled (captures changes in the users' location over time).

Run notebook.ipynb. Look out for the instructions as inline comments.

License

Copyright © 2023 Michael Dorner.

This work is licensed under MIT license.