baumicoder / improving-source-code-readability-theory-and-practice Goto Github PK

• change-distiller.csv

  • contains ChangeDistiller data

• ref-miner.csv

  • contains RefactoringMiner data

• source-file-samples/*

  • contains all before/after code samples (For downloading this content see download-samples.py)

• after-checkstyle.csv

  • checkstyle warnings for all after files

• before-checkstyle.csv

  • checkstyle warnings for all before files

• oracle.csv

  • our oracle of commits, files in the commits and the commit message

• readability-metrics.csv

  • contains readability scores calculated using Scalabrino's model, Dorn's model and Combined Model.

A python script for downloading the before and after version of the files from the commits into the source-file-samples directory. It reads the metadata of the commits and files from the oracle.csv.

Make sure you have the required python packages installed. For this you can run the following command to install them with pip:

pip install "requests==2.*"
pip install "patch==1.*"

Create a GitHub API token with at least the right to read public repositories here. Then call the script with the token as parameter, like this:

./download-samples.py GitHub_API_Token_place_here

The download takes about 30 minutes.

Some commits are not available anymore. Actual known (in March 2024):

Affected rows produce appropriated message on stderr like this:

ERROR: Commit, from row 410 in oracle.csv, not found.

In the oracle.csv are also files listed, which were renamed in the corresponding commit. So there are no content changes, so nothing to evaluate in matter of readability.

Affected rows produce appropriated message on stdout like this:

Row 391 of oracle.csv skipped, because the file was only renamed.