Open the project in PyCharm, setup the vent as Python 3.7.0, and install dependencies:

pip install -r requirements.txt

Edit SmartCommitEvaluation-Viz/config.py to change the raw data directory:

# root dir of raw data
root_path = str(Path.home()) + '/smartcommit/viz/'

For Controlled Open Source Experiment:

• Run rq1/rq1-boxplot-project_method.py to generate rq1-baselines.pdf: Accuracy of SmartCommit and 3 baseline methods (Figure 7(a) in the paper)
• Run rq1/rq1-boxplot-project_length.py to generate rq1-length.pdf: Accuracy of SmartCommit for different numbers of merged atomic commits (Figure 7(b) in the paper)
• Run rq2/rq2-adjustments.py to generate reassign_frequency.txt and reorder_frequency.txt: Number of reassign and reorder steps and proportions.

For Industrial Field Study:

• Run rq3/rq3-scatterplot.py to generate rq3.pdf: the runtime cost with different input sizes;
• Run rq4/preprocess.py first, and other python files then to generate usage distributions (Figure 9 in the paper).