Quickstart

This page walks you through your first run in under five minutes using the included toy example.

Run the toy walkthrough

After installing the requirements (see installation), run:

python3 tutorial/toy_walkthrough.py

This executes the full pipeline on a small synthetic dataset and produces:

an event log (event_log.csv)
a filtered event log (filtered_log.csv)
a Gantt chart visualization (gantt_chart.png)
a text summary of the run (results.txt)

Run the main pipeline

To run the main pipeline on the bundled pasteurization dataset:

python3 main.py

The default configuration in main.py uses an unsupervised HMM and the following feature plan:

feature_plan = {
    'statistical': ['T', 'Q_in', 'Q_out'],
    'temporal':    ['T', 'Q_in', 'Q_out'],
    'stability':   ['T', 'Q_in', 'Q_out'],
    'interaction': [['T', 'Q_in', 'Q_out']],
    'event': [
        '(T_diff_smooth > 1)',
        '(T_diff_smooth < -1)',
        '(Q_out > 0.3)',
        '(T > 70) & (T_stable_flag == 1)',
        '(Q_in > 0.3) AND (T_diff < 0.2)'
    ],
    'contextual': []
}

See Feature Plan Reference for the full DSL reference.