Quickstart

This quickstart shows how to trace a script, create a baseline, and detect a performance regression.

1. Create a sample script

Create a file named my_app.py:

import time


def process_data():
    time.sleep(0.10)
    calculate_results()


def calculate_results():
    time.sleep(0.20)


def main():
    for _ in range(2):
        process_data()


if __name__ == "__main__":
    main()

2. Run a trace

oracletrace my_app.py

You will see:

• Top functions by total time
• Call counts and average time per call
• Call graph logic flow

3. Save a baseline trace

oracletrace my_app.py --json baseline.json

Keep this file as your reference run.

4. Simulate a regression and compare

Increase one of the sleep values in your script, then run:

oracletrace my_app.py --json new.json --compare baseline.json

Comparison output highlights:

• Functions that got slower
• Functions that got faster
• Newly added or removed functions

5. Understand the output quickly

Example flow:

<module>
└── my_app.py:main
    └── my_app.py:process_data
        └── my_app.py:calculate_results

OracleTrace is optimized for relative change detection between runs, which makes it ideal for everyday checks and CI validation.