<aside> 📘

Tutorial 13 of the Kalix tutorial series. You'll run the same Stringybark Creek calibration as Tutorial 12 — but driven from Python with kalix.optimise(), reading the result back as a dictionary, then simulating and plotting the calibrated model in a notebook. Expected time: about 20 minutes.

</aside>

What you'll build

A Jupyter notebook that calibrates the Stringybark catchment model in one call, inspects the optimised parameters as a pandas object, and overlays the calibrated simulation against the observed record. This is the analysis-friendly counterpart to Tutorial 12's command-line workflow — same optimisation, same config file, driven from Python.

Prerequisites

• Tutorial 12 complete — see Tutorial 12 — Optimisation from the commandline. We reuse its exact optimisation config and model, so the concepts (terms, statistics, parameter ranges, algorithms) carry straight over. This tutorial only changes how you launch the run.
• Tutorial 5 complete — see Tutorial 5 — Running Kalix from Python. We build on the same pip install kalix package and the simulate() / pandas / matplotlib pattern you learned there.
• Python 3.9 or newer with Jupyter installed.
• Tutorial files — download the 013/ folder from the KalixTutorials repository. It also ships a fully-worked analysis.ipynb if you'd rather skim than type along.

Project layout

013/
├── data/
│   ├── climate_data.csv
│   └── observed.csv
└── models/
    ├── stringybark.ini              # the starting model
    ├── optimisation_config.ini      # the calibration recipe (same as Tutorial 12)
    └── analysis.ipynb               # the notebook we'll write

The notebook sits next to the model and config files, so the relative paths inside the config (../data/observed.csv) and the notebook resolve cleanly. Launch Jupyter from 013/models/.

The CLI and Python, side by side

The kalix Python package mirrors the CLI. Tutorial 12's command:

kalix optimise optimisation_config.ini stringybark.ini -s stringybark_calibrated.ini

becomes, in Python:

result = kalix.optimise(
    "optimisation_config.ini",
    model_file="stringybark.ini",
    save_model="stringybark_calibrated.ini",
)

Same three pieces — the config, the model, and where to save the calibrated model. The difference: the CLI prints a summary to the terminal, while Python returns the results to you as a dictionary you can work with directly.

<aside> 💡

Everything from Tutorial 12 — Optimisation from the commandline still applies — the [optimisation], [term.*], and [parameters] sections of the config are read exactly the same way. If you want to change the algorithm, statistic, or parameter ranges, you edit optimisation_config.ini, not the Python.

</aside>

Step 1 — Open a notebook in the model folder