Status: MVP auf aare.guru
I bi vou binech das momentan z viu AI-Hype ume isch aber weisch wi geil we dr Aare Guru scho paar Täg im vorus ungfähr wüsst wi warm d Aare de isch u wiviu Wasser si de het?
Füre Bevölkerigsschutz gits scho ä Vorhersag zum Wasserabfluss (d Wassermängi im Aare.guru) vom Bund, i nime a di isch scho sehr guet u chönnt direkt verwendet werde. Übrig blibt auso d Temperaturvorhersag, wo vermuetlech eifacher isch aus d Abflussvorhersag, drum probiere mrs eis :)
Relies on uv and DVC. Uses just to simplify commands. Only tested on Linux with bash and zsh.
First install uv, then use it to install just with:
uv tool install rust-justThen creating the Python environment is as simple as
just syncTo run the pipelines, do this. These currently just calculate the baselines.
just reproTo run the notebooks, make sure to set the kernel to the lokal venv (.venv/bin/python) in your IDE.
In vscode, when it asks after first time execution, set the environment to the recommended ".venv" env.
If you want to use JupyterLab, you need to install the kernelspec first: just install-kernel
Then run jupyter lab. You'll need to change the kernel in the top right for every notebook.
To start the mlflow ui and the optuna dashboard for experiment tracking and hyperparameter tuning, use the following commands in separate terminals:
just mlflow
just optunaOr use
just trackIf you want to work with the code instead of just running it, I suggest you install nbstripout (already in venv) and prek (uv tool install prek).
To set them up, run the following.
uv run nbstripout --install --attributes .gitattributes
prek install -fD Date chöme vom aare.guru säuber oder vom BAFU, aggregiert im wahnsinnig tolle Archiv vom Bureau für digitale Existenz.
- Aare.guru API Description
- InfluxDB Access ❤️
- SwissMetNet Dataset Description
- BAFU Hydrologie Dataset Description
- Hydrologische Vorhersagen vom Bund (nüt dopplet mache)
AGPL-3.0
The deployed API has additional usage restrictions as you can see when you visit it.