agi-app-weather-forecast

agi-app-weather-forecast

agi-app-weather-forecast publishes the weather_forecast_project AGILAB app as a self-contained PyPI payload. It is the public notebook-migration example for turning a forecasting notebook into an executable AGILAB project.

Purpose

Use this package to validate the notebook-to-app path: a small Meteo-France sample dataset is forecasted, metrics are exported, and analysis pages can inspect the resulting predictions.

Installed Project

The distribution name is agi-app-weather-forecast; the AGILAB project name is weather_forecast_project. The package exposes weather_forecast and weather_forecast_project through the agilab.apps entry point group, so AgiEnv(app="weather_forecast_project") works without a monorepo checkout. The retired weather_forecast_legacy and weather_forecast_legacy_project names remain provider and import aliases when this package is installed with its matching AGILAB core release; new work should use the maintained name.

Install

pip install agi-app-weather-forecast

Most users get this package through agi-apps, agilab[ui], or agilab[examples]; direct installation is useful when validating one app package in isolation.

Run In AGILAB

Select weather_forecast_project, open ORCHESTRATE, then run Deploy scheduler & workers and RUN. Open view_forecast_analysis from ANALYSIS; use view_release_decision when you want baseline-versus-candidate promotion evidence.

Expected Inputs

The default run uses a bundled sample weather CSV. No live Meteo-France call, cloud account, private dataset, or API key is required.

Expected Outputs

The run writes forecast metrics, prediction CSV files, reducer summaries, and analysis-ready artifacts under the weather-forecast output paths.

Change One Thing

Change the forecast horizon or input window, then rerun the app. The metrics artifact should make the impact visible before you promote or reject the run.

Scope

This is a migration and reproducibility demo. It does not claim production forecast serving, live weather ingestion, drift monitoring, or model governance.