v0.5.1: Data updates and robustness

New: Configurable update interval and robustness fixes

You can now set how often the integration polls the API (1–24 hours, default 8) via Configure in the integration settings.

Bug fixes

• Allergy risk sensors now update properly — the daily and hourly allergy risk sensors previously read from a snapshot taken at startup and never reflected coordinator refreshes. They now read live data, matching the behaviour of pollen allergen sensors. Sensors created during an API outage (is_stale) automatically recover when data returns.

• Options flow changes actually take effect — country, coordinates, language and API key were previously read only from the initial config. Changes made via Configure are now applied to both the coordinator and sensors.

• Accurate sensor attributes — last_updated now shows the actual API fetch timestamp (not the time you read the attribute), and update_success reflects the coordinator's real status. Timezone-aware throughout.

• Proper integration lifecycle — added async_unload_entry so changing options no longer leaves orphaned coordinators or timers. Coordinator now receives config_entry as required by recent HA versions.

• Better API error reporting — non-JSON responses (HTML error pages, empty bodies) are now detected before parsing, logged with a body preview, and classified correctly as server errors.

• Duplicate detection respects reconfigured locations — adding a new entry no longer ignores location changes made via Configure.

• Options flow consistency — sanitized update interval default (prevents form crash on corrupted config), aligned fallback title format with config flow.

• Complete translations — the options flow UI is now translated in all 15 supported languages.

Other

• Added test infrastructure with 85 unit tests

---

Full Changelog: v0.5.0...v0.5.1