Improve Caelus forecast synopses and hourly metrics

[mot-yelraf]

Summary

Both forecast tiles now show a synopsis without enlarging their existing layout. Open-Meteo and MET Norway summaries use hourly conditions, temperatures, wind, and precipitation probabilities; NWS uses its supplied period narrative when available, with a generated fallback. Open-Meteo weather codes preserve snow and thunder conditions.

Caelus adds humidity and wind ranges and hourly details. Hourly metrics use the same RH/Wind/Rain-or-Snow label order as Looking Ahead, whose precipitation text now matches the other metric colors. Long synopses support keyboard scrolling. Updates the user guide and version to v0.26.253.5.

Verification

• User verified the forecast presentation.
• macOS, Python 3.13.9; host mode, no broker or direct hardware required.
• python3 -m pytest testApparatus/test_weather_forecast.py testApparatus/test_integrated_weather_forecast.py testApparatus/test_compile_python.py -q: 54 passed.
• npm run validate:pr: 25 passed in local Chromium, including desktop/mobile layout limits, metric units, and long synopsis keyboard scrolling.
• git diff --check: passed.

Contributor checklist

• This pull request is focused on one logical change.
• I added or updated tests for behavior changes.
• I updated the canonical documentation when behavior or configuration changed.
• I did not commit credentials, private runtime configuration, databases, or logs.
• I preserved compatibility-sensitive MQTT, settings, switch identity, and persistence behavior, or documented an intentional break.

Maintainer verification

• Required remote checks pass.
• npm run validate:pr passes on a trusted host when the change can affect rendered UI behavior.
• Hardware-, broker-, onboarding-, and platform-specific behavior not covered remotely is recorded below.

Residual risk or unverified areas

Automated provider tests use fixtures and mocked HTTP, including NWS narrative failure and period boundaries. Live provider responses were not exercised by the automated checks. NWS narrative text retains its original units; numeric metrics follow the selected display units.