v1.0.0 - Health Score, Smart Notifications & Apple Home UI

What's New in v1.0.0

This is a major release that transforms Cardio4HA into a premium, intelligent device health monitoring solution.

Health Score (0-100)

• Weighted score combining unavailable devices (40%), low batteries (25%), weak signals (20%), and flaky devices (15%)
• New sensor.cardio4ha_health_score entity for automations

Flaky Device Detection

• Statistical analysis identifies unstable devices (frequent offline/online cycles)
• Devices flagged with purple "Unstable" badge across all views
• New sensor.cardio4ha_flaky_devices entity

Battery Prediction

• Linear regression on 30-day battery readings
• "X days left" prediction shown in battery view
• Expandable rows show battery trend chart

30-Day Device History

• Persistent event tracking (offline/online, battery, signal)
• Statuspage.io-style timeline bar per device
• Signal and battery trend mini-charts in expanded rows

Smart Notifications

• Instant alerts: Device offline, battery critically low, mass offline (3+ devices)
• Recovery notifications: When devices come back online
• Daily digest: Morning summary when critical issues exist
• 3-layer rate limiting: Global hourly limit, per-device cooldown, mass offline grouping
• Configurable via options flow (Step 2: Notifications)

Apple Home Style Panel

• Large animated health score ring
• Expandable device rows with detail panels
• 30-day uptime timeline bars
• 5-bar signal strength indicators
• Battery bars with prediction badges
• Critical issues banner
• Recent notifications feed
• Responsive design (desktop/tablet/mobile)

Other Changes

• Config flow upgraded to v2 with automatic migration for existing users
• 2-step options: Thresholds & Exclusions → Notifications
• New WebSocket commands: get_device_timeline, get_notification_history
• New service: clear_device_history
• Old Lovelace card removed (sidebar panel only)

Upgrade Notes

• Existing v0.3.0 installations will auto-migrate config to v2
• All existing stores (unavailable tracking, maintenance) preserved
• New stores (device_history) initialize empty and populate over time
• Battery predictions require ~5 readings to start (a few hours of operation)
• Flaky detection needs 30 days of history for accurate results