feat: add runtime/duration-based task triggers (sensor value monitoring)

[Seidlm]

Runtime/Duration-Based Task Triggers

Related to #109 - extends the integration with a third trigger type.

Problem

Many smart home devices report cumulative usage metrics - for example, a Husqvarna robotic lawn mower tracks how many hours the blades have been running. Users need a way to trigger maintenance tasks based on these runtime values (e.g., "replace blades after 100 hours").

This can't be solved with time-based intervals (the mower doesn't run every day) or count-based triggers (it's not about on/off cycles, but accumulated hours).

Solution

A new trigger type: Runtime-based - monitors a numeric sensor entity and triggers a task when the accumulated value exceeds a threshold.

How It Works

1. User creates a task with Trigger Type: Runtime
2. Selects a sensor entity (e.g., sensor.husqvarna_blade_hours)
3. Sets a threshold (e.g., 100 hours)
4. The integration records the current sensor value as baseline
5. Task becomes due when: current_value - baseline >= threshold
6. On complete ? baseline resets to current sensor value
7. Cycle repeats

Automatic Sensor Reset Detection

Some devices reset their counters (e.g., after blade replacement the mower resets hours to 0). The integration handles this automatically:

• If current_value < baseline ? external reset detected
• Baseline is automatically adjusted to 0
• Counting continues from the new value

Example:

Step	Sensor Value	Baseline	Delta	Threshold	Due?
Task created	450h	450	0	100	No
Normal use	520h	450	70	100	No
Threshold hit	550h	450	100	100	Yes ?
Task completed	550h	550	0	100	No
Blades replaced, sensor resets	0h	0 (auto)	0	100	No
Normal use resumes	100h	0	100	100	Yes ?

Changes

Backend (Python):

• store.py - 3 new fields: runtime_entity_id, runtime_threshold, runtime_baseline. Auto-reads sensor value as baseline on task creation and completion.
• binary_sensor.py - New _update_state_runtime() with delta calculation, reset detection. Exposes attributes: runtime_entity_id, runtime_threshold, runtime_baseline, runtime_current, runtime_delta
• __init__.py - State change listener for runtime entities, detects external resets
• websocket.py - Accepts runtime fields on add/update, auto-populates baseline

Frontend (Panel):

• "Runtime-based" option in trigger type dropdown (create + edit)
• Entity selector filtered to sensor domain
• Threshold number input
• Table shows delta/threshold (e.g., "72.5 / 100") with red highlighting when due

Backward Compatibility

• Fully backward compatible - existing time-based and count-based tasks work unchanged
• New fields have safe defaults
• No breaking changes to existing services or WebSocket API

Use Cases

• ?? Robotic lawn mower blade replacement (hours)
• ??? Exercise equipment maintenance (usage hours)
• ?? Car service intervals (km via OBD sensor)
• ??? Printer maintenance (page count sensor)
• Any device that reports cumulative numeric usage data

Testing

Tested on Home Assistant (HAOS, Hyper-V) via HACS custom repository. Verified:

• Runtime delta calculation works correctly
• External sensor reset (to 0) is detected and handled
• Task becomes due at threshold
• Baseline resets on completion
• Existing time and count tasks unaffected