Julian Pawlowski 76dc488bb5 feat(sensors): add momentum-based trend detection with two new sensors ... Added intelligent price trend analysis combining historical momentum (weighted 1h lookback) with future outlook for more accurate trend recognition. Introduced two complementary sensors for comprehensive trend monitoring. New sensors: - current_price_trend: Shows active trend direction with duration - next_price_trend_change: Predicts when trend will reverse Momentum analysis (historical perspective): - Weighted 1h lookback (4 × 15-min intervals) - Linear weight progression [0.5, 0.75, 1.0, 1.25] - ±3% threshold for momentum classification - Recognizes ongoing trends earlier than future-only analysis Two-phase trend calculation: - Phase 1: Calculate momentum from weighted trailing average - Phase 2: Validate with volatility-adaptive future comparison - Combines both for final trend determination (rising/falling/stable) - Centralized in _calculate_trend_info() with 60s cache Volatility-adaptive thresholds: - Existing trend sensors (1h-12h) now use adaptive thresholds - calculate_price_trend() adjusted by market volatility: * LOW volatility (<15% CV): factor 0.6 → more sensitive (e.g., 3%→1.8%) * MODERATE volatility (15-30%): factor 1.0 → baseline (3%) * HIGH volatility (≥30%): factor 1.4 → less sensitive (e.g., 3%→4.2%) - Uses same coefficient of variation as volatility sensors - Ensures mathematical consistency across integration Default threshold reduction: - Rising/falling thresholds: 5% → 3% (more responsive) - Momentum-based detection enables lower thresholds without noise - Adaptive adjustment compensates during high volatility Architectural improvements: - Centralized calculation: Single source of truth for both sensors - Eliminates Henne-Ei problem (duplicate calculations) - 60-second cache per coordinator update - Shared helper methods: _calculate_momentum(), _combine_momentum_with_future() Translation updates (all 5 languages): - Documented momentum feature in custom_translations (de/en/nb/nl/sv) - Explained "recognizes ongoing trends earlier" advantage - Added sensor names and state options to standard translations - Updated volatility threshold descriptions (clarify usage by trend sensors) Files changed: - custom_components/tibber_prices/sensor/core.py (930 lines added) * New: _calculate_momentum(), _combine_momentum_with_future() * New: _calculate_trend_info() (centralized with cache) * New: _get_current_trend_value(), _get_next_trend_change_value() * Modified: _get_price_trend_value() (volatility-adaptive thresholds) - custom_components/tibber_prices/sensor/definitions.py * Added: current_price_trend (ENUM sensor) * Added: next_price_trend_change (TIMESTAMP sensor) - custom_components/tibber_prices/sensor/attributes.py * New: _add_cached_trend_attributes() helper * Support for current_trend_attributes, trend_change_attributes - custom_components/tibber_prices/price_utils.py (178 lines added) * New: _calculate_lookahead_volatility_factor() * Modified: calculate_price_trend() with volatility adjustment * Added: VOLATILITY_FACTOR_* constants (0.6/1.0/1.4) - custom_components/tibber_prices/entity_utils/icons.py * Added: Dynamic icon handling for next_price_trend_change - custom_components/tibber_prices/const.py * Changed: DEFAULT_PRICE_TREND_THRESHOLD_RISING/FALLING (5→3%) - custom_components/tibber_prices/translations/*.json (5 files) * Added: Sensor names, state options, descriptions - custom_components/tibber_prices/custom_translations/*.json (5 files) * Added: Long descriptions with momentum feature explanation Impact: Users get significantly more accurate trend detection that understands they're ALREADY in a trend, not just predicting future changes. Momentum-based approach recognizes ongoing movements 15-60 minutes earlier. Adaptive thresholds prevent false signals during volatile periods. Two complementary sensors enable both status display (current trend) and event-based automation (when will it change). Perfect for use cases like "charge EV when next trend change shows falling prices" or dashboard badges showing "Rising for 2.5h".		2025-11-16 12:49:43 +00:00
..
__init__.py	refactor(sensors): Transform sensor platform into package	2025-11-15 11:46:54 +00:00
attributes.py	feat(sensors): add momentum-based trend detection with two new sensors	2025-11-16 12:49:43 +00:00
core.py	feat(sensors): add momentum-based trend detection with two new sensors	2025-11-16 12:49:43 +00:00
definitions.py	feat(sensors): add momentum-based trend detection with two new sensors	2025-11-16 12:49:43 +00:00
helpers.py	refactor(sensors): Transform sensor platform into package	2025-11-15 11:46:54 +00:00