Add .env support, hot water correction, README, and bug fixes#2
Open
Appesteijn wants to merge 40 commits intoRickvdt:mainfrom
Open
Add .env support, hot water correction, README, and bug fixes#2Appesteijn wants to merge 40 commits intoRickvdt:mainfrom
Appesteijn wants to merge 40 commits intoRickvdt:mainfrom
Conversation
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Add README.md with usage guide for heat pump, gas, and transitioning users
- Move HA_URL and TOKEN to .env file (python-dotenv)
- Add hot water correction for gas-only users based on warm-day baseline
- Fix locals() vs globals() checks for inter-cell variable access
- Fix deprecated pandas resample('H') to resample('h')
- Fix timezone mismatch in gas data join
- Initialize df_gas_daily as empty to prevent stale data in downstream cells
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Convert Jupyter notebook analysis into a full HA integration with config flow, sensors, services, and an apexcharts-card dashboard. Move original notebook and files to archive. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
attr_fn was set to None instead of a lambda returning None, causing a TypeError when HA called extra_state_attributes. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Rename component from "Quatt Stooklijn Analyse" to "Quatt Warmteanalyse" across all user-visible locations (manifest, HACS, sensors, config flow, dashboard, README). Internal identifiers (domain, module paths) unchanged. - Fix pandas FutureWarning in quatt.py by dropping all-NA columns before DataFrame concatenation. - Add 94 unit/integration tests covering heat loss regression, stooklijn curve fitting, gas processing, sensor value/attr functions, config flow validation, and coordinator logic. - Add .gitignore entries for __pycache__ and .pytest_cache. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The gas analysis period typically predates the heat pump installation, so outdoor temperature from Quatt hourly data was always empty (no overlapping dates). Now fetches temperature directly from the HA recorder for the gas date range using the configured temp_entities. Falls back to heat pump data if recorder fetch finds nothing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Change the last_analysis sensor to display the date in yyyy-mm-dd format instead of using the timestamp device class which shows a localized human-readable format. Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
- Extend heat_at_temps to include interpolated COP values per temperature - Add new columns to dashboard table: COP and electricity consumption - Calculate accurate electricity usage based on temperature-specific COP - Update tests for new heat_at_temps structure This provides more accurate electricity consumption estimates by using interpolated COP values for each temperature instead of a single average, giving users better insight into heat pump efficiency across different temperatures. Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Major performance and accuracy improvements: CACHING SYSTEM: - Add persistent cache for Quatt insights data to minimize API calls - Implement 30-day initial fetch limit to prevent API abuse for new users - Cache grows organically (1 day per run) to full history over time - 99.6% reduction in API calls after initial run (251 calls → 1 call) - Cache stored in .storage/quatt_stooklijn_insights_cache KNEE DETECTION IMPROVEMENTS: - Use all available Quatt hourly data (grows from 30 to 250+ days) - Smart filtering removes defrost cycles and partial operation hours - Fallback to recorder data if Quatt unavailable - 3x to 25x more data than previous 10-day recorder method - Progressive accuracy as cache grows CHANGES: - Add custom_components/quatt_stooklijn/cache.py (new cache helper) - Update quatt.py with caching and 30-day limit - Update stooklijn.py with improved knee detection - Extensive README documentation on caching and performance - Add comprehensive documentation files BENEFITS: - New users: Safe 30 API calls first run, grows automatically - Existing users: Instant analyses with 1 API call per run - Better accuracy: More historical data for knee detection - No configuration needed: Works transparently Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
After a restart, dashboards showed empty graphs because analysis data is in-memory only. Now automatically triggers analysis after EVENT_HOMEASSISTANT_STARTED, using cached data for fast completion. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Instead of fetching all historical data from the Quatt API (which was limited to 30 days on first run anyway), now uses HA recorder long-term statistics for the full configured period. The Quatt API is only used for the last 30 days of hourly detail data (for knee detection). This gives months of daily data immediately without any API calls, and the API data overwrites recorder data for recent days where it's more accurate. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
With recorder statistics now providing months of data including summer, the COP average and stooklijn regression were skewed by days with no heating demand. Now filters on totalHeatPerHour >= 200W and COP > 0 to only include meaningful heating days. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The recorder's heatpump_total_quatt_cop sensor averages over 24h including off-periods (COP=0), giving artificially low values on warm days when the pump runs fewer hours. Now calculates COP as totalHpHeat/totalHpElectric which matches the Quatt API behavior and gives the correct operational COP. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Previously, cached hourly data older than 30 days was ignored. Now checks cache for the full configured period before fetching from the API, giving more data points for knee detection. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The recorder DataFrame used int64 (0) for totalBoilerGas while the API data contains floats, causing a pandas dtype warning on df_daily.update(). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Scatter data for both stooklijn and heat loss charts included summer days with 0W heating, causing the x-axis to extend to ~29°C. Now uses the same heating_data filter as the regression. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Rewrites the "How it works" and "Performance & Caching" sections to document the recorder + API + cache hybrid approach, adds auto-startup info, and updates troubleshooting with COP tips. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
All trend/stooklijn lines were drawn from -10°C to 25°C regardless of actual data. Now they stop at max(scatter_data temps) + 1°C, so lines don't extend far beyond the data points. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Lines with shallow slopes (like Quatt estimated) extended far beyond the relevant range because y > 0 was the only limit. Now each line stops at min(-intercept/slope + 1, 20)°C, which is where heat demand reaches zero. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
graph_span reduced from 35h to 30h so x-axis stops at 20°C, matching the data range. Beyond 20°C there is no heating demand. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The Quatt API returns hourly averages, which makes the 2500W power filter unreliable (partial operation hours pass the filter). This resulted in an artificially shallow slope (-133 W/°C vs expected -318). Now uses HA recorder state_changes (minute-level granularity) for the stooklijn slope estimation, matching the original notebook approach. Quatt hourly data is still used for knee detection where it works well. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
More data points give a more reliable stooklijn estimation, especially if recent days had mild weather. This is a local database query only (no Quatt API calls). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Documents the 4-source hybrid approach: recorder statistics, recorder state changes (minute-level), Quatt API, and cache. Explains why minute-level data is used for stooklijn estimation. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The dashboard trendlines were recalculating slope/intercept from scatter point averages, which is inaccurate and caused the line to go horizontal at the right end. Now uses the actual regression slope and intercept from the sensor attributes. Also exposes slope, intercept, and balance_point on both heat loss sensors (HP and gas) for use in the dashboard. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Add Dutch, German, and French translations - Fix codeowners to @Appesteijn Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add hassfest, HACS validation, and pytest workflows - Fix conftest.py stubs for recorder.statistics and helpers.storage - All 94 tests pass Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Vervang curve_fit door exhaustieve grid search (-4 tot +4°C, stap 0.25°C) om lokale minima te vermijden; deterministisch en debugbaar - Voeg fysieke beperkingen toe aan grid search: warme kant moet negatieve helling hebben, koude kant substantieel vlakker - Stel recorder minuutdata in als primaire bron voor knikdetectie (geen defrost-verdunning van uurgemiddelden → nauwkeuriger, stabiel op 1.75°C) - Quatt uurdata blijft als fallback bij onvoldoende koude-weer recorder-data - Herschrijf KNEE_DETECTION_IMPROVEMENTS.md met validatieresultaten en onderbouwing van databron-keuze Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
scipy wordt niet meer gebruikt na vervanging van curve_fit door grid search. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
.envsupport — MovedHA_URLandTOKENto.envfile usingpython-dotenv, added.env.exampleHOT_WATER_OUTSIDE_TEMP_THRESHOLD), with DHW summary in outputlocals()vsglobals()for inter-cell variables, deprecated pandasresample('H'), timezone mismatch in joins, stale data from failed fetches.gitignore, moved older notebook versions toarchive/Test plan
python-dotenvis includedHA_URL/TOKENfrom.envlocals()or timezone errors🤖 Generated with Claude Code