Repository navigation
feat(forecast): Phase 4 – PR 4-D PV-Prognose mit Gütemessung, Blending, Kalibrierung und Clipping - #29
Open
PXLJAKE wants to merge 27 commits into
Open
feat(forecast): Phase 4 – PR 4-D PV-Prognose mit Gütemessung, Blending, Kalibrierung und Clipping#29PXLJAKE wants to merge 27 commits into
PXLJAKE wants to merge 27 commits into
Conversation
Snapshot so nothing is lost when the usage limit stops the session; not yet checked, squashed in spirit by the following WP5 commits. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VHXsfH6PkT62o9idByAd6e
…he database closes
…d of asyncio.wait
… a change during the refresh
… (WP5) Template-driven PV forecast sources (Solcast, Forecast.Solar via energy/solar_forecast, Open-Meteo Solar Forecast, generic attribute), refresh service with keep-last-good and hourly-or-material storage, nightly accuracy bookkeeping, API and demo forecasts. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VHXsfH6PkT62o9idByAd6e
… (WP5) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VHXsfH6PkT62o9idByAd6e
…class (WP5) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VHXsfH6PkT62o9idByAd6e
…ted inverter plateau (WP5) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VHXsfH6PkT62o9idByAd6e
…clipping-free data (WP5) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VHXsfH6PkT62o9idByAd6e
…(WP5) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VHXsfH6PkT62o9idByAd6e
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VHXsfH6PkT62o9idByAd6e
…ot every zero (WP5) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VHXsfH6PkT62o9idByAd6e
…source (WP5) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VHXsfH6PkT62o9idByAd6e
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VHXsfH6PkT62o9idByAd6e
… fresh one exists (WP5) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VHXsfH6PkT62o9idByAd6e
…)connects (WP5) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VHXsfH6PkT62o9idByAd6e
…vices first (WP5) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VHXsfH6PkT62o9idByAd6e
…log the runtime (WP5) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VHXsfH6PkT62o9idByAd6e
…ED (WP5) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VHXsfH6PkT62o9idByAd6e
…r-sky model Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VHXsfH6PkT62o9idByAd6e
…ping detection (WP5) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VHXsfH6PkT62o9idByAd6e
…red peaks, not a growing max (WP5) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VHXsfH6PkT62o9idByAd6e
…the boundary docstring (WP5) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VHXsfH6PkT62o9idByAd6e
…; regenerate the compatibility table (WP5) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VHXsfH6PkT62o9idByAd6e
…only (WP5) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VHXsfH6PkT62o9idByAd6e
…st sources The demo now has three PV forecast sources (WP5); the E2E check expected Solcast to be the only one. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VHXsfH6PkT62o9idByAd6e
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
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.
Gestapelt auf #26 (PR 4-B). Er hängt nicht von #27 (Tarife) ab. Die Basis wird nach dem Merge der Vorgänger umgestellt.
Was
WP5 aus
docs/plans/phase-4.md, Paketsuncore/forecast/pv/undsuncore/forecast/accuracy.py:QuantileSeriesin W):detailedForecast, 30 min, P10/P50/P90 aus kW. Die Tagessensoren heute bis d4 werden verbunden. Fehlt ein Tag, entsteht eineNone-Lücke, kein kopierter Nachbarwert.wh_period_15m, energieerhaltend in W umgerechnet.energy/solar_forecast(wh_hoursjeconfig_entry_id). Bei mehreren Einträgen wird nicht geraten. Die Integration lässt nur den 0-Wh-Wert um Mitternacht weg, dieser wird aufgefüllt. Schlüssel außerhalb des Rasters, etwa zum Sonnenaufgang, werden energieerhaltend der umgebenden Periode zugeschlagen; die Quelle fällt deswegen nie aus.PvForecastService:get_config.GET /api/v1/forecast/pv?hours=48: je Quelle die Reihe mit Aktualität, der Mix, die Gewichte und der Kalibrierungsstatus, roh und kalibriert zum Vergleich.GET /api/v1/forecast/accuracy: inklusiveblended.Warum
Der Planer (4-F) und das Risikomodell (4-E) brauchen eine 48-h-PV-Prognose mit Unsicherheit, die aus der gemessenen Güte der Quellen lernt (BRIEF §4.2, Plan §0).
Wie getestet
./scripts/check.shgrün: 3654 Python-Tests, 219 Frontend-Tests, Abdeckung 96 %, Bundle 132,7 kB. markdownlint ohne Fehler.Risiken
feat-Abschlusscommit (wegen der Nutzungslimits). Squash-Merge empfohlen, damit der CHANGELOG sauber bleibt.Safety-Review
n/a laut Reviewer: kein Geräte-Schreibpfad. Der neue WS-Befehl
energy/solar_forecastist nur lesend und durch Grenzregel 9 aufha/simbeschränkt.forecast/importiert keinstorage. Beim Stoppen werden die Geräte freigegeben, bevor auf den Nachtjob gewartet wird; dafür gibt es einen Test.Checkliste
enunddeget_configlatitude/longitude; Vorlagen-Changelog; Plan-Abweichungen)docs/STATUS.mdaktualisiertdocs/research/integrations.md, Offenes alsUNVERIFIEDOffene Fragen an den Maintainer
Für alle gilt ohne Antwort der Standard.
UNVERIFIED:energy/solar_forecastdurchreichtGenerated by Claude Code