Merge pull request 'ACT-001: actuator-first sensor assignment and lifecycle' (#31) from feature/actuator-sensor-lifecycle into main

Closes #21

.env.example

@@ -1,3 +1,9 @@
 SILLYHOME_HA_URL=http://homeassistant.local:8123
 SILLYHOME_HA_TOKEN=REPLACE_ME_WITH_LONG_LIVED_TOKEN
 SILLYHOME_MODEL_STORE=.model_store
+SILLYHOME_AUTOMATION_STORE=.automation_store
+SILLYHOME_ACTUATOR_STORE=.actuator_store
+SILLYHOME_HISTORY_DAYS=14
+SILLYHOME_MIN_TRAINING_POINTS=24
+SILLYHOME_RETRAIN_STALE_HOURS=24
+SILLYHOME_RECONCILE_INTERVAL_SECONDS=900

CHANGELOG.md

@@ -1,6 +1,18 @@
 # Changelog
 
-## Unreleased
+## 0.4.0 - 2026-06-13
+- Aktuator-zentrierte Einrichtung: Nutzer wählen nur noch Aktuatoren, Sensoren werden deterministisch gefunden und bewertet
+- Persistente automatische und manuelle Sensorzuordnungen mit Evidenz, Confidence, Review-Gating und Neustart-Sicherheit
+- Autonomer Modell-Lebenszyklus auf echter HA-Historie: Training, Retraining bei Staleness oder Datenänderung, Archivierung von Waisen
+- Neues Dashboard und API für Aktuatorauswahl, Reconciliation, Overrides, Modellstatus und Audit-Trail
+- Neue Container-/Add-on-Defaults für Aktuator-Store und periodische Reconciliation ohne zusätzliche Gerätesteuerung
+
+## 0.2.0 - 2026-06-13
+- Klassifizierte Home-Assistant-Entity-Discovery mit Lernrelevanz und Filtern
+- Validierter Zugriff auf die Home-Assistant-History-API
+- Normalisierte, chronologisch sortierte numerische Zeitreihen über `/v1/history`
+- Trainierbares statistisches Baseline-Modell mit persistierten Parametern
+- Numerische Vorhersagen mit Confidence sowie MAE-/RMSE-Evaluation
 
 ## 0.1.0 - 2026-06-13
 - Projektinitiierung

Dockerfile

@@ -4,6 +4,12 @@ ENV PYTHONDONTWRITEBYTECODE=1 \
     PYTHONUNBUFFERED=1 \
     PIP_NO_CACHE_DIR=1 \
     SILLYHOME_MODEL_STORE=/app/data/models
+ENV SILLYHOME_AUTOMATION_STORE=/app/data/automations \
+    SILLYHOME_ACTUATOR_STORE=/app/data/actuators \
+    SILLYHOME_HISTORY_DAYS=14 \
+    SILLYHOME_MIN_TRAINING_POINTS=24 \
+    SILLYHOME_RETRAIN_STALE_HOURS=24 \
+    SILLYHOME_RECONCILE_INTERVAL_SECONDS=900
 
 WORKDIR /app
 
@@ -14,7 +20,7 @@ COPY app ./app
 COPY backend ./backend
 RUN python -m pip install --upgrade pip && \
     python -m pip install . && \
-    mkdir -p /app/data/models && \
+    mkdir -p /app/data/models /app/data/automations /app/data/actuators && \
     chown -R sillyhome:sillyhome /app/data
 
 EXPOSE 8000

README.md

@@ -4,10 +4,10 @@ Lokaler, datenschutzfreundlicher API-Prototyp für Home Assistant.
 
 ## Reifegrad
 
-Version `0.1.0` stellt eine gehärtete technische Basis bereit: Home-Assistant-Entities
+Die aktuelle Entwicklungslinie ist aktor-zentriert: Nutzer konfigurieren nur
-lesen, regelbasierte Bausteine und eine persistente Modell-Artefakt-Registry. Die
+noch Home-Assistant-Aktuatoren. SillyHome Next findet dazu passende numerische
-aktuelle Trainings- und Vorhersagelogik ist noch eine deterministische
+Sensoren und Kontext-Entities, zeigt Evidenz und Review-Bedarf an und hält
-Schnittstellen-Implementierung und **kein produktives Machine-Learning-Modell**.
+passende Modelle lokal und autonom aktuell.
 
 ## Motivation
 TheSillyHome zeigte die Idee: statt statischer Regeln das Zuhause aus Verhaltensmustern verstehen. Diese Architektur modernisiert den Ansatz in Richtung Explainable AI, hybride Intelligenzebenen und langlebige Wartbarkeit.
@@ -39,11 +39,19 @@ uvicorn app.main:app --reload
 ```
 
 4. Erreichbar unter:
+- `http://127.0.0.1:8000/` - lokales Dashboard
 - `http://127.0.0.1:8000/health` - Health-Check
 - `http://127.0.0.1:8000/docs/` - OpenAPI-Dokumentation
 - `http://127.0.0.1:8000/v1/entities` - Home-Assistant-Entities
+- `http://127.0.0.1:8000/v1/discovery` - klassifizierte, filterbare Entities
+- `http://127.0.0.1:8000/v1/history` - normalisierte numerische Zeitreihen
+- `http://127.0.0.1:8000/v1/actuators/discovery` - unterstützte Aktuatoren für den aktor-zentrierten Workflow
+- `POST http://127.0.0.1:8000/v1/actuators` - Aktuator registrieren, Sensorzuordnung prüfen und Modell-Lebenszyklus starten
+- `POST http://127.0.0.1:8000/v1/actuators/reconciliation/run` - globale Reconciliation manuell anstoßen
 - `http://127.0.0.1:8000/ml/health` - Registry-/Serving-Health
 - `POST http://127.0.0.1:8000/ml/retrain` - Modell-Metadaten aktualisieren
+- `POST http://127.0.0.1:8000/ml/evaluate` - MAE/RMSE/Coverage berechnen
+- `POST http://127.0.0.1:8000/v1/automations/proposals` - sicheren Entwurf anlegen
 
 Ohne vollständige HA-Konfiguration liefert `/v1/entities` bewusst `503`.
 
@@ -62,10 +70,39 @@ dem Netz muss ein authentifizierender Reverse Proxy vorgeschaltet werden.
 - `SILLYHOME_HA_URL` – Basis-URL deiner Home-Assistant-Instanz (z. B. `http://homeassistant.local:8123`)
 - `SILLYHOME_HA_TOKEN` – Long-Lived Access Token eines dedizierten HA-Benutzers mit minimalen Rechten
 - `SILLYHOME_MODEL_STORE` – Verzeichnis für persistierte Modell-Metadaten
+- `SILLYHOME_AUTOMATION_STORE` – Verzeichnis für Automation-Entwürfe
+- `SILLYHOME_ACTUATOR_STORE` – Verzeichnis für persistente Aktuator-Zuordnungen, Overrides und Reconciliation-Status
+- `SILLYHOME_HISTORY_DAYS` – Trainingsfenster für HA-History (1 bis 31 Tage)
+- `SILLYHOME_MIN_TRAINING_POINTS` – Mindestanzahl nutzbarer numerischer Messpunkte vor einem Modelltraining
+- `SILLYHOME_RETRAIN_STALE_HOURS` – Staleness-Grenze für automatisches Retraining
+- `SILLYHOME_RECONCILE_INTERVAL_SECONDS` – Intervall für sichere periodische Reconciliation
 
 Niemals Administrator-Tokens oder Passwörter eintragen. `.env` gehört nicht ins
 Versionskontrollsystem.
 
+### Home-Assistant-Add-on
+
+Das Repository ist zugleich ein Home-Assistant-Add-on-Repository. In Home Assistant
+unter **Einstellungen → Add-ons → Add-on-Shop → Repositories** diese URL eintragen:
+
+`http://192.168.6.31:3000/pino/sillyhome-next`
+
+Danach **SillyHome Next** installieren und starten. Das Dashboard wird per Ingress
+geöffnet. Das Add-on nutzt die Supervisor-API nur lesend; Automation-Entwürfe werden
+lokal gespeichert und niemals automatisch ausgeführt.
+
+### Normaler Workflow
+1. Im Dashboard oder per API einen Aktuator auswählen, zum Beispiel `light.abstellkammer`.
+2. SillyHome Next bewertet passende numerische Sensoren und binäre Kontext-Entities anhand von Bereich, Gerät, Namen, Domain und `device_class`.
+3. Starke und eindeutige Zuordnungen werden automatisch genutzt; schwache oder knappe Kandidaten bleiben mit Review-Hinweis sichtbar.
+4. Manuelle Overrides haben Vorrang, bleiben persistent und überstehen Neustarts.
+5. Sobald genügend numerische HA-Historie vorhanden ist, trainiert das System automatisch ein lokales Modell pro Aktuator-Zuordnung und retrainiert es bei relevanten Datenänderungen oder Staleness.
+
+Vor einem Update sollte in Home Assistant unter **Einstellungen → System → Backups**
+eine Teil-Sicherung des Add-ons erstellt werden. Zur Wiederherstellung das gewünschte
+Backup öffnen, **SillyHome Next** auswählen und wiederherstellen. Der erste produktive
+Teststand `v0.3.0` wurde als HA-Backup `7df0fca0` gesichert.
+
 ### Tests
 ```bash
 pytest

addon/Dockerfile

@@ -0,0 +1,19 @@
+FROM python:3.13-slim
+
+ENV PYTHONDONTWRITEBYTECODE=1 \
+    PYTHONUNBUFFERED=1 \
+    PIP_NO_CACHE_DIR=1
+
+RUN apt-get update \
+    && apt-get install -y --no-install-recommends git \
+    && git clone --depth 1 --branch main \
+       http://192.168.6.31:3000/pino/sillyhome-next.git /app \
+    && python -m pip install --upgrade pip \
+    && python -m pip install /app \
+    && rm -rf /var/lib/apt/lists/* /app/.git
+
+COPY run.sh /run.sh
+RUN chmod 0755 /run.sh
+
+EXPOSE 8000
+CMD ["/run.sh"]

addon/config.yaml

@@ -0,0 +1,31 @@
+name: SillyHome Next
+version: "0.4.0"
+slug: sillyhome_next
+description: Lokale HA-Analyse, Vorhersagen und sichere Automation-Entwürfe
+url: http://192.168.6.31:3000/pino/sillyhome-next
+arch:
+  - amd64
+startup: application
+boot: auto
+init: false
+ingress: true
+ingress_port: 8000
+panel_title: SillyHome Next
+panel_icon: mdi:home-analytics
+panel_admin: true
+homeassistant_api: true
+hassio_api: false
+auth_api: false
+options:
+  history_days: 14
+  min_training_points: 24
+  retrain_stale_hours: 24
+  reconcile_interval_seconds: 900
+schema:
+  history_days: "int(1,31)"
+  min_training_points: "int(2,10000)"
+  retrain_stale_hours: "int(1,720)"
+  reconcile_interval_seconds: "int(60,86400)"
+map:
+  - type: addon_config
+    read_only: false

addon/run.sh

@@ -0,0 +1,19 @@
+#!/bin/sh
+set -eu
+
+export SILLYHOME_HA_URL="${SILLYHOME_HA_URL:-http://supervisor/core}"
+export SILLYHOME_HA_TOKEN="${SILLYHOME_HA_TOKEN:-${SUPERVISOR_TOKEN:-}}"
+export SILLYHOME_MODEL_STORE=/data/models
+export SILLYHOME_AUTOMATION_STORE=/data/automations
+export SILLYHOME_ACTUATOR_STORE=/data/actuators
+
+if [ -f /data/options.json ]; then
+  export SILLYHOME_HISTORY_DAYS="$(python -c 'import json; print(json.load(open("/data/options.json")).get("history_days", 14))')"
+  export SILLYHOME_MIN_TRAINING_POINTS="$(python -c 'import json; print(json.load(open("/data/options.json")).get("min_training_points", 24))')"
+  export SILLYHOME_RETRAIN_STALE_HOURS="$(python -c 'import json; print(json.load(open("/data/options.json")).get("retrain_stale_hours", 24))')"
+  export SILLYHOME_RECONCILE_INTERVAL_SECONDS="$(python -c 'import json; print(json.load(open("/data/options.json")).get("reconcile_interval_seconds", 900))')"
+fi
+
+mkdir -p "$SILLYHOME_MODEL_STORE" "$SILLYHOME_AUTOMATION_STORE" "$SILLYHOME_ACTUATOR_STORE"
+exec uvicorn app.main:app --app-dir /app --host 0.0.0.0 --port 8000 \
+  --proxy-headers --forwarded-allow-ips='*'

app/actuators/__init__.py

@@ -0,0 +1,27 @@
+from app.actuators.lifecycle import (
+    ActuatorReconciliationService,
+)
+from app.actuators.models import (
+    ActuatorRecord,
+    AssignmentCandidate,
+    AssignmentSelection,
+    LifecycleAuditEntry,
+    LifecycleStatus,
+    ManualOverride,
+    ReconciliationState,
+    model_id_for_actuator,
+)
+from app.actuators.store import ActuatorStore
+
+__all__ = [
+    "ActuatorReconciliationService",
+    "ActuatorRecord",
+    "ActuatorStore",
+    "AssignmentCandidate",
+    "AssignmentSelection",
+    "LifecycleAuditEntry",
+    "LifecycleStatus",
+    "ManualOverride",
+    "ReconciliationState",
+    "model_id_for_actuator",
+]

app/actuators/lifecycle.py

@@ -0,0 +1,607 @@
+from __future__ import annotations
+
+import hashlib
+import logging
+import re
+from collections.abc import Iterable
+from datetime import datetime, timedelta, timezone
+
+from app.actuators.models import (
+    ActuatorRecord,
+    AssignmentCandidate,
+    AssignmentSelection,
+    AssignmentSource,
+    LifecycleAuditEntry,
+    LifecycleStatus,
+    ManualOverride,
+    ModelLifecycleState,
+    ReconciliationState,
+    model_id_for_actuator,
+)
+from app.actuators.store import ActuatorStore
+from app.config import Settings
+from app.ha.discovery import DiscoveredEntity, EntityRole
+from app.ha.history import EntityHistorySeries, NumericHistoryPoint
+from app.ha.models import HaEntitySummary
+from app.ha.reader import HaReader
+from app.ml.feature_store import FeatureVector
+from app.ml.registry.model_registry import ModelRegistry
+from app.ml.retraining import retrain_model
+from app.ml.training import TrainedArtifact
+
+logger = logging.getLogger(__name__)
+
+_TOKEN_PATTERN = re.compile(r"[a-z0-9]+", re.IGNORECASE)
+_STOPWORDS = frozenset(
+    {
+        "actuator",
+        "battery",
+        "bin",
+        "binary",
+        "brightness",
+        "current",
+        "door",
+        "energy",
+        "entity",
+        "humidity",
+        "illuminance",
+        "light",
+        "power",
+        "sensor",
+        "state",
+        "switch",
+        "temperature",
+        "value",
+    }
+)
+_NUMERIC_AUTO_ACCEPT_SCORE = 0.82
+_NUMERIC_MIN_MARGIN = 0.18
+_CONTEXT_AUTO_ACCEPT_SCORE = 0.78
+_MAX_CONTEXT_SELECTIONS = 3
+_AUDIT_LIMIT = 20
+
+
+class ActuatorReconciliationService:
+    def __init__(
+        self,
+        *,
+        ha_reader: HaReader,
+        store: ActuatorStore,
+        registry: ModelRegistry,
+        settings: Settings,
+    ) -> None:
+        self._ha_reader = ha_reader
+        self._store = store
+        self._registry = registry
+        self._settings = settings
+
+    def list_configured(self) -> list[ActuatorRecord]:
+        return self._store.list()
+
+    def configure_actuator(self, actuator_entity_id: str, *, enabled: bool = True) -> ActuatorRecord:
+        self._store.configure(actuator_entity_id, enabled=enabled)
+        return self.reconcile_actuator(actuator_entity_id, trigger="configuration")
+
+    def get_actuator(self, actuator_entity_id: str) -> ActuatorRecord:
+        return self._store.get(actuator_entity_id)
+
+    def set_override(
+        self,
+        actuator_entity_id: str,
+        override: ManualOverride | None,
+    ) -> ActuatorRecord:
+        record = self._store.get(actuator_entity_id)
+        updated = record.model_copy(
+            update={
+                "manual_override": override,
+                "updated_at": datetime.now(timezone.utc),
+            }
+        )
+        self._store.upsert(updated)
+        return self.reconcile_actuator(actuator_entity_id, trigger="override")
+
+    def delete_actuator(self, actuator_entity_id: str) -> None:
+        model_id = model_id_for_actuator(actuator_entity_id)
+        self._registry.archive(model_id)
+        self._store.delete(actuator_entity_id)
+
+    def reconcile_all(self, trigger: str = "manual") -> ReconciliationState:
+        state = self._store.load_reconciliation_state().model_copy(
+            update={
+                "running": True,
+                "last_started_at": datetime.now(timezone.utc),
+                "last_trigger": trigger,
+            }
+        )
+        self._store.save_reconciliation_state(state)
+        records = self._store.list()
+        for record in records:
+            self.reconcile_actuator(record.actuator_entity_id, trigger=trigger)
+        self._archive_orphan_models({model_id_for_actuator(record.actuator_entity_id) for record in records})
+        refreshed = self._store.list()
+        summary = ReconciliationState(
+            last_started_at=state.last_started_at,
+            last_completed_at=datetime.now(timezone.utc),
+            last_trigger=trigger,
+            running=False,
+            configured_actuators=len(refreshed),
+            review_required=sum(1 for record in refreshed if record.assignment.review_required),
+            trained_models=sum(
+                1 for record in refreshed if record.lifecycle.status is LifecycleStatus.TRAINED
+            ),
+            last_summary=(
+                f"{len(refreshed)} Aktuatoren geprüft, "
+                f"{sum(1 for record in refreshed if record.assignment.review_required)} "
+                "mit Prüfbedarf."
+            ),
+        )
+        self._store.save_reconciliation_state(summary)
+        return summary
+
+    def reconcile_actuator(self, actuator_entity_id: str, trigger: str = "manual") -> ActuatorRecord:
+        now = datetime.now(timezone.utc)
+        record = self._store.get(actuator_entity_id)
+        entities = {entity.entity_id: entity for entity in self._ha_reader.read_entities()}
+        discovered = {entity.entity_id: entity for entity in self._ha_reader.discover()}
+        actuator = entities.get(actuator_entity_id)
+        descriptor = discovered.get(actuator_entity_id)
+        lifecycle = record.lifecycle.model_copy(update={"last_reconciled_at": now})
+
+        if not record.enabled:
+            lifecycle = self._archive_state(
+                lifecycle,
+                "Aktuator ist deaktiviert; Modell bleibt archiviert.",
+                now=now,
+            )
+            updated = record.model_copy(
+                update={
+                    "assignment": AssignmentSelection(
+                        selected_numeric_entity_id=None,
+                        selected_context_entity_ids=[],
+                        source=AssignmentSource.NONE,
+                        confidence=0.0,
+                        review_required=False,
+                        reason="Aktuator ist deaktiviert.",
+                    ),
+                    "numeric_candidates": [],
+                    "context_candidates": [],
+                    "lifecycle": lifecycle,
+                    "updated_at": now,
+                }
+            )
+            return self._store.upsert(updated)
+
+        if actuator is None or descriptor is None or descriptor.role is not EntityRole.ACTUATOR:
+            lifecycle = self._archive_state(
+                lifecycle,
+                "Aktuator ist in Home Assistant nicht mehr als Aktor vorhanden.",
+                now=now,
+                status=LifecycleStatus.ORPHANED,
+            )
+            updated = record.model_copy(
+                update={
+                    "assignment": AssignmentSelection(
+                        selected_numeric_entity_id=None,
+                        selected_context_entity_ids=[],
+                        source=AssignmentSource.NONE,
+                        confidence=0.0,
+                        review_required=True,
+                        reason="Aktuator fehlt oder ist kein unterstützter Aktor mehr.",
+                    ),
+                    "numeric_candidates": [],
+                    "context_candidates": [],
+                    "lifecycle": lifecycle,
+                    "updated_at": now,
+                }
+            )
+            return self._store.upsert(updated)
+
+        numeric_candidates = self._rank_candidates(
+            actuator=actuator,
+            candidates=_filter_candidates(entities, discovered, {EntityRole.MEASUREMENT}),
+            context=False,
+        )
+        context_candidates = self._rank_candidates(
+            actuator=actuator,
+            candidates=_filter_candidates(
+                entities,
+                discovered,
+                {EntityRole.BINARY_CONTEXT, EntityRole.CONTEXT},
+            ),
+            context=True,
+        )
+        assignment = self._select_assignment(
+            actuator=actuator,
+            numeric_candidates=numeric_candidates,
+            context_candidates=context_candidates,
+            override=record.manual_override,
+        )
+        lifecycle = self._reconcile_lifecycle(
+            actuator=actuator,
+            assignment=assignment,
+            lifecycle=lifecycle,
+            now=now,
+        )
+        updated = record.model_copy(
+            update={
+                "assignment": assignment,
+                "numeric_candidates": numeric_candidates,
+                "context_candidates": context_candidates,
+                "lifecycle": lifecycle,
+                "updated_at": now,
+            }
+        )
+        self._store.upsert(updated)
+        logger.info(
+            "Actuator %s reconciled via %s -> %s",
+            actuator_entity_id,
+            trigger,
+            lifecycle.status,
+        )
+        return updated
+
+    def _select_assignment(
+        self,
+        *,
+        actuator: HaEntitySummary,
+        numeric_candidates: list[AssignmentCandidate],
+        context_candidates: list[AssignmentCandidate],
+        override: ManualOverride | None,
+    ) -> AssignmentSelection:
+        if override is not None:
+            selected_numeric = override.numeric_entity_id
+            selected_contexts = list(dict.fromkeys(override.context_entity_ids))
+            return AssignmentSelection(
+                selected_numeric_entity_id=selected_numeric,
+                selected_context_entity_ids=selected_contexts,
+                source=AssignmentSource.MANUAL,
+                confidence=1.0 if selected_numeric else 0.6,
+                review_required=False,
+                reason=(
+                    "Manuelle Zuordnung überschreibt die automatische Heuristik dauerhaft."
+                ),
+            )
+
+        top_numeric = numeric_candidates[0] if numeric_candidates else None
+        top_contexts = [
+            candidate.entity_id
+            for candidate in context_candidates
+            if candidate.auto_accepted
+        ][: _MAX_CONTEXT_SELECTIONS]
+        if top_numeric is None:
+            return AssignmentSelection(
+                selected_numeric_entity_id=None,
+                selected_context_entity_ids=top_contexts,
+                source=AssignmentSource.NONE,
+                confidence=0.0,
+                review_required=True,
+                reason=f"Kein numerischer Sensor konnte für {display_name(actuator)} bestimmt werden.",
+            )
+
+        return AssignmentSelection(
+            selected_numeric_entity_id=top_numeric.entity_id,
+            selected_context_entity_ids=top_contexts,
+            source=AssignmentSource.AUTOMATIC,
+            confidence=top_numeric.confidence,
+            review_required=not top_numeric.auto_accepted,
+            reason=(
+                "Automatisch akzeptiert."
+                if top_numeric.auto_accepted
+                else "Top-Kandidat gefunden, aber Zuordnung ist noch nicht eindeutig genug."
+            ),
+        )
+
+    def _reconcile_lifecycle(
+        self,
+        *,
+        actuator: HaEntitySummary,
+        assignment: AssignmentSelection,
+        lifecycle: ModelLifecycleState,
+        now: datetime,
+    ) -> ModelLifecycleState:
+        model_id = lifecycle.model_id
+        if assignment.selected_numeric_entity_id is None:
+            return self._archive_state(
+                lifecycle,
+                "Ohne numerische Sensorzuordnung wird kein Modell aktiv gehalten.",
+                now=now,
+            )
+        if assignment.review_required and assignment.source is not AssignmentSource.MANUAL:
+            return self._archive_state(
+                lifecycle,
+                "Zuordnung ist nicht eindeutig; Modell wartet auf Review.",
+                now=now,
+                status=LifecycleStatus.REVIEW_REQUIRED,
+            )
+
+        sensor_id = assignment.selected_numeric_entity_id
+        series = self._read_history(sensor_id, now)
+        points = series.points if series is not None else []
+        if len(points) < self._settings.min_training_points:
+            return self._with_audit(
+                lifecycle.model_copy(
+                    update={
+                        "status": LifecycleStatus.PENDING_HISTORY,
+                        "last_reconciled_at": now,
+                        "reason": (
+                            f"{len(points)} von mindestens {self._settings.min_training_points} "
+                            f"Messpunkten für {sensor_id} vorhanden."
+                        ),
+                        "next_action": "Mehr Historie sammeln und Reconciliation erneut ausführen.",
+                        "last_history_point_count": len(points),
+                    }
+                ),
+                action="history_wait",
+                reason=(
+                    f"Training für {display_name(actuator)} verschoben: zu wenig numerische Historie."
+                ),
+                now=now,
+            )
+
+        signature = _history_signature(sensor_id, points)
+        artifact = self._registry.get_optional(model_id)
+        needs_retrain = artifact is None
+        retrain_reason = "Noch kein Modell vorhanden."
+        if artifact is not None:
+            valid, reason = _artifact_valid_for_sensor(artifact, sensor_id)
+            if not valid:
+                self._registry.archive(model_id)
+                needs_retrain = True
+                retrain_reason = reason
+            elif lifecycle.last_history_signature != signature:
+                needs_retrain = True
+                retrain_reason = "Historie hat sich seit dem letzten Training materiell geändert."
+            elif lifecycle.last_trained_at is None or (
+                now - lifecycle.last_trained_at
+            ) >= timedelta(hours=self._settings.retrain_stale_hours):
+                needs_retrain = True
+                retrain_reason = "Modell gilt als veraltet und wird präventiv neu trainiert."
+
+        if needs_retrain:
+            vectors = [FeatureVector(sensor_id=sensor_id, values={"value": point.value}) for point in points]
+            result = retrain_model(self._registry, model_id, vectors)
+            return self._with_audit(
+                lifecycle.model_copy(
+                    update={
+                        "status": LifecycleStatus.TRAINED,
+                        "last_reconciled_at": now,
+                        "last_trained_at": now,
+                        "last_history_signature": signature,
+                        "last_history_point_count": len(points),
+                        "reason": retrain_reason,
+                        "next_action": "Automatisch überwachen und bei neuen Daten neu trainieren.",
+                    }
+                ),
+                action="retrained" if result.replaced else "trained",
+                reason=f"{retrain_reason} Modell {model_id} aktualisiert.",
+                now=now,
+            )
+
+        return self._with_audit(
+            lifecycle.model_copy(
+                update={
+                    "status": LifecycleStatus.TRAINED,
+                    "last_reconciled_at": now,
+                    "last_history_signature": signature,
+                    "last_history_point_count": len(points),
+                    "reason": "Modell ist aktuell und passt zur bestätigten Sensorzuordnung.",
+                    "next_action": "Auf neue Historie oder Staleness warten.",
+                }
+            ),
+            action="kept",
+            reason=f"Modell {model_id} blieb unverändert.",
+            now=now,
+        )
+
+    def _read_history(self, sensor_id: str, now: datetime) -> EntityHistorySeries | None:
+        start = now - timedelta(days=self._settings.history_days)
+        history = list(self._ha_reader.read_history([sensor_id], start, now))
+        for series in history:
+            if series.entity_id == sensor_id:
+                return series
+        return None
+
+    def _archive_orphan_models(self, configured_model_ids: set[str]) -> None:
+        for artifact in self._registry.list_models():
+            if not artifact.artifact_id.startswith("actuator."):
+                continue
+            if artifact.artifact_id not in configured_model_ids:
+                self._registry.archive(artifact.artifact_id)
+
+    def _archive_state(
+        self,
+        lifecycle: ModelLifecycleState,
+        reason: str,
+        *,
+        now: datetime,
+        status: LifecycleStatus = LifecycleStatus.ARCHIVED,
+    ) -> ModelLifecycleState:
+        self._registry.archive(lifecycle.model_id)
+        return self._with_audit(
+            lifecycle.model_copy(
+                update={
+                    "status": status,
+                    "last_reconciled_at": now,
+                    "reason": reason,
+                    "next_action": "Review oder neue Zuordnung erforderlich.",
+                }
+            ),
+            action="archived",
+            reason=reason,
+            now=now,
+        )
+
+    def _rank_candidates(
+        self,
+        *,
+        actuator: HaEntitySummary,
+        candidates: Iterable[tuple[HaEntitySummary, DiscoveredEntity]],
+        context: bool,
+    ) -> list[AssignmentCandidate]:
+        scored: list[AssignmentCandidate] = []
+        all_scores: list[float] = []
+        for entity, discovered in candidates:
+            score, evidence = _score_candidate(actuator, entity, discovered.role, context=context)
+            if score <= 0:
+                continue
+            all_scores.append(score)
+            scored.append(
+                AssignmentCandidate(
+                    entity_id=entity.entity_id,
+                    domain=entity.domain,
+                    role=discovered.role,
+                    device_class=entity.device_class,
+                    state_class=entity.state_class,
+                    unit_of_measurement=entity.unit_of_measurement,
+                    friendly_name=entity.friendly_name,
+                    area_name=entity.area_name,
+                    device_name=entity.device_name,
+                    score=score,
+                    confidence=0.0,
+                    evidence=evidence,
+                )
+            )
+        if not scored:
+            return []
+        highest = max(all_scores)
+        sorted_candidates = sorted(scored, key=lambda item: (-item.score, item.entity_id))
+        second_score = sorted_candidates[1].score if len(sorted_candidates) > 1 else 0.0
+        for index, candidate in enumerate(sorted_candidates):
+            confidence = candidate.score / highest if highest else 0.0
+            margin = candidate.score - second_score if index == 0 else 0.0
+            auto_score = _CONTEXT_AUTO_ACCEPT_SCORE if context else _NUMERIC_AUTO_ACCEPT_SCORE
+            auto_accepted = confidence >= auto_score and (
+                context or margin >= _NUMERIC_MIN_MARGIN
+            )
+            sorted_candidates[index] = candidate.model_copy(
+                update={
+                    "confidence": round(confidence, 4),
+                    "auto_accepted": auto_accepted,
+                }
+            )
+        return sorted_candidates
+
+    @staticmethod
+    def _with_audit(
+        lifecycle: ModelLifecycleState,
+        *,
+        action: str,
+        reason: str,
+        now: datetime,
+    ) -> ModelLifecycleState:
+        audit = list(lifecycle.audit)
+        entry = LifecycleAuditEntry(at=now, action=action, reason=reason)
+        if not audit or audit[-1].action != action or audit[-1].reason != reason:
+            audit.append(entry)
+        if len(audit) > _AUDIT_LIMIT:
+            audit = audit[-_AUDIT_LIMIT:]
+        return lifecycle.model_copy(update={"audit": audit})
+
+
+def display_name(entity: HaEntitySummary) -> str:
+    return entity.friendly_name or entity.device_name or entity.entity_id
+
+
+def _filter_candidates(
+    entities: dict[str, HaEntitySummary],
+    discovered: dict[str, DiscoveredEntity],
+    roles: set[EntityRole],
+) -> list[tuple[HaEntitySummary, DiscoveredEntity]]:
+    result: list[tuple[HaEntitySummary, DiscoveredEntity]] = []
+    for entity_id, summary in entities.items():
+        candidate = discovered.get(entity_id)
+        if candidate is None or candidate.role not in roles:
+            continue
+        result.append((summary, candidate))
+    return result
+
+
+def _score_candidate(
+    actuator: HaEntitySummary,
+    entity: HaEntitySummary,
+    role: EntityRole,
+    *,
+    context: bool,
+) -> tuple[float, list[str]]:
+    evidence: list[str] = []
+    score = 0.0
+    actuator_tokens = _metadata_tokens(actuator)
+    entity_tokens = _metadata_tokens(entity)
+    overlap = sorted(actuator_tokens.intersection(entity_tokens))
+    if overlap:
+        score += min(0.4, 0.1 * len(overlap))
+        evidence.append(f"Gemeinsame Tokens: {', '.join(overlap[:4])}")
+    if actuator.area_name and entity.area_name and actuator.area_name == entity.area_name:
+        score += 0.35
+        evidence.append(f"Gleicher Bereich: {actuator.area_name}")
+    if actuator.device_id and entity.device_id and actuator.device_id == entity.device_id:
+        score += 0.2
+        evidence.append("Gleiche Home-Assistant-Geräte-ID")
+    if actuator.device_name and entity.device_name and actuator.device_name == entity.device_name:
+        score += 0.15
+        evidence.append(f"Gleicher Gerätename: {actuator.device_name}")
+    if actuator.friendly_name and entity.friendly_name and actuator.friendly_name == entity.friendly_name:
+        score += 0.1
+        evidence.append("Gleicher Friendly Name")
+    preferred_device_classes = _preferred_device_classes(actuator.domain, context=context)
+    if entity.device_class in preferred_device_classes:
+        score += 0.2
+        evidence.append(f"Passende device_class: {entity.device_class}")
+    if not context and entity.unit_of_measurement is not None:
+        score += 0.05
+        evidence.append(f"Numerische Einheit vorhanden: {entity.unit_of_measurement}")
+    if context and role is EntityRole.BINARY_CONTEXT:
+        score += 0.05
+        evidence.append("Binärer Kontextsensor bevorzugt für Zusatzkontext.")
+    return round(min(score, 1.0), 4), evidence
+
+
+def _preferred_device_classes(domain: str, *, context: bool) -> frozenset[str]:
+    if context:
+        return frozenset({"door", "garage_door", "motion", "occupancy", "opening", "presence"})
+    mapping = {
+        "climate": {"temperature", "humidity", "power"},
+        "cover": {"illuminance", "temperature", "wind_speed"},
+        "fan": {"temperature", "humidity", "power"},
+        "humidifier": {"humidity", "temperature", "power"},
+        "light": {"illuminance", "power", "energy"},
+        "switch": {"power", "energy", "current"},
+        "valve": {"temperature", "pressure", "humidity"},
+    }
+    return frozenset(mapping.get(domain, {"power", "energy", "temperature"}))
+
+
+def _metadata_tokens(entity: HaEntitySummary) -> set[str]:
+    raw_values = [
+        entity.entity_id,
+        entity.friendly_name,
+        entity.area_name,
+        entity.device_name,
+    ]
+    tokens: set[str] = set()
+    for value in raw_values:
+        if value is None:
+            continue
+        for token in _TOKEN_PATTERN.findall(value.lower().replace("_", " ")):
+            if len(token) < 3 or token in _STOPWORDS:
+                continue
+            tokens.add(token)
+    return tokens
+
+
+def _history_signature(sensor_id: str, points: list[NumericHistoryPoint]) -> str:
+    digest = hashlib.sha256()
+    digest.update(sensor_id.encode("utf-8"))
+    for point in points:
+        digest.update(point.timestamp.isoformat().encode("utf-8"))
+        digest.update(f"{point.value:.6f}".encode("utf-8"))
+    return digest.hexdigest()
+
+
+def _artifact_valid_for_sensor(artifact: TrainedArtifact, sensor_id: str) -> tuple[bool, str]:
+    if sensor_id not in artifact.supported_sensors:
+        return False, "Vorhandenes Modell passt nicht mehr zur aktuellen Sensorzuordnung."
+    feature_models = artifact.feature_models.get(sensor_id, {})
+    if "value" not in feature_models:
+        return False, "Vorhandenes Modell enthält kein numerisches Trainingsmerkmal 'value'."
+    return True, "Modell ist kompatibel."

app/actuators/models.py

@@ -0,0 +1,102 @@
+from __future__ import annotations
+
+from datetime import datetime, timezone
+from enum import StrEnum
+
+from pydantic import BaseModel, Field
+
+from app.ha.discovery import EntityRole
+
+
+class AssignmentSource(StrEnum):
+    NONE = "none"
+    AUTOMATIC = "automatic"
+    MANUAL = "manual"
+
+
+class LifecycleStatus(StrEnum):
+    PENDING_ASSIGNMENT = "pending_assignment"
+    REVIEW_REQUIRED = "review_required"
+    PENDING_HISTORY = "pending_history"
+    TRAINED = "trained"
+    STALE = "stale"
+    INVALID = "invalid"
+    ORPHANED = "orphaned"
+    ARCHIVED = "archived"
+
+
+class AssignmentCandidate(BaseModel):
+    entity_id: str
+    domain: str
+    role: EntityRole
+    device_class: str | None = None
+    state_class: str | None = None
+    unit_of_measurement: str | None = None
+    friendly_name: str | None = None
+    area_name: str | None = None
+    device_name: str | None = None
+    score: float = Field(ge=0.0)
+    confidence: float = Field(ge=0.0, le=1.0)
+    auto_accepted: bool = False
+    evidence: list[str] = Field(default_factory=list)
+
+
+class AssignmentSelection(BaseModel):
+    selected_numeric_entity_id: str | None = None
+    selected_context_entity_ids: list[str] = Field(default_factory=list)
+    source: AssignmentSource = AssignmentSource.NONE
+    confidence: float = Field(default=0.0, ge=0.0, le=1.0)
+    review_required: bool = True
+    reason: str = "Noch keine Zuordnung vorhanden."
+
+
+class ManualOverride(BaseModel):
+    numeric_entity_id: str | None = None
+    context_entity_ids: list[str] = Field(default_factory=list)
+    updated_at: datetime = Field(default_factory=lambda: datetime.now(timezone.utc))
+    note: str | None = None
+
+
+class LifecycleAuditEntry(BaseModel):
+    at: datetime = Field(default_factory=lambda: datetime.now(timezone.utc))
+    action: str = Field(min_length=1, max_length=120)
+    reason: str = Field(min_length=1, max_length=500)
+
+
+class ModelLifecycleState(BaseModel):
+    model_id: str
+    status: LifecycleStatus = LifecycleStatus.PENDING_ASSIGNMENT
+    last_reconciled_at: datetime | None = None
+    last_trained_at: datetime | None = None
+    last_history_signature: str | None = None
+    last_history_point_count: int = Field(default=0, ge=0)
+    reason: str = "Noch keine Trainingsdaten ausgewertet."
+    next_action: str = "Aktuator auswählen und Zuordnung prüfen."
+    audit: list[LifecycleAuditEntry] = Field(default_factory=list)
+
+
+class ActuatorRecord(BaseModel):
+    actuator_entity_id: str = Field(pattern=r"^[a-z0-9_]+\.[a-z0-9_]+$")
+    enabled: bool = True
+    created_at: datetime = Field(default_factory=lambda: datetime.now(timezone.utc))
+    updated_at: datetime = Field(default_factory=lambda: datetime.now(timezone.utc))
+    assignment: AssignmentSelection = Field(default_factory=AssignmentSelection)
+    manual_override: ManualOverride | None = None
+    numeric_candidates: list[AssignmentCandidate] = Field(default_factory=list)
+    context_candidates: list[AssignmentCandidate] = Field(default_factory=list)
+    lifecycle: ModelLifecycleState
+
+
+class ReconciliationState(BaseModel):
+    last_started_at: datetime | None = None
+    last_completed_at: datetime | None = None
+    last_trigger: str | None = None
+    running: bool = False
+    configured_actuators: int = Field(default=0, ge=0)
+    review_required: int = Field(default=0, ge=0)
+    trained_models: int = Field(default=0, ge=0)
+    last_summary: str = "Noch keine Reconciliation ausgeführt."
+
+
+def model_id_for_actuator(actuator_entity_id: str) -> str:
+    return f"actuator.{actuator_entity_id}"

app/actuators/store.py

@@ -0,0 +1,116 @@
+from __future__ import annotations
+
+import json
+import os
+from datetime import datetime, timezone
+from pathlib import Path
+from threading import RLock
+
+from app.actuators.models import (
+    ActuatorRecord,
+    LifecycleStatus,
+    ModelLifecycleState,
+    ReconciliationState,
+    model_id_for_actuator,
+)
+
+
+class ActuatorStore:
+    def __init__(self, root: str | Path) -> None:
+        self._root = Path(root).resolve()
+        self._actuators_root = self._root / "actuators"
+        self._actuators_root.mkdir(parents=True, exist_ok=True)
+        self._lock = RLock()
+        self._reconciliation_state_path = self._root / "reconciliation_state.json"
+
+    def list(self) -> list[ActuatorRecord]:
+        with self._lock:
+            return [self._load(path) for path in sorted(self._actuators_root.glob("*.json"))]
+
+    def get(self, actuator_entity_id: str) -> ActuatorRecord:
+        with self._lock:
+            target = self._target(actuator_entity_id)
+            if not target.exists():
+                raise KeyError("Aktuator-Konfiguration nicht gefunden.")
+            return self._load(target)
+
+    def upsert(self, record: ActuatorRecord) -> ActuatorRecord:
+        with self._lock:
+            self._persist(record)
+            return record
+
+    def configure(self, actuator_entity_id: str, *, enabled: bool = True) -> ActuatorRecord:
+        with self._lock:
+            target = self._target(actuator_entity_id)
+            if target.exists():
+                record = self._load(target)
+                updated = record.model_copy(
+                    update={
+                        "enabled": enabled,
+                        "updated_at": datetime.now(timezone.utc),
+                    }
+                )
+                self._persist(updated)
+                return updated
+            record = ActuatorRecord(
+                actuator_entity_id=actuator_entity_id,
+                enabled=enabled,
+                lifecycle=ModelLifecycleState(
+                    model_id=model_id_for_actuator(actuator_entity_id),
+                    status=LifecycleStatus.PENDING_ASSIGNMENT,
+                ),
+            )
+            self._persist(record)
+            return record
+
+    def delete(self, actuator_entity_id: str) -> None:
+        with self._lock:
+            target = self._target(actuator_entity_id)
+            if target.exists():
+                target.unlink()
+
+    def load_reconciliation_state(self) -> ReconciliationState:
+        with self._lock:
+            if not self._reconciliation_state_path.exists():
+                return ReconciliationState()
+            try:
+                return ReconciliationState.model_validate_json(
+                    self._reconciliation_state_path.read_text(encoding="utf-8")
+                )
+            except ValueError as exc:
+                raise ValueError("Ungültiger Reconciliation-Status.") from exc
+
+    def save_reconciliation_state(self, state: ReconciliationState) -> ReconciliationState:
+        with self._lock:
+            self._persist_reconciliation_state(state)
+            return state
+
+    def _target(self, actuator_entity_id: str) -> Path:
+        if "." not in actuator_entity_id:
+            raise ValueError("Ungültige actuator_entity_id.")
+        safe_name = actuator_entity_id.replace(".", "__")
+        return self._actuators_root / f"{safe_name}.json"
+
+    def _persist(self, record: ActuatorRecord) -> None:
+        target = self._target(record.actuator_entity_id)
+        temporary = target.with_suffix(".json.tmp")
+        temporary.write_text(
+            json.dumps(record.model_dump(mode="json"), ensure_ascii=True, sort_keys=True) + "\n",
+            encoding="utf-8",
+        )
+        os.replace(temporary, target)
+
+    def _persist_reconciliation_state(self, state: ReconciliationState) -> None:
+        temporary = self._reconciliation_state_path.with_suffix(".json.tmp")
+        temporary.write_text(
+            json.dumps(state.model_dump(mode="json"), ensure_ascii=True, sort_keys=True) + "\n",
+            encoding="utf-8",
+        )
+        os.replace(temporary, self._reconciliation_state_path)
+
+    @staticmethod
+    def _load(path: Path) -> ActuatorRecord:
+        try:
+            return ActuatorRecord.model_validate_json(path.read_text(encoding="utf-8"))
+        except ValueError as exc:
+            raise ValueError(f"Ungültige Aktuator-Konfiguration: {path.name}") from exc

app/api/v1/actuators.py

@@ -0,0 +1,122 @@
+from __future__ import annotations
+
+from fastapi import APIRouter, Depends, HTTPException, Query, Request, status
+from pydantic import BaseModel, Field
+
+from app.actuators.lifecycle import ActuatorReconciliationService
+from app.actuators.models import ActuatorRecord, ManualOverride, ReconciliationState
+from app.actuators.store import ActuatorStore
+from app.dependencies import get_ha_reader
+from app.ha.discovery import EntityRole
+from app.ha.models import HaEntitySummary
+from app.ha.reader import HaReader
+
+router = APIRouter(prefix="/v1/actuators", tags=["actuators"])
+
+
+class ConfigureActuatorRequest(BaseModel):
+    actuator_entity_id: str = Field(pattern=r"^[a-z0-9_]+\.[a-z0-9_]+$")
+    enabled: bool = True
+
+
+class OverrideRequest(BaseModel):
+    numeric_entity_id: str | None = Field(default=None, pattern=r"^[a-z0-9_]+\.[a-z0-9_]+$")
+    context_entity_ids: list[str] = Field(default_factory=list)
+    note: str | None = Field(default=None, max_length=300)
+    clear: bool = False
+
+
+@router.get("/discovery", response_model=list[HaEntitySummary])
+def discover_actuators(ha_reader: HaReader = Depends(get_ha_reader)) -> list[HaEntitySummary]:
+    entities = {entity.entity_id: entity for entity in ha_reader.read_entities()}
+    discovered = ha_reader.discover()
+    actuator_ids = sorted(
+        entity.entity_id for entity in discovered if entity.role is EntityRole.ACTUATOR
+    )
+    return [entities[entity_id] for entity_id in actuator_ids if entity_id in entities]
+
+
+@router.get("", response_model=list[ActuatorRecord])
+def list_configured(request: Request) -> list[ActuatorRecord]:
+    return _service(request).list_configured()
+
+
+@router.post("", response_model=ActuatorRecord, status_code=201)
+def configure(payload: ConfigureActuatorRequest, request: Request) -> ActuatorRecord:
+    try:
+        return _service(request).configure_actuator(
+            payload.actuator_entity_id,
+            enabled=payload.enabled,
+        )
+    except KeyError as exc:
+        raise HTTPException(status_code=404, detail=str(exc)) from exc
+
+
+@router.get("/{actuator_entity_id}", response_model=ActuatorRecord)
+def get_actuator(actuator_entity_id: str, request: Request) -> ActuatorRecord:
+    try:
+        return _service(request).get_actuator(actuator_entity_id)
+    except KeyError as exc:
+        raise HTTPException(status_code=404, detail=str(exc)) from exc
+
+
+@router.delete("/{actuator_entity_id}", status_code=204)
+def delete_actuator(actuator_entity_id: str, request: Request) -> None:
+    _service(request).delete_actuator(actuator_entity_id)
+
+
+@router.post("/{actuator_entity_id}/override", response_model=ActuatorRecord)
+def set_override(
+    actuator_entity_id: str,
+    payload: OverrideRequest,
+    request: Request,
+) -> ActuatorRecord:
+    override = None if payload.clear else ManualOverride(
+        numeric_entity_id=payload.numeric_entity_id,
+        context_entity_ids=payload.context_entity_ids,
+        note=payload.note,
+    )
+    try:
+        return _service(request).set_override(actuator_entity_id, override)
+    except KeyError as exc:
+        raise HTTPException(status_code=404, detail=str(exc)) from exc
+
+
+@router.post("/{actuator_entity_id}/reconcile", response_model=ActuatorRecord)
+def reconcile_actuator(
+    actuator_entity_id: str,
+    request: Request,
+) -> ActuatorRecord:
+    try:
+        return _service(request).reconcile_actuator(actuator_entity_id, trigger="manual")
+    except KeyError as exc:
+        raise HTTPException(status_code=404, detail=str(exc)) from exc
+
+
+@router.get("/reconciliation/state", response_model=ReconciliationState)
+def get_reconciliation_state(request: Request) -> ReconciliationState:
+    store = getattr(request.app.state, "actuator_store", None)
+    if not isinstance(store, ActuatorStore):
+        raise HTTPException(
+            status_code=status.HTTP_503_SERVICE_UNAVAILABLE,
+            detail="Actuator Store nicht initialisiert.",
+        )
+    return store.load_reconciliation_state()
+
+
+@router.post("/reconciliation/run", response_model=ReconciliationState)
+def run_reconciliation(
+    request: Request,
+    trigger: str = Query(default="manual", pattern=r"^[a-z0-9_-]{1,32}$"),
+) -> ReconciliationState:
+    return _service(request).reconcile_all(trigger=trigger)
+
+
+def _service(request: Request) -> ActuatorReconciliationService:
+    service = getattr(request.app.state, "actuator_service", None)
+    if not isinstance(service, ActuatorReconciliationService):
+        raise HTTPException(
+            status_code=status.HTTP_503_SERVICE_UNAVAILABLE,
+            detail="Actuator-Reconciliation nicht initialisiert.",
+        )
+    return service

app/api/v1/automations.py

@@ -0,0 +1,77 @@
+from __future__ import annotations
+
+from fastapi import APIRouter, HTTPException, Request, Response, status
+
+from app.automations.models import (
+    AutomationProposal,
+    ProposalDecision,
+    ProposalStatus,
+)
+from app.automations.store import AutomationStore
+
+router = APIRouter(prefix="/v1/automations", tags=["automations"])
+
+
+@router.post("/proposals", response_model=AutomationProposal, status_code=201)
+def create_proposal(payload: AutomationProposal, request: Request) -> AutomationProposal:
+    if payload.trigger.above is None and payload.trigger.below is None:
+        raise HTTPException(status_code=422, detail="Trigger benötigt above oder below.")
+    return _store(request).create(payload.model_copy(update={"status": ProposalStatus.DRAFT}))
+
+
+@router.get("/proposals", response_model=list[AutomationProposal])
+def list_proposals(request: Request) -> list[AutomationProposal]:
+    return _store(request).list()
+
+
+@router.post("/proposals/{proposal_id}/approve", response_model=AutomationProposal)
+def approve(
+    proposal_id: str,
+    payload: ProposalDecision,
+    request: Request,
+) -> AutomationProposal:
+    return _decide(request, proposal_id, ProposalStatus.APPROVED, payload.expected_revision)
+
+
+@router.post("/proposals/{proposal_id}/reject", response_model=AutomationProposal)
+def reject(
+    proposal_id: str,
+    payload: ProposalDecision,
+    request: Request,
+) -> AutomationProposal:
+    return _decide(request, proposal_id, ProposalStatus.REJECTED, payload.expected_revision)
+
+
+@router.get("/proposals/{proposal_id}/yaml")
+def export_yaml(proposal_id: str, request: Request) -> Response:
+    try:
+        content = _store(request).export_yaml(proposal_id)
+    except KeyError as exc:
+        raise HTTPException(status_code=404, detail=str(exc)) from exc
+    except ValueError as exc:
+        raise HTTPException(status_code=409, detail=str(exc)) from exc
+    return Response(content=content, media_type="application/yaml")
+
+
+def _decide(
+    request: Request,
+    proposal_id: str,
+    decision: ProposalStatus,
+    expected_revision: int,
+) -> AutomationProposal:
+    try:
+        return _store(request).decide(proposal_id, decision, expected_revision)
+    except KeyError as exc:
+        raise HTTPException(status_code=404, detail=str(exc)) from exc
+    except ValueError as exc:
+        raise HTTPException(status_code=409, detail=str(exc)) from exc
+
+
+def _store(request: Request) -> AutomationStore:
+    store = getattr(request.app.state, "automation_store", None)
+    if not isinstance(store, AutomationStore):
+        raise HTTPException(
+            status_code=status.HTTP_503_SERVICE_UNAVAILABLE,
+            detail="Automation Store nicht initialisiert.",
+        )
+    return store

app/api/v1/entities.py

@@ -1,10 +1,13 @@
 from __future__ import annotations
 
+from datetime import datetime
 from typing import List
 
-from fastapi import APIRouter, Depends
+from fastapi import APIRouter, Depends, HTTPException, Query, status
 
 from app.dependencies import get_ha_reader
+from app.ha.discovery import DiscoveredEntity
+from app.ha.history import EntityHistorySeries
 from app.ha.models import HaEntitySummary
 from app.ha.reader import HaReader
 
@@ -19,3 +22,43 @@ router = APIRouter(prefix="/v1", tags=["entities"])
 )
 def list_entities(ha_reader: HaReader = Depends(get_ha_reader)) -> List[HaEntitySummary]:
     return list(ha_reader.read_entities())
+
+
+@router.get(
+    "/discovery",
+    summary="Home-Assistant-Entities klassifizieren",
+    description="Klassifiziert Entities nach Lernrelevanz, Kontextquelle und Aktor-Rolle.",
+    response_model=List[DiscoveredEntity],
+)
+def discovery(
+    domain: List[str] | None = Query(default=None),
+    learnable: bool | None = None,
+    ha_reader: HaReader = Depends(get_ha_reader),
+) -> List[DiscoveredEntity]:
+    return list(
+        ha_reader.discover(
+            domains=set(domain) if domain else None,
+            learnable=learnable,
+        )
+    )
+
+
+@router.get(
+    "/history",
+    summary="Numerische Home-Assistant-Historie lesen",
+    description="Lädt und normalisiert numerische Zustände ausgewählter Entities.",
+    response_model=List[EntityHistorySeries],
+)
+def history(
+    entity_id: List[str] = Query(),
+    start_time: datetime = Query(),
+    end_time: datetime = Query(),
+    ha_reader: HaReader = Depends(get_ha_reader),
+) -> List[EntityHistorySeries]:
+    try:
+        return list(ha_reader.read_history(entity_id, start_time, end_time))
+    except ValueError as exc:
+        raise HTTPException(
+            status_code=status.HTTP_422_UNPROCESSABLE_CONTENT,
+            detail=str(exc),
+        ) from exc

app/automations/__init__.py

@@ -0,0 +1,3 @@
+from app.automations.store import AutomationStore
+
+__all__ = ["AutomationStore"]

app/automations/models.py

@@ -0,0 +1,41 @@
+from __future__ import annotations
+
+from datetime import datetime, timezone
+from enum import StrEnum
+from uuid import uuid4
+
+from pydantic import BaseModel, Field
+
+
+class ProposalStatus(StrEnum):
+    DRAFT = "draft"
+    APPROVED = "approved"
+    REJECTED = "rejected"
+
+
+class NumericStateTrigger(BaseModel):
+    entity_id: str = Field(pattern=r"^sensor\.[a-z0-9_]+$")
+    above: float | None = None
+    below: float | None = None
+
+
+class ServiceAction(BaseModel):
+    service: str = Field(pattern=r"^(light|switch|climate|fan|cover)\.[a-z0-9_]+$")
+    entity_id: str = Field(pattern=r"^(light|switch|climate|fan|cover)\.[a-z0-9_]+$")
+    data: dict[str, str | int | float | bool] = Field(default_factory=dict)
+
+
+class AutomationProposal(BaseModel):
+    proposal_id: str = Field(default_factory=lambda: uuid4().hex)
+    alias: str = Field(min_length=1, max_length=120)
+    description: str = Field(min_length=1, max_length=500)
+    trigger: NumericStateTrigger
+    action: ServiceAction
+    status: ProposalStatus = ProposalStatus.DRAFT
+    created_at: datetime = Field(default_factory=lambda: datetime.now(timezone.utc))
+    updated_at: datetime = Field(default_factory=lambda: datetime.now(timezone.utc))
+    revision: int = 1
+
+
+class ProposalDecision(BaseModel):
+    expected_revision: int = Field(ge=1)

app/automations/store.py

@@ -0,0 +1,124 @@
+from __future__ import annotations
+
+import json
+import os
+from datetime import datetime, timezone
+from pathlib import Path
+from threading import RLock
+
+from app.automations.models import AutomationProposal, ProposalStatus
+
+
+class AutomationStore:
+    def __init__(self, root: str | Path) -> None:
+        self._root = Path(root).resolve()
+        self._root.mkdir(parents=True, exist_ok=True)
+        self._lock = RLock()
+
+    def create(self, proposal: AutomationProposal) -> AutomationProposal:
+        with self._lock:
+            target = self._target(proposal.proposal_id)
+            if target.exists():
+                raise ValueError("Automation-Vorschlag existiert bereits.")
+            self._persist(proposal)
+        return proposal
+
+    def list(self) -> list[AutomationProposal]:
+        with self._lock:
+            return [self._load(path) for path in sorted(self._root.glob("*.json"))]
+
+    def get(self, proposal_id: str) -> AutomationProposal:
+        with self._lock:
+            target = self._target(proposal_id)
+            if not target.exists():
+                raise KeyError("Automation-Vorschlag nicht gefunden.")
+            return self._load(target)
+
+    def decide(
+        self,
+        proposal_id: str,
+        status: ProposalStatus,
+        expected_revision: int,
+    ) -> AutomationProposal:
+        if status is ProposalStatus.DRAFT:
+            raise ValueError("Entscheidung darf nicht auf draft gesetzt werden.")
+        with self._lock:
+            proposal = self.get(proposal_id)
+            if proposal.revision != expected_revision:
+                raise ValueError("Revision stimmt nicht mit dem aktuellen Vorschlag überein.")
+            if proposal.status is not ProposalStatus.DRAFT:
+                raise ValueError("Über den Vorschlag wurde bereits entschieden.")
+            updated = proposal.model_copy(
+                update={
+                    "status": status,
+                    "updated_at": datetime.now(timezone.utc),
+                    "revision": proposal.revision + 1,
+                }
+            )
+            self._persist(updated)
+            return updated
+
+    def export_yaml(self, proposal_id: str) -> str:
+        proposal = self.get(proposal_id)
+        if proposal.status is not ProposalStatus.APPROVED:
+            raise ValueError("Nur freigegebene Vorschläge dürfen exportiert werden.")
+        trigger_lines = [
+            "trigger:",
+            "  - platform: numeric_state",
+            f"    entity_id: {proposal.trigger.entity_id}",
+        ]
+        if proposal.trigger.above is not None:
+            trigger_lines.append(f"    above: {proposal.trigger.above}")
+        if proposal.trigger.below is not None:
+            trigger_lines.append(f"    below: {proposal.trigger.below}")
+        action_lines = [
+            "action:",
+            f"  - service: {proposal.action.service}",
+            "    target:",
+            f"      entity_id: {proposal.action.entity_id}",
+        ]
+        if proposal.action.data:
+            action_lines.append("    data:")
+            action_lines.extend(
+                f"      {key}: {_yaml_scalar(value)}"
+                for key, value in sorted(proposal.action.data.items())
+            )
+        return "\n".join(
+            [
+                f"alias: {_yaml_scalar(proposal.alias)}",
+                f"description: {_yaml_scalar(proposal.description)}",
+                *trigger_lines,
+                *action_lines,
+                "mode: single",
+                "",
+            ]
+        )
+
+    def _target(self, proposal_id: str) -> Path:
+        if len(proposal_id) != 32 or not proposal_id.isalnum():
+            raise ValueError("Ungültige proposal_id.")
+        return self._root / f"{proposal_id}.json"
+
+    def _persist(self, proposal: AutomationProposal) -> None:
+        target = self._target(proposal.proposal_id)
+        temporary = target.with_suffix(".json.tmp")
+        temporary.write_text(
+            json.dumps(proposal.model_dump(mode="json"), ensure_ascii=True, sort_keys=True) + "\n",
+            encoding="utf-8",
+        )
+        os.replace(temporary, target)
+
+    @staticmethod
+    def _load(path: Path) -> AutomationProposal:
+        try:
+            return AutomationProposal.model_validate_json(path.read_text(encoding="utf-8"))
+        except ValueError as exc:
+            raise ValueError(f"Ungültiger Automation-Vorschlag: {path.name}") from exc
+
+
+def _yaml_scalar(value: str | int | float | bool) -> str:
+    if isinstance(value, bool):
+        return "true" if value else "false"
+    if isinstance(value, (int, float)):
+        return str(value)
+    return json.dumps(value, ensure_ascii=True)

app/config.py

@@ -9,6 +9,12 @@ class Settings:
     ha_url: str | None = None
     ha_token: str | None = None
     model_store: str = ".model_store"
+    automation_store: str = ".automation_store"
+    actuator_store: str = ".actuator_store"
+    history_days: int = 14
+    min_training_points: int = 24
+    retrain_stale_hours: int = 24
+    reconcile_interval_seconds: int = 900
 
     @property
     def ha_configured(self) -> bool:
@@ -20,4 +26,12 @@ def load_settings() -> Settings:
         ha_url=os.getenv("SILLYHOME_HA_URL") or os.getenv("HA_URL"),
         ha_token=os.getenv("SILLYHOME_HA_TOKEN") or os.getenv("HA_TOKEN"),
         model_store=os.getenv("SILLYHOME_MODEL_STORE", ".model_store"),
+        automation_store=os.getenv("SILLYHOME_AUTOMATION_STORE", ".automation_store"),
+        actuator_store=os.getenv("SILLYHOME_ACTUATOR_STORE", ".actuator_store"),
+        history_days=max(1, min(31, int(os.getenv("SILLYHOME_HISTORY_DAYS", "14")))),
+        min_training_points=max(2, int(os.getenv("SILLYHOME_MIN_TRAINING_POINTS", "24"))),
+        retrain_stale_hours=max(1, int(os.getenv("SILLYHOME_RETRAIN_STALE_HOURS", "24"))),
+        reconcile_interval_seconds=max(
+            60, int(os.getenv("SILLYHOME_RECONCILE_INTERVAL_SECONDS", "900"))
+        ),
     )

app/ha/client.py

@@ -2,6 +2,10 @@ from __future__ import annotations
 
 import logging
 from dataclasses import dataclass
+from datetime import datetime
+import json
+import re
+from urllib.parse import quote
 
 import requests
 
@@ -14,6 +18,9 @@ from app.ha.exceptions import (
 
 logger = logging.getLogger(__name__)
 
+_ENTITY_ID_PATTERN = re.compile(r"^[a-z0-9_]+\.[a-z0-9_]+$")
+_MAX_HISTORY_SECONDS = 31 * 24 * 60 * 60
+
 
 @dataclass(frozen=True)
 class HaClientSettings:
@@ -35,9 +42,84 @@ class HaClient:
         self._session.close()
 
     def list_entities(self) -> list[dict[str, object]]:
+        payload = self._get_json("/api/states")
+        if not isinstance(payload, list):
+            raise HaUnexpectedPayloadError(
+                "Antwort von Home Assistant hat unerwartetes Format."
+            )
+        return payload
+
+    def get_history(
+        self,
+        entity_ids: list[str],
+        start_time: datetime,
+        end_time: datetime,
+    ) -> list[object]:
+        if not entity_ids:
+            raise ValueError("Mindestens eine entity_id ist erforderlich.")
+        if len(entity_ids) > 100:
+            raise ValueError("Es können höchstens 100 Entities abgefragt werden.")
+        if any(not _ENTITY_ID_PATTERN.fullmatch(entity_id) for entity_id in entity_ids):
+            raise ValueError("entity_id enthält ein ungültiges Format.")
+        if start_time.tzinfo is None or end_time.tzinfo is None:
+            raise ValueError("start_time und end_time müssen eine Zeitzone enthalten.")
+        if end_time <= start_time:
+            raise ValueError("end_time muss nach start_time liegen.")
+        if (end_time - start_time).total_seconds() > _MAX_HISTORY_SECONDS:
+            raise ValueError("History-Abfragen sind auf 31 Tage begrenzt.")
+
+        start = quote(start_time.isoformat(), safe=":+")
+        payload = self._get_json(
+            f"/api/history/period/{start}",
+            params={
+                "filter_entity_id": ",".join(entity_ids),
+                "end_time": end_time.isoformat(),
+                "minimal_response": "1",
+                "no_attributes": "1",
+            },
+        )
+        if not isinstance(payload, list):
+            raise HaUnexpectedPayloadError(
+                "History-Antwort von Home Assistant hat unerwartetes Format."
+            )
+        return payload
+
+    def list_entity_metadata(self, entity_ids: list[str]) -> dict[str, dict[str, str | None]]:
+        if not entity_ids:
+            return {}
+        if any(not _ENTITY_ID_PATTERN.fullmatch(entity_id) for entity_id in entity_ids):
+            raise ValueError("entity_id enthält ein ungültiges Format.")
+        template = _metadata_template(entity_ids)
+        rendered = self._post_text("/api/template", {"template": template})
+        try:
+            payload = json.loads(rendered)
+        except json.JSONDecodeError as exc:
+            raise HaUnexpectedPayloadError("Entity-Metadaten konnten nicht gelesen werden.") from exc
+        if not isinstance(payload, list):
+            raise HaUnexpectedPayloadError("Entity-Metadaten haben ein unerwartetes Format.")
+        result: dict[str, dict[str, str | None]] = {}
+        for item in payload:
+            if not isinstance(item, dict):
+                raise HaUnexpectedPayloadError("Entity-Metadaten haben ein unerwartetes Format.")
+            entity_id = item.get("entity_id")
+            if not isinstance(entity_id, str) or "." not in entity_id:
+                raise HaUnexpectedPayloadError("Entity-Metadaten enthalten ungültige entity_id.")
+            result[entity_id] = {
+                key: _optional_string(item.get(key))
+                for key in ("area_id", "area_name", "device_id", "device_name")
+            }
+        return result
+
+    def _get_json(
+        self,
+        path: str,
+        *,
+        params: dict[str, str] | None = None,
+    ) -> object:
         try:
             response = self._session.get(
-                f"{self._settings.url.rstrip('/')}/api/states",
+                f"{self._settings.url.rstrip('/')}{path}",
+                params=params,
                 timeout=self._settings.timeout_seconds,
             )
         except requests.Timeout as exc:
@@ -69,9 +151,58 @@ class HaClient:
                 "Antwort von Home Assistant ist kein gültiges JSON."
             ) from exc
 
-        if not isinstance(payload, list):
-            raise HaUnexpectedPayloadError(
-                "Antwort von Home Assistant hat unerwartetes Format."
-            )
-
         return payload
+
+    def _post_text(self, path: str, payload: dict[str, str]) -> str:
+        try:
+            response = self._session.post(
+                f"{self._settings.url.rstrip('/')}{path}",
+                json=payload,
+                timeout=self._settings.timeout_seconds,
+            )
+        except requests.Timeout as exc:
+            raise HaTimeoutError("Zeitüberschreitung beim Zugriff auf Home Assistant.") from exc
+        except requests.RequestException as exc:
+            raise HaHttpError(
+                getattr(getattr(exc, "response", None), "status_code", 502),
+                "Netzwerkfehler beim Zugriff auf Home Assistant.",
+            ) from exc
+
+        if response.status_code in (401, 403):
+            raise HaAuthError(
+                response.status_code,
+                "Authentifizierung bei Home Assistant fehlgeschlagen.",
+            )
+        try:
+            response.raise_for_status()
+        except requests.HTTPError as exc:
+            raise HaHttpError(response.status_code, "Home Assistant meldet einen Fehler.") from exc
+        return response.text
+
+
+def _metadata_template(entity_ids: list[str]) -> str:
+    ids = json.dumps(entity_ids, ensure_ascii=True)
+    return (
+        "{% set ids = "
+        f"{ids}"
+        " %}["
+        "{% for entity_id in ids %}"
+        "{% set device = device_id(entity_id) %}"
+        "{{ "
+        "{"
+        "\"entity_id\": entity_id,"
+        "\"area_id\": area_id(entity_id),"
+        "\"area_name\": area_name(entity_id),"
+        "\"device_id\": device,"
+        "\"device_name\": device_attr(device, 'name') if device else none"
+        "}"
+        " | tojson }}"
+        "{% if not loop.last %},{% endif %}"
+        "{% endfor %}]"
+    )
+
+
+def _optional_string(value: object) -> str | None:
+    if value is None or value == "":
+        return None
+    return str(value)

app/ha/discovery.py

@@ -0,0 +1,184 @@
+from __future__ import annotations
+
+from enum import StrEnum
+
+from pydantic import BaseModel
+
+from app.ha.models import HaEntitySummary
+
+
+class EntityRole(StrEnum):
+    MEASUREMENT = "measurement"
+    BINARY_CONTEXT = "binary_context"
+    CONTEXT = "context"
+    ACTUATOR = "actuator"
+    UNSUPPORTED = "unsupported"
+
+
+class DiscoveredEntity(BaseModel):
+    entity_id: str
+    domain: str
+    device_class: str | None = None
+    state_class: str | None = None
+    unit_of_measurement: str | None = None
+    role: EntityRole
+    learnable: bool
+    reason: str
+
+
+_MEASUREMENT_CLASSES = frozenset({
+    "apparent_power",
+    "atmospheric_pressure",
+    "battery",
+    "carbon_dioxide",
+    "carbon_monoxide",
+    "current",
+    "distance",
+    "duration",
+    "energy",
+    "frequency",
+    "gas",
+    "humidity",
+    "illuminance",
+    "moisture",
+    "monetary",
+    "nitrogen_dioxide",
+    "nitrogen_monoxide",
+    "nitrous_oxide",
+    "ozone",
+    "pm1",
+    "pm10",
+    "pm25",
+    "power",
+    "precipitation",
+    "pressure",
+    "reactive_power",
+    "signal_strength",
+    "sound_pressure",
+    "speed",
+    "sulphur_dioxide",
+    "temperature",
+    "volatile_organic_compounds",
+    "voltage",
+    "volume",
+    "volume_flow_rate",
+    "water",
+    "weight",
+    "wind_speed",
+})
+_BINARY_CONTEXT_CLASSES = frozenset({
+    "door",
+    "garage_door",
+    "lock",
+    "motion",
+    "occupancy",
+    "opening",
+    "presence",
+    "problem",
+    "safety",
+    "smoke",
+    "sound",
+    "vibration",
+    "window",
+})
+_ACTUATOR_DOMAINS = frozenset({
+    "button",
+    "climate",
+    "cover",
+    "fan",
+    "humidifier",
+    "light",
+    "lock",
+    "scene",
+    "select",
+    "siren",
+    "switch",
+    "valve",
+})
+_CONTEXT_DOMAINS = frozenset({"device_tracker", "person", "sun", "weather", "zone"})
+_LEARNABLE_CONTEXT_DOMAINS = frozenset({"device_tracker", "person", "weather"})
+_NUMERIC_STATE_CLASSES = frozenset({"measurement", "total", "total_increasing"})
+
+
+def classify_entity(entity: HaEntitySummary) -> DiscoveredEntity:
+    if entity.domain == "sensor" and (
+        entity.state_class in _NUMERIC_STATE_CLASSES
+        or entity.device_class in _MEASUREMENT_CLASSES
+        or entity.unit_of_measurement is not None
+    ):
+        return _result(
+            entity,
+            EntityRole.MEASUREMENT,
+            learnable=True,
+            reason="Numerischer Messsensor für Zeitreihen und Training.",
+        )
+
+    if entity.domain == "binary_sensor" and entity.device_class in _BINARY_CONTEXT_CLASSES:
+        return _result(
+            entity,
+            EntityRole.BINARY_CONTEXT,
+            learnable=True,
+            reason="Binärer Kontextsensor für Zustands- und Anwesenheitsmuster.",
+        )
+
+    if entity.domain in _CONTEXT_DOMAINS:
+        learnable = entity.domain in _LEARNABLE_CONTEXT_DOMAINS
+        return _result(
+            entity,
+            EntityRole.CONTEXT,
+            learnable=learnable,
+            reason=(
+                "Kontextquelle für Training und Erklärungen."
+                if learnable
+                else "Kontextquelle ohne direkte Trainingsfreigabe."
+            ),
+        )
+
+    if entity.domain in _ACTUATOR_DOMAINS:
+        return _result(
+            entity,
+            EntityRole.ACTUATOR,
+            learnable=False,
+            reason="Aktor ist ein mögliches Automationsziel, aber kein Trainingssensor.",
+        )
+
+    return _result(
+        entity,
+        EntityRole.UNSUPPORTED,
+        learnable=False,
+        reason="Entity-Typ ist noch nicht für Lernen oder Automationen klassifiziert.",
+    )
+
+
+def discover_entities(
+    entities: list[HaEntitySummary],
+    domains: set[str] | None = None,
+    learnable: bool | None = None,
+) -> list[DiscoveredEntity]:
+    normalized_domains = {domain.strip().lower() for domain in domains or set() if domain.strip()}
+    discovered = [classify_entity(entity) for entity in entities]
+    return [
+        entity
+        for entity in discovered
+        if (not normalized_domains or entity.domain in normalized_domains)
+        and (learnable is None or entity.learnable is learnable)
+    ]
+
+
+def _result(
+    entity: HaEntitySummary,
+    role: EntityRole,
+    *,
+    learnable: bool,
+    reason: str,
+) -> DiscoveredEntity:
+    return DiscoveredEntity(
+        entity_id=entity.entity_id,
+        domain=entity.domain,
+        device_class=entity.device_class,
+        state_class=entity.state_class,
+        unit_of_measurement=entity.unit_of_measurement,
+        role=role,
+        learnable=learnable,
+        reason=reason,
+    )

app/ha/history.py

@@ -0,0 +1,91 @@
+from __future__ import annotations
+
+import math
+from datetime import datetime
+
+from pydantic import BaseModel
+
+from app.ha.exceptions import HaUnexpectedPayloadError
+
+
+class NumericHistoryPoint(BaseModel):
+    timestamp: datetime
+    value: float
+
+
+class EntityHistorySeries(BaseModel):
+    entity_id: str
+    points: list[NumericHistoryPoint]
+
+
+def normalize_history_payload(payload: object) -> list[EntityHistorySeries]:
+    if not isinstance(payload, list):
+        raise HaUnexpectedPayloadError("History-Payload muss eine Liste sein.")
+
+    normalized: list[EntityHistorySeries] = []
+    for raw_series in payload:
+        if not isinstance(raw_series, list):
+            raise HaUnexpectedPayloadError("History-Serie muss eine Liste sein.")
+        series = _normalize_series(raw_series)
+        if series is not None:
+            normalized.append(series)
+
+    return sorted(normalized, key=lambda item: item.entity_id)
+
+
+def _normalize_series(raw_series: list[object]) -> EntityHistorySeries | None:
+    entity_id: str | None = None
+    points: list[NumericHistoryPoint] = []
+
+    for raw_entry in raw_series:
+        if not isinstance(raw_entry, dict):
+            raise HaUnexpectedPayloadError("History-Eintrag muss ein Objekt sein.")
+
+        raw_entity_id = raw_entry.get("entity_id")
+        if raw_entity_id is not None:
+            if not isinstance(raw_entity_id, str) or "." not in raw_entity_id:
+                raise HaUnexpectedPayloadError("History-Eintrag enthält ungültige entity_id.")
+            if entity_id is not None and entity_id != raw_entity_id:
+                raise HaUnexpectedPayloadError("History-Serie enthält mehrere Entities.")
+            entity_id = raw_entity_id
+
+        raw_state = raw_entry.get("state")
+        value = _finite_float(raw_state)
+        if value is None:
+            continue
+        if entity_id is None:
+            raise HaUnexpectedPayloadError("History-Serie enthält keine entity_id.")
+
+        raw_timestamp = raw_entry.get("last_changed") or raw_entry.get("last_updated")
+        timestamp = _parse_timestamp(raw_timestamp)
+        points.append(NumericHistoryPoint(timestamp=timestamp, value=value))
+
+    if entity_id is None or not points:
+        return None
+
+    points.sort(key=lambda point: point.timestamp)
+    return EntityHistorySeries(entity_id=entity_id, points=points)
+
+
+def _finite_float(value: object) -> float | None:
+    if isinstance(value, bool) or value is None:
+        return None
+    if not isinstance(value, (str, int, float)):
+        return None
+    try:
+        converted = float(value)
+    except (TypeError, ValueError):
+        return None
+    return converted if math.isfinite(converted) else None
+
+
+def _parse_timestamp(value: object) -> datetime:
+    if not isinstance(value, str):
+        raise HaUnexpectedPayloadError("Numerischer History-Eintrag enthält keinen Zeitstempel.")
+    try:
+        parsed = datetime.fromisoformat(value.replace("Z", "+00:00"))
+    except ValueError as exc:
+        raise HaUnexpectedPayloadError("History-Eintrag enthält ungültigen Zeitstempel.") from exc
+    if parsed.tzinfo is None:
+        raise HaUnexpectedPayloadError("History-Zeitstempel muss eine Zeitzone enthalten.")
+    return parsed

app/ha/models.py

@@ -17,3 +17,8 @@ class HaEntitySummary(BaseModel):
     state_class: str | None = None
     device_class: str | None = None
     unit_of_measurement: str | None = None
+    friendly_name: str | None = None
+    area_id: str | None = None
+    area_name: str | None = None
+    device_id: str | None = None
+    device_name: str | None = None

app/ha/reader.py

@@ -1,11 +1,19 @@
 from __future__ import annotations
 
 from collections.abc import Sequence
+from datetime import datetime
 from typing import Any
+import logging
+
+from app.ha.exceptions import HaClientError
 
 from app.ha.client import HaClient
+from app.ha.discovery import DiscoveredEntity, discover_entities
+from app.ha.history import EntityHistorySeries, normalize_history_payload
 from app.ha.models import HaEntitySummary
 
+logger = logging.getLogger(__name__)
+
 
 class HaReader:
     def __init__(self, client: HaClient) -> None:
@@ -13,6 +21,16 @@ class HaReader:
 
     def read_entities(self) -> Sequence[HaEntitySummary]:
         entities = self._client.list_entities()
+        entity_ids = [
+            raw_entity_id
+            for item in entities
+            if isinstance((raw_entity_id := item.get("entity_id")), str) and "." in raw_entity_id
+        ]
+        try:
+            metadata_by_entity = self._client.list_entity_metadata(entity_ids)
+        except (HaClientError, ValueError) as exc:
+            logger.warning("HA metadata enrichment skipped: %s", exc)
+            metadata_by_entity = {}
         summaries: list[HaEntitySummary] = []
         for item in entities:
             raw_entity_id = item.get("entity_id")
@@ -22,6 +40,7 @@ class HaReader:
             domain = entity_id.split(".", 1)[0]
             raw_attributes = item.get("attributes") or {}
             attributes: dict[str, Any] = raw_attributes if isinstance(raw_attributes, dict) else {}
+            metadata = metadata_by_entity.get(entity_id, {})
             summaries.append(
                 HaEntitySummary(
                     entity_id=entity_id,
@@ -29,10 +48,35 @@ class HaReader:
                     state_class=_optional_str(attributes.get("state_class")),
                     device_class=_optional_str(attributes.get("device_class")),
                     unit_of_measurement=_optional_str(attributes.get("unit_of_measurement")),
+                    friendly_name=_optional_str(attributes.get("friendly_name")),
+                    area_id=_optional_str(metadata.get("area_id") or attributes.get("area_id")),
+                    area_name=_optional_str(metadata.get("area_name") or attributes.get("area_name")),
+                    device_id=_optional_str(metadata.get("device_id") or attributes.get("device_id")),
+                    device_name=_optional_str(
+                        metadata.get("device_name")
+                        or attributes.get("device_name")
+                        or attributes.get("device")
+                    ),
                 )
             )
         return summaries
 
+    def discover(
+        self,
+        domains: set[str] | None = None,
+        learnable: bool | None = None,
+    ) -> Sequence[DiscoveredEntity]:
+        return discover_entities(list(self.read_entities()), domains=domains, learnable=learnable)
+
+    def read_history(
+        self,
+        entity_ids: list[str],
+        start_time: datetime,
+        end_time: datetime,
+    ) -> Sequence[EntityHistorySeries]:
+        payload = self._client.get_history(entity_ids, start_time, end_time)
+        return normalize_history_payload(payload)
+
 
 def _optional_str(value: object) -> str | None:
     if value is None or value == "":

app/main.py

@@ -1,10 +1,19 @@
-from contextlib import asynccontextmanager
+import asyncio
+from contextlib import asynccontextmanager, suppress
 from collections.abc import AsyncIterator
+from pathlib import Path
 from typing import cast
 
 from fastapi import FastAPI
+from fastapi.responses import FileResponse
+from fastapi.staticfiles import StaticFiles
 
+from app.actuators.lifecycle import ActuatorReconciliationService
+from app.actuators.store import ActuatorStore
+from app.api.v1.actuators import router as actuators_router
 from app.api.v1.entities import router as entities_router
+from app.api.v1.automations import router as automations_router
+from app.automations.store import AutomationStore
 from app.config import load_settings
 from app.core.exception_handlers import register_exception_handlers
 from app.ha.client import HaClient, HaClientSettings
@@ -17,9 +26,14 @@ from backend.routes.ml import init_ml_routes
 async def lifespan(app: FastAPI) -> AsyncIterator[None]:
     settings = app.state.settings
     client: HaClient | None = None
+    reconcile_task: asyncio.Task[None] | None = None
     app.state.registry = ModelRegistry(settings.model_store)
+    app.state.automation_store = AutomationStore(settings.automation_store)
+    app.state.actuator_store = ActuatorStore(settings.actuator_store)
     if hasattr(app.state, "ha_reader"):
         del app.state.ha_reader
+    if hasattr(app.state, "actuator_service"):
+        del app.state.actuator_service
     if settings.ha_configured:
         client = HaClient(
             settings=HaClientSettings(
@@ -28,9 +42,21 @@ async def lifespan(app: FastAPI) -> AsyncIterator[None]:
             )
         )
         app.state.ha_reader = HaReader(client=client)
+        app.state.actuator_service = ActuatorReconciliationService(
+            ha_reader=app.state.ha_reader,
+            store=app.state.actuator_store,
+            registry=app.state.registry,
+            settings=settings,
+        )
+        await asyncio.to_thread(app.state.actuator_service.reconcile_all, "startup")
+        reconcile_task = asyncio.create_task(_periodic_reconciliation(app))
     try:
         yield
     finally:
+        if reconcile_task is not None:
+            reconcile_task.cancel()
+            with suppress(asyncio.CancelledError):
+                await reconcile_task
         if client is not None:
             client.close()
 
@@ -38,14 +64,19 @@ async def lifespan(app: FastAPI) -> AsyncIterator[None]:
 app = FastAPI(
     title="SillyHome Next API",
     description="Lokales Smart-Home-Intelligenzsystem für Home Assistant.",
-    version="0.1.0",
+    version="0.4.0",
     lifespan=lifespan,
 )
 app.state.settings = load_settings()
 register_exception_handlers(app)
 app.include_router(entities_router)
+app.include_router(automations_router)
+app.include_router(actuators_router)
 init_ml_routes(app, model_store=app.state.settings.model_store)
 
+STATIC_DIR = Path(__file__).with_name("static")
+app.mount("/static", StaticFiles(directory=STATIC_DIR), name="static")
+
 
 @app.get("/health")
 def health() -> dict[str, str]:
@@ -53,5 +84,14 @@ def health() -> dict[str, str]:
 
 
 @app.get("/")
-def root() -> dict[str, str]:
+def root() -> FileResponse:
-    return {"service": "sillyhome-next", "docs": "/docs"}
+    return FileResponse(STATIC_DIR / "index.html")
+
+
+async def _periodic_reconciliation(app: FastAPI) -> None:
+    while True:
+        await asyncio.sleep(app.state.settings.reconcile_interval_seconds)
+        service = getattr(app.state, "actuator_service", None)
+        if not isinstance(service, ActuatorReconciliationService):
+            continue
+        await asyncio.to_thread(service.reconcile_all, "scheduled")

app/ml/__init__.py

@@ -3,6 +3,10 @@
 __all__ = [
     "FeatureStore",
     "FeatureVector",
+    "FeatureModel",
+    "FeatureExplanation",
+    "PredictionResult",
+    "Predictor",
     "RetrainingResult",
     "RetrainingService",
     "TrainedArtifact",
@@ -10,5 +14,7 @@ __all__ = [
     "retrain_model",
 ]
 from app.ml.feature_store import FeatureStore, FeatureVector
+from app.ml.explanation import FeatureExplanation
+from app.ml.predictor import PredictionResult, Predictor
 from app.ml.retraining import RetrainingResult, RetrainingService, retrain_model
-from app.ml.training import TrainedArtifact, TrainingPipeline
+from app.ml.training import FeatureModel, TrainedArtifact, TrainingPipeline

app/ml/evaluation.py

@@ -1,9 +1,13 @@
 from __future__ import annotations
 
 import logging
+import math
 from collections.abc import Sequence
 from dataclasses import dataclass
 
+from app.ml.feature_store import FeatureVector
+from app.ml.predictor import Predictor
+from app.ml.registry.model_registry import ModelRegistry
 from app.ml.training import TrainingPipeline
 
 logger = logging.getLogger(__name__)
@@ -24,41 +28,62 @@ class EvalReport:
 
 
 class Evaluator:
-    def __init__(self, pipeline: TrainingPipeline) -> None:
+    def __init__(
+        self,
+        pipeline: TrainingPipeline | None = None,
+        registry: ModelRegistry | None = None,
+    ) -> None:
+        if isinstance(pipeline, ModelRegistry) and registry is None:
+            registry = pipeline
+            pipeline = None
+        if pipeline is None and registry is None:
+            raise ValueError("Evaluator erfordert TrainingPipeline oder ModelRegistry.")
         self._pipeline = pipeline
+        self._registry = registry
+        self._predictor = Predictor(pipeline=pipeline, registry=registry)
 
-    def evaluate(self, artifact_id: str, predictions: Sequence[str]) -> EvalReport:
+    def evaluate(self, artifact_id: str, samples: Sequence[FeatureVector]) -> EvalReport:
         try:
-            supported_sensors = set(self._pipeline.export(artifact_id).supported_sensors)
+            if self._registry is not None:
+                self._registry.load_artifact(artifact_id)
+            elif self._pipeline is not None:
+                self._pipeline.export(artifact_id)
         except KeyError as exc:
             raise ValueError("Kein trainiertes Modell für Evaluation vorhanden.") from exc
 
-        parsed_sensors = [_prediction_sensor(prediction) for prediction in predictions]
+        absolute_errors: list[float] = []
-        supported_hits = sum(sensor in supported_sensors for sensor in parsed_sensors)
+        squared_errors: list[float] = []
-        unknown_hits = sum(sensor not in supported_sensors for sensor in parsed_sensors)
+        for sample in samples:
-        sample_size = len(predictions)
+            try:
-        coverage = supported_hits / sample_size if sample_size else 0.0
+                prediction = self._predictor.predict(artifact_id, sample)
-        unknown_rate = unknown_hits / sample_size if sample_size else 0.0
+            except ValueError:
+                continue
+            for feature_name, predicted in prediction.predictions.items():
+                actual = float(sample.values[feature_name])
+                error = predicted - actual
+                absolute_errors.append(abs(error))
+                squared_errors.append(error**2)
 
-        coverage_metric = Metric(name="coverage", value=coverage, threshold=0.8)
+        sample_size = len(absolute_errors)
-        unknown_metric = Metric(name="unknown_rate", value=unknown_rate, threshold=0.1)
+        mae = sum(absolute_errors) / sample_size if sample_size else 0.0
+        rmse = math.sqrt(sum(squared_errors) / sample_size) if sample_size else 0.0
+        expected_values = sum(len(sample.values) for sample in samples)
+        coverage = sample_size / expected_values if expected_values else 0.0
 
         report = EvalReport(
             artifact_id=artifact_id,
             sample_size=sample_size,
-            metrics=[coverage_metric, unknown_metric],
+            metrics=[
+                Metric(name="mae", value=mae),
+                Metric(name="rmse", value=rmse),
+                Metric(name="coverage", value=coverage, threshold=0.8),
+            ],
         )
         logger.info(
-            "Evaluation %s -> coverage=%.2f, unknown_rate=%.2f",
+            "Evaluation %s -> mae=%.4f, rmse=%.4f, coverage=%.2f",
             artifact_id,
+            mae,
+            rmse,
             coverage,
-            unknown_rate,
         )
         return report
-
-
-def _prediction_sensor(prediction: str) -> str | None:
-    parts = prediction.split(":", 2)
-    if len(parts) != 3 or not parts[0] or not parts[1]:
-        return None
-    return parts[1]

app/ml/explanation.py

@@ -0,0 +1,57 @@
+from __future__ import annotations
+
+from dataclasses import dataclass
+
+from app.ml.training import FeatureModel
+
+
+@dataclass(frozen=True)
+class FeatureExplanation:
+    feature: str
+    current_value: float
+    predicted_value: float
+    change: float
+    direction: str
+    sample_count: int
+    historical_mean: float
+    historical_range: tuple[float, float]
+    standard_deviation: float
+    trend_per_step: float
+    confidence: float
+    summary: str
+
+
+def explain_feature(
+    feature_name: str,
+    current_value: float,
+    predicted_value: float,
+    model: FeatureModel,
+) -> FeatureExplanation:
+    change = predicted_value - current_value
+    direction = _direction(change)
+    summary = (
+        f"{feature_name}: {direction}; Prognose {predicted_value:.3f} "
+        f"aus aktuellem Wert {current_value:.3f} und Trend {model.slope:+.3f}. "
+        f"Basis: {model.sample_count} Messwerte, Mittelwert {model.mean:.3f}, "
+        f"Confidence {model.confidence:.0%}."
+    )
+    return FeatureExplanation(
+        feature=feature_name,
+        current_value=current_value,
+        predicted_value=predicted_value,
+        change=change,
+        direction=direction,
+        sample_count=model.sample_count,
+        historical_mean=model.mean,
+        historical_range=(model.minimum, model.maximum),
+        standard_deviation=model.standard_deviation,
+        trend_per_step=model.slope,
+        confidence=model.confidence,
+        summary=summary,
+    )
+
+
+def _direction(change: float) -> str:
+    if abs(change) < 1e-12:
+        return "stabil"
+    return "steigend" if change > 0 else "fallend"

app/ml/predictor.py

@@ -1,8 +1,11 @@
 from __future__ import annotations
 
 import logging
+import math
+from dataclasses import dataclass
 from typing import Sequence
 
+from app.ml.explanation import FeatureExplanation, explain_feature
 from app.ml.feature_store import FeatureVector
 from app.ml.registry.model_registry import ModelRegistry
 from app.ml.training import TrainedArtifact, TrainingPipeline
@@ -10,6 +13,16 @@ from app.ml.training import TrainedArtifact, TrainingPipeline
 logger = logging.getLogger(__name__)
 
 
+@dataclass(frozen=True)
+class PredictionResult:
+    artifact_id: str
+    sensor_id: str
+    predictions: dict[str, float]
+    confidence: float
+    model_type: str
+    explanations: dict[str, FeatureExplanation]
+
+
 class Predictor:
     def __init__(
         self,
@@ -24,15 +37,54 @@ class Predictor:
         self._pipeline = pipeline
         self._registry = registry
 
-    def predict(self, artifact_id: str, entity: FeatureVector) -> str:
+    def predict(self, artifact_id: str, entity: FeatureVector) -> PredictionResult:
         artifact = self._get_artifact(artifact_id)
         if entity.sensor_id not in artifact.supported_sensors:
             raise ValueError(
                 f"Sensor '{entity.sensor_id}' wird vom Modell '{artifact_id}' nicht unterstützt."
             )
-        return f"{artifact_id}:{entity.sensor_id}:{entity.values}"
+        sensor_models = artifact.feature_models.get(entity.sensor_id, {})
+        if not sensor_models:
+            raise ValueError(f"Modell '{artifact_id}' enthält keine statistischen Parameter.")
 
-    def predict_batch(self, artifact_id: str, entities: Sequence[FeatureVector]) -> list[str]:
+        feature_names = sorted(set(sensor_models).intersection(entity.values))
+        if not feature_names:
+            raise ValueError(
+                f"Keine Eingabemerkmale werden vom Modell '{artifact_id}' unterstützt."
+            )
+
+        predictions: dict[str, float] = {}
+        explanations: dict[str, FeatureExplanation] = {}
+        confidences: list[float] = []
+        for feature_name in feature_names:
+            model = sensor_models[feature_name]
+            current_value = float(entity.values[feature_name])
+            if not math.isfinite(current_value):
+                raise ValueError("Vorhersagewerte müssen endlich sein.")
+            predicted_value = model.forecast(current_value)
+            predictions[feature_name] = predicted_value
+            explanations[feature_name] = explain_feature(
+                feature_name,
+                current_value,
+                predicted_value,
+                model,
+            )
+            confidences.append(model.confidence)
+
+        return PredictionResult(
+            artifact_id=artifact_id,
+            sensor_id=entity.sensor_id,
+            predictions=predictions,
+            confidence=sum(confidences) / len(confidences),
+            model_type=artifact.model_type,
+            explanations=explanations,
+        )
+
+    def predict_batch(
+        self,
+        artifact_id: str,
+        entities: Sequence[FeatureVector],
+    ) -> list[PredictionResult]:
         return [self.predict(artifact_id, entity) for entity in entities]
 
     @staticmethod

app/ml/registry/model_registry.py

@@ -2,13 +2,14 @@ from __future__ import annotations
 
 import json
 import logging
+import math
 import os
 from pathlib import Path
 import re
 from threading import RLock
 from collections.abc import Iterable
 
-from app.ml.training import TrainedArtifact
+from app.ml.training import FeatureModel, TrainedArtifact
 
 logger = logging.getLogger(__name__)
 
@@ -19,6 +20,8 @@ class ModelRegistry:
     def __init__(self, root: str | Path) -> None:
         self._root = Path(root).resolve()
         self._root.mkdir(parents=True, exist_ok=True)
+        self._archive_root = self._root / "archive"
+        self._archive_root.mkdir(parents=True, exist_ok=True)
         self._artifacts: dict[str, TrainedArtifact] = {}
         self._lock = RLock()
         self._load_existing()
@@ -42,29 +45,53 @@ class ModelRegistry:
                 raise KeyError(f"Artifact '{artifact_id}' nicht registriert.")
             return self._artifacts[artifact_id]
 
+    def get_optional(self, artifact_id: str) -> TrainedArtifact | None:
+        self._validate_artifact_id(artifact_id)
+        with self._lock:
+            return self._artifacts.get(artifact_id)
+
     def list_models(self) -> Iterable[TrainedArtifact]:
         with self._lock:
             return [self._artifacts[key] for key in sorted(self._artifacts)]
 
+    def archive(self, artifact_id: str) -> bool:
+        self._validate_artifact_id(artifact_id)
+        with self._lock:
+            artifact = self._artifacts.pop(artifact_id, None)
+            source = self._root / f"{artifact_id}.json"
+            if not source.exists():
+                return artifact is not None
+            target = self._archive_root / f"{artifact_id}.json"
+            os.replace(source, target)
+            logger.info("Modell archiviert: %s", target)
+            return True
+
     def _load_existing(self) -> None:
         for source in sorted(self._root.glob("*.json")):
             try:
                 raw = json.loads(source.read_text(encoding="utf-8"))
                 artifact_id = raw["artifact_id"]
                 supported_sensors = raw["supported_sensors"]
+                model_type = raw.get("model_type", "metadata")
+                raw_feature_models = raw.get("feature_models", {})
                 if not isinstance(artifact_id, str) or not isinstance(supported_sensors, list):
                     raise ValueError("invalid artifact structure")
+                if not isinstance(model_type, str):
+                    raise ValueError("model_type must be a string")
                 self._validate_artifact_id(artifact_id)
                 if source.name != f"{artifact_id}.json":
                     raise ValueError("artifact id does not match filename")
                 if not all(isinstance(sensor, str) for sensor in supported_sensors):
                     raise ValueError("supported_sensors must contain strings")
+                feature_models = _deserialize_feature_models(raw_feature_models)
             except (KeyError, TypeError, ValueError, json.JSONDecodeError) as exc:
                 raise ValueError(f"Ungültiges Modell-Artefakt: {source.name}") from exc
 
             self._artifacts[artifact_id] = TrainedArtifact(
                 artifact_id=artifact_id,
                 supported_sensors=tuple(supported_sensors),
+                feature_models=feature_models,
+                model_type=model_type,
             )
 
     def _persist(self, artifact: TrainedArtifact) -> None:
@@ -73,6 +100,22 @@ class ModelRegistry:
         payload = {
             "artifact_id": artifact.artifact_id,
             "supported_sensors": list(artifact.supported_sensors),
+            "model_type": artifact.model_type,
+            "feature_models": {
+                sensor_id: {
+                    feature_name: {
+                        "sample_count": model.sample_count,
+                        "mean": model.mean,
+                        "standard_deviation": model.standard_deviation,
+                        "minimum": model.minimum,
+                        "maximum": model.maximum,
+                        "slope": model.slope,
+                        "intercept": model.intercept,
+                    }
+                    for feature_name, model in sorted(models.items())
+                }
+                for sensor_id, models in sorted(artifact.feature_models.items())
+            },
         }
         temporary.write_text(
             json.dumps(payload, ensure_ascii=True, sort_keys=True) + "\n",
@@ -88,3 +131,51 @@ class ModelRegistry:
                 "artifact_id darf nur Buchstaben, Ziffern, Punkt, Unterstrich "
                 "und Bindestrich enthalten."
             )
+
+
+def _deserialize_feature_models(raw: object) -> dict[str, dict[str, FeatureModel]]:
+    if not isinstance(raw, dict):
+        raise ValueError("feature_models must be an object")
+
+    result: dict[str, dict[str, FeatureModel]] = {}
+    for sensor_id, raw_features in raw.items():
+        if not isinstance(sensor_id, str) or not isinstance(raw_features, dict):
+            raise ValueError("invalid sensor feature models")
+        features: dict[str, FeatureModel] = {}
+        for feature_name, raw_model in raw_features.items():
+            if not isinstance(feature_name, str) or not isinstance(raw_model, dict):
+                raise ValueError("invalid feature model")
+            sample_count = raw_model.get("sample_count")
+            if not isinstance(sample_count, int) or isinstance(sample_count, bool) or sample_count < 1:
+                raise ValueError("sample_count must be a positive integer")
+            values = {
+                key: _finite_number(raw_model.get(key))
+                for key in (
+                    "mean",
+                    "standard_deviation",
+                    "minimum",
+                    "maximum",
+                    "slope",
+                    "intercept",
+                )
+            }
+            features[feature_name] = FeatureModel(
+                sample_count=sample_count,
+                mean=values["mean"],
+                standard_deviation=values["standard_deviation"],
+                minimum=values["minimum"],
+                maximum=values["maximum"],
+                slope=values["slope"],
+                intercept=values["intercept"],
+            )
+        result[sensor_id] = features
+    return result
+
+
+def _finite_number(value: object) -> float:
+    if not isinstance(value, (int, float)) or isinstance(value, bool):
+        raise ValueError("feature model values must be finite numbers")
+    converted = float(value)
+    if not math.isfinite(converted):
+        raise ValueError("feature model values must be finite numbers")
+    return converted

app/ml/training.py

@@ -1,17 +1,44 @@
 from __future__ import annotations
 
 import logging
-from dataclasses import dataclass
+import math
+from collections import defaultdict
+from dataclasses import dataclass, field
 
 from app.ml.feature_store import FeatureStore
 
 logger = logging.getLogger(__name__)
 
 
-@dataclass
+@dataclass(frozen=True)
+class FeatureModel:
+    sample_count: int
+    mean: float
+    standard_deviation: float
+    minimum: float
+    maximum: float
+    slope: float
+    intercept: float
+
+    def forecast(self, current_value: float | None = None) -> float:
+        if current_value is not None:
+            return current_value + self.slope
+        return self.intercept + self.slope * self.sample_count
+
+    @property
+    def confidence(self) -> float:
+        sample_score = self.sample_count / (self.sample_count + 2)
+        scale = abs(self.mean) if abs(self.mean) > 1e-9 else 1.0
+        stability_score = 1.0 / (1.0 + self.standard_deviation / scale)
+        return min(0.99, max(0.05, sample_score * stability_score))
+
+
+@dataclass(frozen=True)
 class TrainedArtifact:
     artifact_id: str
     supported_sensors: tuple[str, ...]
+    feature_models: dict[str, dict[str, FeatureModel]] = field(default_factory=dict)
+    model_type: str = "statistical_baseline"
 
 
 class TrainingPipeline:
@@ -24,8 +51,33 @@ class TrainingPipeline:
         if not vectors:
             raise ValueError("FeatureStore enthält keine Trainingsdaten.")
 
-        sensors = tuple(sorted({vector.sensor_id for vector in vectors}))
+        samples: dict[str, dict[str, list[float]]] = defaultdict(lambda: defaultdict(list))
-        artifact = TrainedArtifact(artifact_id=artifact_id, supported_sensors=sensors)
+        for vector in vectors:
+            for feature_name, raw_value in vector.values.items():
+                value = float(raw_value)
+                if math.isfinite(value):
+                    samples[vector.sensor_id][feature_name].append(value)
+
+        feature_models = {
+            sensor_id: {
+                feature_name: _fit_feature(values)
+                for feature_name, values in sorted(features.items())
+                if values
+            }
+            for sensor_id, features in sorted(samples.items())
+        }
+        feature_models = {
+            sensor_id: models for sensor_id, models in feature_models.items() if models
+        }
+        if not feature_models:
+            raise ValueError("Trainingsdaten enthalten keine endlichen numerischen Werte.")
+
+        sensors = tuple(feature_models)
+        artifact = TrainedArtifact(
+            artifact_id=artifact_id,
+            supported_sensors=sensors,
+            feature_models=feature_models,
+        )
         self._artifacts[artifact_id] = artifact
         logger.info("Training abgeschlossen für %s mit %d Sensoren", artifact_id, len(sensors))
         return artifact
@@ -34,3 +86,32 @@ class TrainingPipeline:
         if artifact_id not in self._artifacts:
             raise KeyError(f"Artifact '{artifact_id}' nicht gefunden.")
         return self._artifacts[artifact_id]
+
+
+def _fit_feature(values: list[float]) -> FeatureModel:
+    sample_count = len(values)
+    mean = sum(values) / sample_count
+    variance = sum((value - mean) ** 2 for value in values) / sample_count
+    standard_deviation = math.sqrt(variance)
+
+    if sample_count == 1:
+        slope = 0.0
+        intercept = mean
+    else:
+        x_mean = (sample_count - 1) / 2
+        denominator = sum((index - x_mean) ** 2 for index in range(sample_count))
+        numerator = sum(
+            (index - x_mean) * (value - mean) for index, value in enumerate(values)
+        )
+        slope = numerator / denominator
+        intercept = mean - slope * x_mean
+
+    return FeatureModel(
+        sample_count=sample_count,
+        mean=mean,
+        standard_deviation=standard_deviation,
+        minimum=min(values),
+        maximum=max(values),
+        slope=slope,
+        intercept=intercept,
+    )

app/static/index.html

@@ -0,0 +1,355 @@
+<!doctype html>
+<html lang="de">
+<head>
+  <meta charset="utf-8">
+  <meta name="viewport" content="width=device-width,initial-scale=1">
+  <title>SillyHome Next</title>
+  <style>
+    :root { color-scheme: dark; font-family: system-ui, sans-serif; background: #10151c; color: #eaf1f8; }
+    body { margin: 0; }
+    header { padding: 20px; background: linear-gradient(135deg,#142b3a,#193f36); }
+    h1,h2,h3 { margin: 0 0 12px; }
+    header p { margin: 4px 0; color: #b9c9d6; }
+    main { display: grid; grid-template-columns: repeat(auto-fit,minmax(320px,1fr)); gap: 14px; padding: 14px; }
+    section { background: #18212b; border: 1px solid #2d3a47; border-radius: 12px; padding: 16px; }
+    .wide { grid-column: 1 / -1; }
+    .ok { color: #66dfa9; }
+    .warn { color: #f3c969; }
+    .bad { color: #ff8f8f; }
+    label { display: block; margin: 9px 0 4px; color: #b9c9d6; }
+    input,select,textarea,button { box-sizing: border-box; width: 100%; border-radius: 7px; border: 1px solid #3b4b5b; padding: 9px; background: #101820; color: #fff; }
+    button { margin-top: 10px; background: #23715b; border: 0; font-weight: 700; cursor: pointer; }
+    button.secondary { background: #37495c; }
+    pre { white-space: pre-wrap; overflow: auto; background: #0d141b; padding: 10px; border-radius: 7px; }
+    table { width: 100%; border-collapse: collapse; font-size: .9rem; }
+    td,th { padding: 7px; border-bottom: 1px solid #2d3a47; text-align: left; vertical-align: top; }
+    ul { margin: 8px 0; padding-left: 18px; }
+    .notice { border-left: 4px solid #e8b34b; padding-left: 10px; }
+    .grid-two { display:grid; grid-template-columns:repeat(auto-fit,minmax(220px,1fr)); gap:8px; }
+    .chips { display:flex; flex-wrap:wrap; gap:6px; margin-top:8px; }
+    .chip { padding:4px 8px; border-radius:999px; background:#22303c; border:1px solid #31404d; font-size:.85rem; }
+  </style>
+</head>
+<body>
+<header>
+  <h1>SillyHome Next</h1>
+  <p>Aktuator-zentrierte Home-Assistant-Analyse mit nachvollziehbarer Sensorzuordnung und kontrolliertem Modell-Lebenszyklus.</p>
+  <p class="notice">Sicherheitsmodus: SillyHome führt niemals selbst Aktor-Services aus. Automationen bleiben manuell freizugebende YAML-Entwürfe.</p>
+</header>
+<main>
+  <section>
+    <h2>Systemstatus</h2>
+    <div id="status">Prüfung läuft ...</div>
+    <div class="chips" id="status-chips"></div>
+    <button class="secondary" onclick="loadOverview()">Neu laden</button>
+    <button onclick="runReconciliation()">Reconciliation ausführen</button>
+  </section>
+
+  <section>
+    <h2>Aktuator wählen</h2>
+    <label for="actuator-select">Home-Assistant-Aktor</label>
+    <select id="actuator-select"></select>
+    <button onclick="configureActuator()">Aktuator übernehmen</button>
+    <pre id="actuator-config-result">Noch kein Aktuator konfiguriert.</pre>
+  </section>
+
+  <section class="wide">
+    <h2>Konfigurierte Aktuatoren</h2>
+    <div id="configured-actuators">Noch nicht geladen.</div>
+  </section>
+
+  <section class="wide">
+    <h2>Zuordnung und Modellstatus</h2>
+    <div id="actuator-detail">Einen konfigurierten Aktuator auswählen.</div>
+  </section>
+
+  <section class="wide">
+    <h2>Automation-Entwurf</h2>
+    <p>Der Entwurf muss explizit freigegeben werden. Auch danach wird nur YAML exportiert, nichts geschaltet.</p>
+    <div class="grid-two">
+      <div><label for="alias">Name</label><input id="alias" value="Licht bei Dunkelheit"></div>
+      <div><label for="trigger">Trigger-Entity</label><input id="trigger" placeholder="sensor.flur_illuminance"></div>
+      <div><label for="below">Unter Grenzwert</label><input id="below" type="number" value="10"></div>
+      <div><label for="service">Dienst</label><select id="service"><option>light.turn_on</option><option>light.turn_off</option><option>switch.turn_on</option><option>switch.turn_off</option></select></div>
+      <div><label for="target">Ziel-Entity</label><input id="target" placeholder="light.flur"></div>
+    </div>
+    <button onclick="createProposal()">Entwurf speichern</button>
+    <button class="secondary" onclick="loadProposals()">Entwürfe aktualisieren</button>
+    <div id="proposals"></div>
+  </section>
+</main>
+<script>
+const pretty = value => JSON.stringify(value, null, 2);
+let currentActuatorId = null;
+
+async function api(path, options = {}) {
+  const response = await fetch(path, {headers: {"Content-Type": "application/json"}, ...options});
+  const body = await response.json().catch(() => ({}));
+  if (!response.ok) throw new Error(body.detail || `${response.status} ${response.statusText}`);
+  return body;
+}
+
+function statusClass(record) {
+  if (record.assignment.review_required) return "warn";
+  if (record.lifecycle.status === "trained") return "ok";
+  if (record.lifecycle.status === "review_required" || record.lifecycle.status === "invalid") return "warn";
+  return "bad";
+}
+
+function renderEvidence(evidence) {
+  return evidence.length ? `<ul>${evidence.map(item => `<li>${item}</li>`).join("")}</ul>` : "<span class='bad'>Keine Evidenz</span>";
+}
+
+async function loadOverview() {
+  const status = document.getElementById("status");
+  const chips = document.getElementById("status-chips");
+  try {
+    const [health, ml, reconciliation, actuators] = await Promise.all([
+      api("health"),
+      api("ml/health"),
+      api("v1/actuators/reconciliation/state"),
+      api("v1/actuators"),
+    ]);
+    status.innerHTML = `<p class="ok">API und ML bereit</p><p>Letzte Reconciliation: ${reconciliation.last_completed_at || "noch nie"}</p><p>${reconciliation.last_summary}</p>`;
+    chips.innerHTML = [
+      `<span class="chip">Health: ${health.status}</span>`,
+      `<span class="chip">ML: ${ml.status}</span>`,
+      `<span class="chip">Aktuatoren: ${actuators.length}</span>`,
+      `<span class="chip">Trainierte Modelle: ${reconciliation.trained_models}</span>`,
+    ].join("");
+  } catch (error) {
+    status.innerHTML = `<p class="bad">${error.message}</p>`;
+    chips.innerHTML = "";
+  }
+  await Promise.all([loadActuatorDiscovery(), loadConfiguredActuators(), loadProposals()]);
+}
+
+async function loadActuatorDiscovery() {
+  const select = document.getElementById("actuator-select");
+  try {
+    const actuators = await api("v1/actuators/discovery");
+    select.innerHTML = actuators.length
+      ? actuators.map(entity => `<option value="${entity.entity_id}">${entity.friendly_name || entity.entity_id}${entity.area_name ? ` (${entity.area_name})` : ""}</option>`).join("")
+      : "<option value=''>Keine Aktuatoren gefunden</option>";
+  } catch (error) {
+    select.innerHTML = `<option value="">${error.message}</option>`;
+  }
+}
+
+async function configureActuator() {
+  const actuatorId = document.getElementById("actuator-select").value;
+  const box = document.getElementById("actuator-config-result");
+  if (!actuatorId) return;
+  try {
+    const record = await api("v1/actuators", {
+      method: "POST",
+      body: JSON.stringify({actuator_entity_id: actuatorId}),
+    });
+    currentActuatorId = record.actuator_entity_id;
+    box.textContent = pretty(record);
+    await loadOverview();
+    await showActuator(record.actuator_entity_id);
+  } catch (error) {
+    box.textContent = error.message;
+  }
+}
+
+async function runReconciliation() {
+  try {
+    await api("v1/actuators/reconciliation/run", {method: "POST"});
+    await loadOverview();
+    if (currentActuatorId) await showActuator(currentActuatorId);
+  } catch (error) {
+    alert(error.message);
+  }
+}
+
+async function loadConfiguredActuators() {
+  const box = document.getElementById("configured-actuators");
+  try {
+    const rows = await api("v1/actuators");
+    box.innerHTML = rows.length ? `
+      <table>
+        <tr><th>Aktuator</th><th>Numerischer Sensor</th><th>Review</th><th>Modellstatus</th><th>Letztes Training</th><th>Aktion</th></tr>
+        ${rows.map(record => `
+          <tr>
+            <td>${record.actuator_entity_id}</td>
+            <td>${record.assignment.selected_numeric_entity_id || "-"}</td>
+            <td class="${record.assignment.review_required ? "warn" : "ok"}">${record.assignment.review_required ? "erforderlich" : "nein"}</td>
+            <td class="${statusClass(record)}">${record.lifecycle.status}</td>
+            <td>${record.lifecycle.last_trained_at || "-"}</td>
+            <td><button onclick="showActuator('${record.actuator_entity_id}')">Details</button></td>
+          </tr>
+        `).join("")}
+      </table>` : "<p>Keine konfigurierten Aktuatoren.</p>";
+  } catch (error) {
+    box.textContent = error.message;
+  }
+}
+
+async function showActuator(actuatorId) {
+  currentActuatorId = actuatorId;
+  const box = document.getElementById("actuator-detail");
+  try {
+    const record = await api(`v1/actuators/${encodeURIComponent(actuatorId)}`);
+    const numericRows = record.numeric_candidates.map(candidate => `
+      <tr>
+        <td>${candidate.entity_id}</td>
+        <td>${candidate.score.toFixed(3)} / ${candidate.confidence.toFixed(2)}</td>
+        <td>${candidate.auto_accepted ? "<span class='ok'>auto</span>" : "<span class='warn'>review</span>"}</td>
+        <td>${renderEvidence(candidate.evidence)}</td>
+      </tr>
+    `).join("");
+    const contextRows = record.context_candidates.map(candidate => `
+      <tr>
+        <td>${candidate.entity_id}</td>
+        <td>${candidate.score.toFixed(3)} / ${candidate.confidence.toFixed(2)}</td>
+        <td>${candidate.auto_accepted ? "<span class='ok'>auto</span>" : "<span class='warn'>optional</span>"}</td>
+        <td>${renderEvidence(candidate.evidence)}</td>
+      </tr>
+    `).join("");
+    box.innerHTML = `
+      <div class="grid-two">
+        <div>
+          <h3>Auswahl</h3>
+          <p><strong>Aktuator:</strong> ${record.actuator_entity_id}</p>
+          <p><strong>Numerischer Sensor:</strong> ${record.assignment.selected_numeric_entity_id || "-"}</p>
+          <p><strong>Kontext:</strong> ${record.assignment.selected_context_entity_ids.join(", ") || "-"}</p>
+          <p><strong>Quelle:</strong> ${record.assignment.source}</p>
+          <p><strong>Review:</strong> <span class="${record.assignment.review_required ? "warn" : "ok"}">${record.assignment.review_required ? "erforderlich" : "nicht erforderlich"}</span></p>
+          <p><strong>Begruendung:</strong> ${record.assignment.reason}</p>
+        </div>
+        <div>
+          <h3>Modell-Lebenszyklus</h3>
+          <p><strong>Status:</strong> <span class="${statusClass(record)}">${record.lifecycle.status}</span></p>
+          <p><strong>Letztes Training:</strong> ${record.lifecycle.last_trained_at || "-"}</p>
+          <p><strong>Messpunkte:</strong> ${record.lifecycle.last_history_point_count}</p>
+          <p><strong>Grund:</strong> ${record.lifecycle.reason}</p>
+          <p><strong>Nächste Aktion:</strong> ${record.lifecycle.next_action}</p>
+          <button onclick="reconcileActuator('${record.actuator_entity_id}')">Diesen Aktuator erneut prüfen</button>
+        </div>
+      </div>
+      <div class="grid-two">
+        <div>
+          <h3>Manuelle Overrides</h3>
+          <label for="override-numeric">Numerischer Sensor</label>
+          <input id="override-numeric" value="${record.manual_override?.numeric_entity_id || record.assignment.selected_numeric_entity_id || ""}">
+          <label for="override-context">Kontext-Entities (kommagetrennt)</label>
+          <textarea id="override-context">${(record.manual_override?.context_entity_ids || record.assignment.selected_context_entity_ids || []).join(", ")}</textarea>
+          <label for="override-note">Notiz</label>
+          <input id="override-note" value="${record.manual_override?.note || ""}">
+          <button onclick="saveOverride('${record.actuator_entity_id}')">Override speichern</button>
+          <button class="secondary" onclick="clearOverride('${record.actuator_entity_id}')">Override löschen</button>
+        </div>
+        <div>
+          <h3>Audit</h3>
+          <pre>${pretty(record.lifecycle.audit)}</pre>
+        </div>
+      </div>
+      <h3>Numerische Kandidaten</h3>
+      ${numericRows ? `<table><tr><th>Entity</th><th>Score / Confidence</th><th>Auto</th><th>Evidenz</th></tr>${numericRows}</table>` : "<p>Keine Kandidaten.</p>"}
+      <h3>Kontext-Kandidaten</h3>
+      ${contextRows ? `<table><tr><th>Entity</th><th>Score / Confidence</th><th>Auto</th><th>Evidenz</th></tr>${contextRows}</table>` : "<p>Keine Kandidaten.</p>"}
+    `;
+  } catch (error) {
+    box.textContent = error.message;
+  }
+}
+
+async function reconcileActuator(actuatorId) {
+  try {
+    await api(`v1/actuators/${encodeURIComponent(actuatorId)}/reconcile`, {method: "POST"});
+    await loadOverview();
+    await showActuator(actuatorId);
+  } catch (error) {
+    alert(error.message);
+  }
+}
+
+async function saveOverride(actuatorId) {
+  const numeric = document.getElementById("override-numeric").value.trim() || null;
+  const contexts = document.getElementById("override-context").value
+    .split(",")
+    .map(item => item.trim())
+    .filter(Boolean);
+  const note = document.getElementById("override-note").value.trim() || null;
+  try {
+    await api(`v1/actuators/${encodeURIComponent(actuatorId)}/override`, {
+      method: "POST",
+      body: JSON.stringify({
+        numeric_entity_id: numeric,
+        context_entity_ids: contexts,
+        note,
+      }),
+    });
+    await loadOverview();
+    await showActuator(actuatorId);
+  } catch (error) {
+    alert(error.message);
+  }
+}
+
+async function clearOverride(actuatorId) {
+  try {
+    await api(`v1/actuators/${encodeURIComponent(actuatorId)}/override`, {
+      method: "POST",
+      body: JSON.stringify({clear: true}),
+    });
+    await loadOverview();
+    await showActuator(actuatorId);
+  } catch (error) {
+    alert(error.message);
+  }
+}
+
+async function createProposal() {
+  try {
+    await api("v1/automations/proposals", {method: "POST", body: JSON.stringify({
+      alias: document.getElementById("alias").value,
+      description: "Manuell im SillyHome-Dashboard erstellter und nicht automatisch ausgeführter Entwurf.",
+      trigger: {entity_id: document.getElementById("trigger").value, below: Number(document.getElementById("below").value)},
+      action: {service: document.getElementById("service").value, entity_id: document.getElementById("target").value, data: {}}
+    })});
+    await loadProposals();
+  } catch (error) {
+    alert(error.message);
+  }
+}
+
+async function decide(id, revision, action) {
+  try {
+    await api(`v1/automations/proposals/${id}/${action}`, {method: "POST", body: JSON.stringify({expected_revision: revision})});
+    await loadProposals();
+  } catch (error) {
+    alert(error.message);
+  }
+}
+
+async function loadProposals() {
+  const box = document.getElementById("proposals");
+  try {
+    const rows = await api("v1/automations/proposals");
+    box.innerHTML = rows.length ? `
+      <table>
+        <tr><th>Name</th><th>Status</th><th>Aktion</th></tr>
+        ${rows.map(item => `
+          <tr>
+            <td>${item.alias}</td>
+            <td>${item.status}</td>
+            <td>${item.status === "draft"
+              ? `<button onclick="decide('${item.proposal_id}',${item.revision},'approve')">Freigeben</button><button class="secondary" onclick="decide('${item.proposal_id}',${item.revision},'reject')">Ablehnen</button>`
+              : item.status === "approved"
+                ? `<a href="v1/automations/proposals/${item.proposal_id}/yaml">YAML laden</a>`
+                : "-"}</td>
+          </tr>
+        `).join("")}
+      </table>` : "<p>Keine Entwürfe.</p>";
+  } catch (error) {
+    box.textContent = error.message;
+  }
+}
+
+loadOverview();
+</script>
+</body>
+</html>

backend/routes/ml.py

@@ -7,6 +7,7 @@ from collections.abc import Sequence
 from fastapi import APIRouter, FastAPI, HTTPException, Request, status
 from pydantic import BaseModel, Field
 
+from app.ml.evaluation import Evaluator
 from app.ml.feature_store import FeatureVector
 from app.ml.predictor import Predictor
 from app.ml.registry.model_registry import ModelRegistry
@@ -31,7 +32,25 @@ class PredictRequest(BaseModel):
 class PredictResponse(BaseModel):
     model_id: str
     sensor_id: str
-    prediction: str
+    predictions: dict[str, float]
+    confidence: float
+    model_type: str
+    explanations: dict[str, "FeatureExplanationResponse"]
+
+
+class FeatureExplanationResponse(BaseModel):
+    feature: str
+    current_value: float
+    predicted_value: float
+    change: float
+    direction: str
+    sample_count: int
+    historical_mean: float
+    historical_range: tuple[float, float]
+    standard_deviation: float
+    trend_per_step: float
+    confidence: float
+    summary: str
 
 
 class BatchRequest(BaseModel):
@@ -60,9 +79,28 @@ class RetrainRequest(BaseModel):
 class RetrainResponse(BaseModel):
     model_id: str
     supported_sensors: list[str]
+    trained_features: int
+    model_type: str
     replaced: bool
 
 
+class EvaluateRequest(BaseModel):
+    model_id: str = Field(..., alias="modelId", min_length=1, max_length=128)
+    samples: list[TrainingSample] = Field(min_length=1)
+
+
+class MetricResponse(BaseModel):
+    name: str
+    value: float
+    threshold: float | None = None
+
+
+class EvaluateResponse(BaseModel):
+    model_id: str
+    sample_size: int
+    metrics: list[MetricResponse]
+
+
 @router.get("/health", response_model=HealthResponse, status_code=200)
 def health() -> HealthResponse:
     return HealthResponse(status="ok")
@@ -96,10 +134,50 @@ def retrain(payload: RetrainRequest, request: Request) -> RetrainResponse:
     return RetrainResponse(
         model_id=result.artifact.artifact_id,
         supported_sensors=list(result.artifact.supported_sensors),
+        trained_features=sum(
+            len(feature_models)
+            for feature_models in result.artifact.feature_models.values()
+        ),
+        model_type=result.artifact.model_type,
         replaced=result.replaced,
     )
 
 
+@router.post("/evaluate", response_model=EvaluateResponse, status_code=200)
+def evaluate(payload: EvaluateRequest, request: Request) -> EvaluateResponse:
+    registry = _require_registry(request)
+    vectors = [
+        FeatureVector(
+            sensor_id=sample.sensor_id,
+            values=sample.values,
+            label=sample.label,
+        )
+        for sample in payload.samples
+    ]
+    try:
+        report = Evaluator(registry=registry).evaluate(payload.model_id, vectors)
+    except ValueError as exc:
+        try:
+            registry.load_artifact(payload.model_id)
+        except KeyError:
+            raise HTTPException(
+                status_code=status.HTTP_404_NOT_FOUND,
+                detail=str(exc),
+            ) from exc
+        raise HTTPException(
+            status_code=status.HTTP_422_UNPROCESSABLE_CONTENT,
+            detail=str(exc),
+        ) from exc
+    return EvaluateResponse(
+        model_id=report.artifact_id,
+        sample_size=report.sample_size,
+        metrics=[
+            MetricResponse(name=metric.name, value=metric.value, threshold=metric.threshold)
+            for metric in report.metrics
+        ],
+    )
+
+
 @router.post("/predict", response_model=PredictResponse, status_code=200)
 def predict(payload: PredictRequest, request: Request) -> PredictResponse:
     registry = _require_registry(request)
@@ -117,7 +195,13 @@ def predict(payload: PredictRequest, request: Request) -> PredictResponse:
     return PredictResponse(
         model_id=payload.model_id,
         sensor_id=payload.sensor_id,
-        prediction=prediction,
+        predictions=prediction.predictions,
+        confidence=prediction.confidence,
+        model_type=prediction.model_type,
+        explanations={
+            name: FeatureExplanationResponse(**explanation.__dict__)
+            for name, explanation in prediction.explanations.items()
+        },
     )
 
 
@@ -138,7 +222,17 @@ def predict_batch(payload: BatchRequest, request: Request) -> BatchResponse:
                 detail=str(exc),
             ) from exc
         responses.append(
-            PredictResponse(model_id=item.model_id, sensor_id=item.sensor_id, prediction=prediction)
+            PredictResponse(
+                model_id=item.model_id,
+                sensor_id=item.sensor_id,
+                predictions=prediction.predictions,
+                confidence=prediction.confidence,
+                model_type=prediction.model_type,
+                explanations={
+                    name: FeatureExplanationResponse(**explanation.__dict__)
+                    for name, explanation in prediction.explanations.items()
+                },
+            )
         )
     return BatchResponse(predictions=responses)
 

docker-compose.yml

@@ -8,8 +8,16 @@ services:
         required: false
     environment:
       SILLYHOME_MODEL_STORE: /app/data/models
+      SILLYHOME_AUTOMATION_STORE: /app/data/automations
+      SILLYHOME_ACTUATOR_STORE: /app/data/actuators
+      SILLYHOME_HISTORY_DAYS: 14
+      SILLYHOME_MIN_TRAINING_POINTS: 24
+      SILLYHOME_RETRAIN_STALE_HOURS: 24
+      SILLYHOME_RECONCILE_INTERVAL_SECONDS: 900
     volumes:
       - model-data:/app/data/models
+      - automation-data:/app/data/automations
+      - actuator-data:/app/data/actuators
     read_only: true
     tmpfs:
       - /tmp
@@ -21,3 +29,5 @@ services:
 
 volumes:
   model-data:
+  automation-data:
+  actuator-data:

docs/automations.md

@@ -0,0 +1,14 @@
+# Automation-Vorschläge
+
+SillyHome Next führt Automationen niemals automatisch aus. Der Workflow ist:
+
+1. Vorschlag als `draft` erstellen.
+2. Inhalt und Ziel-Entity prüfen.
+3. Mit aktueller Revision explizit freigeben oder ablehnen.
+4. Nur freigegebene Vorschläge als Home-Assistant-YAML exportieren.
+5. Das YAML außerhalb von SillyHome Next in Home Assistant importieren.
+
+Erlaubt sind numerische Sensor-Trigger und Aktionsdienste aus den Domains
+`light`, `switch`, `climate`, `fan` und `cover`. Shell-Kommandos, Skripte und
+beliebige Service-Domains werden abgewiesen. Eine einmal getroffene Entscheidung
+kann nicht überschrieben werden; Änderungen benötigen einen neuen Vorschlag.

docs/ha_data.md

@@ -0,0 +1,52 @@
+# Home-Assistant-Datenpipeline
+
+SillyHome Next trennt aktuelle Entity-Metadaten, Discovery und historische
+Messwerte. Dadurch gelangen nur klassifizierte, geeignete Daten in spätere
+Trainings- und Erklärungsprozesse.
+
+## Entity Discovery
+
+`GET /v1/discovery` klassifiziert Home-Assistant-Entities in:
+
+- `measurement`: numerische Messsensoren, für Training geeignet
+- `binary_context`: binäre Kontextsensoren wie Bewegung oder Anwesenheit
+- `context`: Personen-, Wetter- und Standortkontext
+- `actuator`: mögliche Automationsziele, nicht als Trainingssensor verwendet
+- `unsupported`: noch nicht klassifizierte Entity-Typen
+
+Zusätzlich reichert `HaReader` verfügbare Metadaten wie `friendly_name`,
+Bereich und Gerät aus Home Assistant an. Für die aktor-zentrierte Zuordnung
+nutzt SillyHome Next bevorzugt:
+
+- `area_id` und `area_name`
+- `device_id` und `device_name`
+- Friendly Names und Entity-ID-Tokens
+- Domain und `device_class`
+
+Optionale Query-Parameter:
+
+- `domain=sensor` kann mehrfach angegeben werden
+- `learnable=true|false` filtert nach Trainingsrelevanz
+
+## Historische Daten
+
+Historische Zustände werden über Home Assistants
+`/api/history/period/<start>`-Schnittstelle geladen. Abfragen verlangen:
+
+- mindestens eine Entity-ID, maximal 100
+- zeitzonenbehaftete Start- und Endzeit
+- ein Enddatum nach dem Startdatum
+- maximal 31 Tage pro Abfrage
+
+Die Normalisierung übernimmt nur endliche numerische Zustände. `unknown`,
+`unavailable`, nichtnumerische Werte, `NaN` und unendliche Werte werden nicht
+als Trainingsdaten verwendet. Ergebnisse werden je Entity chronologisch
+sortiert. Binäre Kontext-Entities werden bewusst nicht in numerische
+Trainingsreihen konvertiert.
+
+## Datenschutz und Betrieb
+
+Die Daten bleiben lokal. Home-Assistant-Tokens gehören ausschließlich in die
+Umgebungskonfiguration und dürfen nicht protokolliert oder versioniert werden.
+Die API sollte nur lokal oder hinter einem authentifizierenden Reverse Proxy
+erreichbar sein.

docs/ml_api.md

@@ -1,11 +1,9 @@
 # ML-Serving-API
 
 Diese Dokumentation beschreibt die REST-Endpunkte der aktuellen
-Modell-Artefakt- und Vorhersage-Schnittstelle.
+Modell-Artefakt-, Vorhersage- und aktor-zentrierten Lifecycle-Schnittstelle.
 
-> Hinweis: Version 0.1.0 enthält noch kein statistisch trainiertes ML-Modell.
+Das Serving verwendet ein lokal trainiertes statistisches Baseline-Modell.
-> Die Vorhersage ist eine deterministische Referenzimplementierung für den
-> späteren Modellvertrag.
 
 ## Basis-URL
 
@@ -13,9 +11,12 @@ Modell-Artefakt- und Vorhersage-Schnittstelle.
 - Health: `/health`
 - Modelle: `/models`
 - Retraining: `/retrain`
+- Evaluation: `/evaluate`
 - Einzelvorhersage: `/predict`
 - Batchvorhersage: `/batch`
 
+Die aktor-zentrierte API liegt unter `/v1/actuators`.
+
 Der Standard-Start erfolgt über `uvicorn app.main:app`, danach stehen HA- und
 ML-Routen in derselben Anwendung bereit.
 
@@ -62,10 +63,27 @@ Einzelne Vorhersage für einen Sensor.
 {
   "model_id": "default",
   "sensor_id": "sensor.kitchen",
-  "prediction": "default:sensor.kitchen:{'temperature': 21.0}"
+  "predictions": {"temperature": 21.4},
+  "confidence": 0.78,
+  "model_type": "statistical_baseline",
+  "explanations": {
+    "temperature": {
+      "direction": "steigend",
+      "change": 0.4,
+      "sample_count": 24,
+      "historical_mean": 20.7,
+      "trend_per_step": 0.4,
+      "summary": "temperature: steigend; Prognose ..."
+    }
+  }
 }
 ```
 
+Die Erklärung nennt pro Merkmal den aktuellen und prognostizierten Wert,
+Richtung, Veränderung, Datenbasis, historischen Bereich, Streuung, Trend und
+Confidence. Sie wird deterministisch aus den gespeicherten Modellparametern
+erzeugt.
+
 ### `POST /ml/retrain`
 
 Trainiert die Artefakt-Metadaten aus neuen Sensordaten. Existiert `modelId`
@@ -91,10 +109,17 @@ dem Modellverzeichnis geladen.
 {
   "model_id": "home-model",
   "supported_sensors": ["sensor.kitchen"],
+  "trained_features": 1,
+  "model_type": "statistical_baseline",
   "replaced": false
 }
 ```
 
+### `POST /ml/evaluate`
+
+Vergleicht Modellvorhersagen mit Validierungsdaten und liefert MAE, RMSE und
+Coverage. Der Request verwendet dasselbe Sample-Format wie `/ml/retrain`.
+
 ### `POST /ml/batch`
 
 Batch-Vorhersage für mehrere Sensorwerte.
@@ -124,12 +149,16 @@ Batch-Vorhersage für mehrere Sensorwerte.
     {
       "model_id": "default",
       "sensor_id": "sensor.kitchen",
-      "prediction": "default:sensor.kitchen:{'temperature': 21.0}"
+      "predictions": {"temperature": 21.4},
+      "confidence": 0.78,
+      "model_type": "statistical_baseline"
     },
     {
       "model_id": "default",
       "sensor_id": "sensor.bedroom",
-      "prediction": "default:sensor.bedroom:{'temperature': 18.5}"
+      "predictions": {"temperature": 18.3},
+      "confidence": 0.74,
+      "model_type": "statistical_baseline"
     }
   ]
 }
@@ -141,14 +170,45 @@ Batch-Vorhersage für mehrere Sensorwerte.
 - `422 Unprocessable Content`: Sensor wird vom Modell nicht unterstützt oder Eingabe ist ungültig.
 - `503 Service Unavailable`: Registry ist nicht initialisiert.
 
+## Aktuator-zentrierte API
+
+### `GET /v1/actuators/discovery`
+
+Listet unterstützte Aktuatoren mit angereicherter HA-Metadatenbasis.
+
+### `POST /v1/actuators`
+
+Registriert einen Aktuator, ermittelt passende numerische Sensoren und
+Kontext-Entities, trainiert bei ausreichender History automatisch ein Modell und
+liefert Zuordnung, Confidence, Evidenz und Lifecycle-Status zurück.
+
+**Request**
+```json
+{
+  "actuator_entity_id": "light.abstellkammer",
+  "enabled": true
+}
+```
+
+### `POST /v1/actuators/{actuator_entity_id}/override`
+
+Persistiert manuelle Overrides. Diese haben Vorrang vor der automatischen
+Heuristik und überstehen Neustarts.
+
+### `POST /v1/actuators/reconciliation/run`
+
+Führt eine sichere globale Reconciliation aus. Die periodische Add-on-Schleife
+ruft denselben idempotenten Ablauf auf, startet aber keine Services in Home
+Assistant.
+
 ## Betrieb
 
-Die produktive App lädt Artefakte aus `SILLYHOME_MODEL_STORE`. Neue Artefakte
+Die produktive App lädt Artefakte aus `SILLYHOME_MODEL_STORE`. Aktuator-,
-werden über `/ml/retrain`, `RetrainingService` oder direkt über
+Override- und Reconciliation-Zustände liegen atomisch in
-`ModelRegistry.register(...)` registriert. Die Registry speichert validiertes
+`SILLYHOME_ACTUATOR_STORE`. Neue Artefakte werden über `/ml/retrain`,
-JSON atomisch und lädt es beim Neustart. Die API sollte nur in einem
+`RetrainingService` oder den aktor-zentrierten Lifecycle registriert. Die API
-vertrauenswürdigen Netz oder hinter einem authentifizierenden Reverse Proxy
+sollte nur in einem vertrauenswürdigen Netz oder hinter einem
-erreichbar sein.
+authentifizierenden Reverse Proxy erreichbar sein.
 
 ## Verweise
 

docs/ml_training.md

@@ -1,14 +1,22 @@
 # ML Training- und Evaluations-Workflow
 
-Dieser Workflow beschreibt den aktuellen Platzhalter für Modell-Metadaten,
+SillyHome Next trainiert ein lokales statistisches Baseline-Modell pro Sensor
-Evaluation und Serving. Er trainiert in Version 0.1.0 noch kein statistisches
+und Merkmal. Es benötigt keine Cloud und keine externe ML-Laufzeit.
-Modell.
+
+Seit `v0.4.0` ist der bevorzugte Weg aktor-zentriert: ein bestätigter Aktuator
+wird mit einem numerischen Primärsensor verknüpft, die Historie dieses Sensors
+wird automatisch geladen und in ein deterministisches Artefakt überführt.
 
 ## 1. Daten sammeln
 
 Alle Trainingsvektoren werden über `FeatureStore.add(...)` oder `add_batch(...)` eingepflegt. Jeder Vektor enthält eine Sensor-ID sowie ein Dictionary mit Merkmalen.
 
-## 2. Artefakt-Metadaten erzeugen
+Im Normalbetrieb erzeugt die Reconciliation diese Vektoren selbst aus realer
+Home-Assistant-History. Das Trainingsmerkmal heißt dabei immer `value`.
+Binäre Kontextsensoren bleiben Kontext und werden nicht als numerische Samples
+missverstanden.
+
+## 2. Statistisches Artefakt erzeugen
 
 ```python
 store = FeatureStore()
@@ -18,21 +26,30 @@ artifact = pipeline.run("my_artifact")
 pipeline.export("my_artifact")
 ```
 
-`TrainingPipeline.run(...)` erzeugt ein `TrainedArtifact` mit den unterstützten
+`TrainingPipeline.run(...)` berechnet für jedes numerische Merkmal:
-Sensor-IDs. Gewichte, Parameter oder ein echtes Modell werden noch nicht
+
-berechnet.
+- Stichprobenzahl
+- Mittelwert und Standardabweichung
+- Minimum und Maximum
+- linearen Trend mit Steigung und Achsenabschnitt
+
+Die nächste Vorhersage kombiniert den letzten beobachteten Wert mit der
+trainierten Trendsteigung. Die Confidence berücksichtigt Datenmenge und
+Stabilität.
 
 ## 3. Modell evaluieren
 
 ```python
 evaluator = Evaluator(pipeline)
-report = evaluator.evaluate(artifact.artifact_id, predictions)
+report = evaluator.evaluate(artifact.artifact_id, validation_samples)
 ```
 
-Der Report enthält:
+Der Report enthält echte numerische Vergleichsmetriken:
 - `artifact_id`
 - `sample_size`
-- Metriken wie `coverage` und `unknown_rate` mit Default-Schwellenwerten
+- `mae` (Mean Absolute Error)
+- `rmse` (Root Mean Squared Error)
+- `coverage` für den Anteil auswertbarer Merkmale
 
 ## 4. Modell registrieren
 
@@ -53,7 +70,23 @@ zustandslose Funktion `retrain_model(registry, artifact_id, vectors)` aufrufen.
 Der Service startet bewusst keinen eigenen Hintergrundprozess. Über
 `POST /ml/retrain` kann derselbe Ablauf per API angestoßen werden.
 
+## 6. Autonomer Lebenszyklus
+
+Der `ActuatorReconciliationService` verwaltet pro konfiguriertem Aktuator:
+
+- die automatische Sensor- und Kontextzuordnung mit Score, Confidence und Evidenz
+- persistente manuelle Overrides
+- den Modellstatus (`trained`, `pending_history`, `review_required`, `archived`, ...)
+- ein Audit-Protokoll mit Gründen für Training, Retraining oder Archivierung
+
+Retraining erfolgt nur, wenn:
+
+- genügend nutzbare numerische Historie vorliegt
+- die aktuelle Zuordnung eindeutig oder manuell bestätigt ist
+- die Historie sich materiell verändert hat oder das Modell als stale gilt
+
 ## Hinweise
 - Für reproduzierbare Sensor-Reihenfolgen wird in `TrainingPipeline.run(...)` eine sortierte Sensor-Liste verwendet.
 - Fehlende Trainingsdaten lösen `ValueError` aus; nicht registrierte Artefakte lösen `KeyError` aus.
-- `coverage` zählt nur exakte Sensor-Referenzen und bleibt im Bereich 0 bis 1.
+- Nur endliche numerische Werte werden trainiert.
+- `coverage` bleibt im Bereich 0 bis 1.

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "sillyhome-next"
-version = "0.1.0"
+version = "0.4.0"
 description = "Lokales Smart-Home-Intelligenzsystem für Home Assistant"
 requires-python = ">=3.11"
 dependencies = [

repository.yaml

@@ -0,0 +1,3 @@
+name: SillyHome Next Add-ons
+url: http://192.168.6.31:3000/pino/sillyhome-next
+maintainer: Pino

tests/actuators/test_actuator_store.py

@@ -0,0 +1,18 @@
+from __future__ import annotations
+
+from pathlib import Path
+
+from app.actuators.models import ReconciliationState
+from app.actuators.store import ActuatorStore
+
+
+def test_actuator_store_persists_record_and_reconciliation_state(tmp_path: Path) -> None:
+    store = ActuatorStore(tmp_path)
+    store.configure("light.abstellkammer")
+    state = ReconciliationState(last_summary="ok", configured_actuators=1)
+
+    store.save_reconciliation_state(state)
+
+    restarted = ActuatorStore(tmp_path)
+    assert restarted.get("light.abstellkammer").actuator_entity_id == "light.abstellkammer"
+    assert restarted.load_reconciliation_state().last_summary == "ok"

tests/actuators/test_lifecycle.py

@@ -0,0 +1,238 @@
+from __future__ import annotations
+
+from datetime import datetime, timedelta, timezone
+from pathlib import Path
+
+from app.actuators.lifecycle import ActuatorReconciliationService
+from app.actuators.models import (
+    AssignmentSource,
+    LifecycleStatus,
+    ManualOverride,
+    model_id_for_actuator,
+)
+from app.actuators.store import ActuatorStore
+from app.config import Settings
+from app.ha.discovery import DiscoveredEntity
+from app.ha.discovery import discover_entities
+from app.ha.history import EntityHistorySeries, NumericHistoryPoint
+from app.ha.models import HaEntitySummary
+from app.ha.reader import HaReader
+from app.ml.registry.model_registry import ModelRegistry
+
+
+class FakeActuatorReader(HaReader):
+    def __init__(
+        self,
+        entities: list[HaEntitySummary],
+        history_by_entity: dict[str, list[NumericHistoryPoint]],
+    ) -> None:
+        self._entities = entities
+        self._history_by_entity = history_by_entity
+
+    def read_entities(self) -> list[HaEntitySummary]:
+        return list(self._entities)
+
+    def discover(
+        self,
+        domains: set[str] | None = None,
+        learnable: bool | None = None,
+    ) -> list[DiscoveredEntity]:
+        return discover_entities(self._entities, domains=domains, learnable=learnable)
+
+    def read_history(
+        self,
+        entity_ids: list[str],
+        start_time: datetime,
+        end_time: datetime,
+    ) -> list[EntityHistorySeries]:
+        series: list[EntityHistorySeries] = []
+        for entity_id in entity_ids:
+            points = [
+                point
+                for point in self._history_by_entity.get(entity_id, [])
+                if start_time <= point.timestamp <= end_time
+            ]
+            if points:
+                series.append(EntityHistorySeries(entity_id=entity_id, points=points))
+        return series
+
+
+def _points(count: int, start: datetime, value: float) -> list[NumericHistoryPoint]:
+    return [
+        NumericHistoryPoint(timestamp=start + timedelta(hours=index), value=value + index)
+        for index in range(count)
+    ]
+
+
+def _service(
+    tmp_path: Path,
+    entities: list[HaEntitySummary],
+    history_by_entity: dict[str, list[NumericHistoryPoint]],
+) -> ActuatorReconciliationService:
+    return ActuatorReconciliationService(
+        ha_reader=FakeActuatorReader(entities, history_by_entity),
+        store=ActuatorStore(tmp_path / "actuators"),
+        registry=ModelRegistry(tmp_path / "models"),
+        settings=Settings(
+            ha_url="http://ha.local",
+            ha_token="token",
+            model_store=str(tmp_path / "models"),
+            automation_store=str(tmp_path / "automations"),
+            actuator_store=str(tmp_path / "actuators"),
+            history_days=14,
+            min_training_points=5,
+            retrain_stale_hours=24,
+            reconcile_interval_seconds=900,
+        ),
+    )
+
+
+def test_reconciliation_auto_assigns_and_trains_numeric_model(tmp_path: Path) -> None:
+    start = datetime(2026, 6, 1, tzinfo=timezone.utc)
+    entities = [
+        HaEntitySummary(
+            entity_id="light.abstellkammer",
+            domain="light",
+            friendly_name="Abstellkammer Licht",
+            area_name="Abstellkammer",
+        ),
+        HaEntitySummary(
+            entity_id="sensor.abstellkammer_illuminance",
+            domain="sensor",
+            device_class="illuminance",
+            state_class="measurement",
+            unit_of_measurement="lx",
+            friendly_name="Abstellkammer Helligkeit",
+            area_name="Abstellkammer",
+        ),
+        HaEntitySummary(
+            entity_id="binary_sensor.abstellkammer_motion",
+            domain="binary_sensor",
+            device_class="motion",
+            friendly_name="Abstellkammer Bewegung",
+            area_name="Abstellkammer",
+        ),
+        HaEntitySummary(
+            entity_id="sensor.kitchen_temperature",
+            domain="sensor",
+            device_class="temperature",
+            state_class="measurement",
+            unit_of_measurement="°C",
+            friendly_name="Kueche Temperatur",
+            area_name="Kueche",
+        ),
+    ]
+    service = _service(
+        tmp_path,
+        entities,
+        {
+            "sensor.abstellkammer_illuminance": _points(8, start, 10.0),
+            "sensor.kitchen_temperature": _points(8, start, 18.0),
+        },
+    )
+
+    record = service.configure_actuator("light.abstellkammer")
+
+    assert record.assignment.selected_numeric_entity_id == "sensor.abstellkammer_illuminance"
+    assert record.assignment.selected_context_entity_ids == ["binary_sensor.abstellkammer_motion"]
+    assert record.assignment.review_required is False
+    assert record.lifecycle.status is LifecycleStatus.TRAINED
+    artifact = service._registry.load_artifact(model_id_for_actuator("light.abstellkammer"))
+    assert artifact.supported_sensors == ("sensor.abstellkammer_illuminance",)
+    assert "binary_sensor.abstellkammer_motion" not in artifact.supported_sensors
+
+
+def test_reconciliation_requires_review_for_ambiguous_sensor_mapping(tmp_path: Path) -> None:
+    start = datetime(2026, 6, 1, tzinfo=timezone.utc)
+    entities = [
+        HaEntitySummary(
+            entity_id="switch.garage_pump",
+            domain="switch",
+            friendly_name="Garage Pumpe",
+            area_name="Garage",
+        ),
+        HaEntitySummary(
+            entity_id="sensor.garage_power",
+            domain="sensor",
+            device_class="power",
+            state_class="measurement",
+            unit_of_measurement="W",
+            friendly_name="Garage Leistung",
+            area_name="Garage",
+        ),
+        HaEntitySummary(
+            entity_id="sensor.garage_energy",
+            domain="sensor",
+            device_class="energy",
+            state_class="measurement",
+            unit_of_measurement="kWh",
+            friendly_name="Garage Energie",
+            area_name="Garage",
+        ),
+    ]
+    service = _service(
+        tmp_path,
+        entities,
+        {
+            "sensor.garage_power": _points(8, start, 10.0),
+            "sensor.garage_energy": _points(8, start, 11.0),
+        },
+    )
+
+    record = service.configure_actuator("switch.garage_pump")
+
+    assert record.assignment.review_required is True
+    assert record.lifecycle.status is LifecycleStatus.REVIEW_REQUIRED
+
+
+def test_manual_override_persists_and_wins_after_restart(tmp_path: Path) -> None:
+    start = datetime(2026, 6, 1, tzinfo=timezone.utc)
+    entities = [
+        HaEntitySummary(
+            entity_id="light.abstellkammer",
+            domain="light",
+            friendly_name="Abstellkammer Licht",
+            area_name="Abstellkammer",
+        ),
+        HaEntitySummary(
+            entity_id="sensor.abstellkammer_illuminance",
+            domain="sensor",
+            device_class="illuminance",
+            state_class="measurement",
+            unit_of_measurement="lx",
+            friendly_name="Abstellkammer Helligkeit",
+            area_name="Abstellkammer",
+        ),
+        HaEntitySummary(
+            entity_id="sensor.abstellkammer_power",
+            domain="sensor",
+            device_class="power",
+            state_class="measurement",
+            unit_of_measurement="W",
+            friendly_name="Abstellkammer Leistung",
+            area_name="Abstellkammer",
+        ),
+    ]
+    history = {
+        "sensor.abstellkammer_illuminance": _points(8, start, 10.0),
+        "sensor.abstellkammer_power": _points(8, start, 30.0),
+    }
+    service = _service(tmp_path, entities, history)
+    service.configure_actuator("light.abstellkammer")
+
+    updated = service.set_override(
+        "light.abstellkammer",
+        ManualOverride(
+            numeric_entity_id="sensor.abstellkammer_power",
+            context_entity_ids=[],
+            note="Manuelle Leistungs-Zuordnung",
+        ),
+    )
+
+    restarted = _service(tmp_path, entities, history)
+    record = restarted.reconcile_actuator("light.abstellkammer")
+
+    assert updated.assignment.source is AssignmentSource.MANUAL
+    assert record.assignment.selected_numeric_entity_id == "sensor.abstellkammer_power"
+    assert record.manual_override is not None
+    assert record.manual_override.numeric_entity_id == "sensor.abstellkammer_power"

tests/api/test_actuators.py

@@ -0,0 +1,135 @@
+from __future__ import annotations
+
+from datetime import datetime, timedelta
+from pathlib import Path
+
+from fastapi.testclient import TestClient
+
+from app.actuators.lifecycle import ActuatorReconciliationService
+from app.actuators.store import ActuatorStore
+from app.config import Settings
+from app.ha.discovery import DiscoveredEntity
+from app.ha.discovery import discover_entities
+from app.ha.history import EntityHistorySeries, NumericHistoryPoint
+from app.ha.models import HaEntitySummary
+from app.ha.reader import HaReader
+from app.main import app
+from app.ml.registry.model_registry import ModelRegistry
+
+
+class FakeHaReader(HaReader):
+    def __init__(self, entities: list[HaEntitySummary], history: dict[str, list[float]]) -> None:
+        self._entities = entities
+        self._history = history
+
+    def read_entities(self) -> list[HaEntitySummary]:
+        return list(self._entities)
+
+    def discover(
+        self,
+        domains: set[str] | None = None,
+        learnable: bool | None = None,
+    ) -> list[DiscoveredEntity]:
+        return discover_entities(self._entities, domains=domains, learnable=learnable)
+
+    def read_history(
+        self,
+        entity_ids: list[str],
+        start_time: datetime,
+        end_time: datetime,
+    ) -> list[EntityHistorySeries]:
+        base = start_time
+        return [
+            EntityHistorySeries(
+                entity_id=entity_id,
+                points=[
+                    NumericHistoryPoint(
+                        timestamp=base + timedelta(hours=index),
+                        value=value,
+                    )
+                    for index, value in enumerate(self._history.get(entity_id, []))
+                ],
+            )
+            for entity_id in entity_ids
+            if entity_id in self._history
+        ]
+
+
+def _install_service(tmp_path: Path) -> None:
+    entities = [
+        HaEntitySummary(
+            entity_id="light.abstellkammer",
+            domain="light",
+            friendly_name="Abstellkammer Licht",
+            area_name="Abstellkammer",
+        ),
+        HaEntitySummary(
+            entity_id="sensor.abstellkammer_illuminance",
+            domain="sensor",
+            device_class="illuminance",
+            state_class="measurement",
+            unit_of_measurement="lx",
+            friendly_name="Abstellkammer Helligkeit",
+            area_name="Abstellkammer",
+        ),
+        HaEntitySummary(
+            entity_id="binary_sensor.abstellkammer_motion",
+            domain="binary_sensor",
+            device_class="motion",
+            friendly_name="Abstellkammer Bewegung",
+            area_name="Abstellkammer",
+        ),
+    ]
+    settings = Settings(
+        ha_url="http://ha.local",
+        ha_token="token",
+        model_store=str(tmp_path / "models"),
+        automation_store=str(tmp_path / "automations"),
+        actuator_store=str(tmp_path / "actuators"),
+        history_days=14,
+        min_training_points=5,
+        retrain_stale_hours=24,
+        reconcile_interval_seconds=900,
+    )
+    app.state.registry = ModelRegistry(tmp_path / "models")
+    app.state.actuator_store = ActuatorStore(tmp_path / "actuators")
+    app.state.ha_reader = FakeHaReader(
+        entities,
+        {"sensor.abstellkammer_illuminance": [10, 11, 12, 13, 14, 15]},
+    )
+    app.state.actuator_service = ActuatorReconciliationService(
+        ha_reader=app.state.ha_reader,
+        store=app.state.actuator_store,
+        registry=app.state.registry,
+        settings=settings,
+    )
+
+
+def test_actuator_api_configures_reconciles_and_overrides(tmp_path: Path) -> None:
+    with TestClient(app) as client:
+        _install_service(tmp_path)
+
+        created = client.post("/v1/actuators", json={"actuator_entity_id": "light.abstellkammer"})
+        assert created.status_code == 201
+        assert created.json()["assignment"]["selected_numeric_entity_id"] == (
+            "sensor.abstellkammer_illuminance"
+        )
+
+        listed = client.get("/v1/actuators")
+        assert listed.status_code == 200
+        assert listed.json()[0]["lifecycle"]["status"] == "trained"
+
+        override = client.post(
+            "/v1/actuators/light.abstellkammer/override",
+            json={
+                "numeric_entity_id": "sensor.abstellkammer_illuminance",
+                "context_entity_ids": ["binary_sensor.abstellkammer_motion"],
+                "note": "Explizit bestaetigt",
+            },
+        )
+        assert override.status_code == 200
+        assert override.json()["assignment"]["source"] == "manual"
+
+        reconciliation = client.post("/v1/actuators/reconciliation/run")
+        assert reconciliation.status_code == 200
+        assert reconciliation.json()["trained_models"] == 1

tests/api/test_automations.py

@@ -0,0 +1,59 @@
+from pathlib import Path
+
+from fastapi.testclient import TestClient
+
+from app.automations.store import AutomationStore
+from app.main import app
+
+
+def _payload() -> dict[str, object]:
+    return {
+        "alias": "Licht bei Dunkelheit",
+        "description": "Schaltet das Flurlicht unter dem Helligkeitsgrenzwert ein.",
+        "trigger": {"entity_id": "sensor.hall_illuminance", "below": 10},
+        "action": {
+            "service": "light.turn_on",
+            "entity_id": "light.hall",
+            "data": {"brightness_pct": 40},
+        },
+    }
+
+
+def test_proposal_requires_explicit_approval_before_yaml(tmp_path: Path) -> None:
+    with TestClient(app) as client:
+        app.state.automation_store = AutomationStore(tmp_path)
+        created = client.post("/v1/automations/proposals", json=_payload())
+        proposal_id = created.json()["proposal_id"]
+        blocked = client.get(f"/v1/automations/proposals/{proposal_id}/yaml")
+        approved = client.post(
+            f"/v1/automations/proposals/{proposal_id}/approve",
+            json={"expected_revision": 1},
+        )
+        exported = client.get(f"/v1/automations/proposals/{proposal_id}/yaml")
+    assert created.status_code == 201
+    assert created.json()["status"] == "draft"
+    assert blocked.status_code == 409
+    assert approved.json()["status"] == "approved"
+    assert "service: light.turn_on" in exported.text
+
+
+def test_proposal_rejects_unsafe_service_domain(tmp_path: Path) -> None:
+    payload = _payload()
+    payload["action"] = {
+        "service": "shell_command.run",
+        "entity_id": "light.hall",
+        "data": {},
+    }
+    with TestClient(app) as client:
+        app.state.automation_store = AutomationStore(tmp_path)
+        response = client.post("/v1/automations/proposals", json=payload)
+    assert response.status_code == 422
+
+
+def test_proposal_requires_numeric_threshold(tmp_path: Path) -> None:
+    payload = _payload()
+    payload["trigger"] = {"entity_id": "sensor.hall_illuminance"}
+    with TestClient(app) as client:
+        app.state.automation_store = AutomationStore(tmp_path)
+        response = client.post("/v1/automations/proposals", json=payload)
+    assert response.status_code == 422

tests/api/test_entities.py

@@ -1,8 +1,11 @@
 from collections.abc import Sequence
+from datetime import datetime
 
 from fastapi.testclient import TestClient
 
 from app.ha.exceptions import HaTimeoutError
+from app.ha.discovery import DiscoveredEntity, EntityRole
+from app.ha.history import EntityHistorySeries, NumericHistoryPoint
 from app.ha.models import HaEntitySummary
 from app.ha.reader import HaReader
 from app.main import app
@@ -15,6 +18,38 @@ class FakeHaReader(HaReader):
     def read_entities(self) -> Sequence[HaEntitySummary]:
         return [HaEntitySummary(entity_id="sensor.temperature", domain="sensor")]
 
+    def discover(
+        self,
+        domains: set[str] | None = None,
+        learnable: bool | None = None,
+    ) -> Sequence[DiscoveredEntity]:
+        result = DiscoveredEntity(
+            entity_id="sensor.temperature",
+            domain="sensor",
+            device_class="temperature",
+            role=EntityRole.MEASUREMENT,
+            learnable=True,
+            reason="Numerischer Messsensor für Zeitreihen und Training.",
+        )
+        if domains and result.domain not in domains:
+            return []
+        if learnable is not None and result.learnable is not learnable:
+            return []
+        return [result]
+
+    def read_history(
+        self,
+        entity_ids: list[str],
+        start_time: datetime,
+        end_time: datetime,
+    ) -> Sequence[EntityHistorySeries]:
+        return [
+            EntityHistorySeries(
+                entity_id=entity_ids[0],
+                points=[NumericHistoryPoint(timestamp=start_time, value=21.5)],
+            )
+        ]
+
 
 class TimeoutHaReader(HaReader):
     def __init__(self) -> None:
@@ -43,6 +78,11 @@ def test_entities_returns_reader_data() -> None:
             "state_class": None,
             "device_class": None,
             "unit_of_measurement": None,
+            "friendly_name": None,
+            "area_id": None,
+            "area_name": None,
+            "device_id": None,
+            "device_name": None,
         }
     ]
 
@@ -59,3 +99,44 @@ def test_entities_maps_ha_errors_without_leaking_details() -> None:
         response = client.get("/v1/entities")
     assert response.status_code == 504
     assert response.json() == {"detail": "Home Assistant request timed out."}
+
+
+def test_discovery_filters_entities() -> None:
+    with TestClient(app) as client:
+        app.state.ha_reader = FakeHaReader()
+        response = client.get("/v1/discovery?domain=sensor&learnable=true")
+
+    assert response.status_code == 200
+    assert response.json() == [
+        {
+            "entity_id": "sensor.temperature",
+            "domain": "sensor",
+            "device_class": "temperature",
+            "state_class": None,
+            "unit_of_measurement": None,
+            "role": "measurement",
+            "learnable": True,
+            "reason": "Numerischer Messsensor für Zeitreihen und Training.",
+        }
+    ]
+
+
+def test_history_returns_normalized_series() -> None:
+    with TestClient(app) as client:
+        app.state.ha_reader = FakeHaReader()
+        response = client.get(
+            "/v1/history",
+            params=[
+                ("entity_id", "sensor.temperature"),
+                ("start_time", "2026-06-01T00:00:00Z"),
+                ("end_time", "2026-06-02T00:00:00Z"),
+            ],
+        )
+
+    assert response.status_code == 200
+    assert response.json() == [
+        {
+            "entity_id": "sensor.temperature",
+            "points": [{"timestamp": "2026-06-01T00:00:00Z", "value": 21.5}],
+        }
+    ]

tests/api/test_ml_routes.py

@@ -87,12 +87,16 @@ def test_retrain_creates_and_replaces_persisted_model(tmp_path: Path) -> None:
     assert created.json() == {
         "model_id": "home-model",
         "supported_sensors": ["sensor.kitchen"],
+        "trained_features": 1,
+        "model_type": "statistical_baseline",
         "replaced": False,
     }
     assert replaced.status_code == 200
     assert replaced.json() == {
         "model_id": "home-model",
         "supported_sensors": ["sensor.bedroom"],
+        "trained_features": 1,
+        "model_type": "statistical_baseline",
         "replaced": True,
     }
     restarted = ModelRegistry(tmp_path)
@@ -107,3 +111,74 @@ def test_retrain_rejects_empty_samples() -> None:
         )
 
     assert response.status_code == 422
+
+
+def test_predict_returns_numeric_forecast_and_confidence(tmp_path: Path) -> None:
+    from app.ml.feature_store import FeatureStore, FeatureVector
+    from app.ml.registry.model_registry import ModelRegistry
+    from app.ml.training import TrainingPipeline
+
+    store = FeatureStore()
+    store.add_batch(
+        [
+            FeatureVector("sensor.kitchen", {"temperature": 19.0}),
+            FeatureVector("sensor.kitchen", {"temperature": 20.0}),
+        ]
+    )
+    registry = ModelRegistry(tmp_path)
+    registry.register(TrainingPipeline(store).run("home-model"))
+
+    with TestClient(app) as client:
+        app.state.registry = registry
+        response = client.post(
+            "/ml/predict",
+            json={
+                "modelId": "home-model",
+                "sensor_id": "sensor.kitchen",
+                "values": {"temperature": 21.0},
+            },
+        )
+
+    assert response.status_code == 200
+    assert response.json()["predictions"] == {"temperature": 22.0}
+    assert 0.0 < response.json()["confidence"] <= 1.0
+    assert response.json()["model_type"] == "statistical_baseline"
+    explanation = response.json()["explanations"]["temperature"]
+    assert explanation["direction"] == "steigend"
+    assert explanation["change"] == 1.0
+    assert explanation["sample_count"] == 2
+
+
+def test_evaluate_returns_real_error_metrics(tmp_path: Path) -> None:
+    from app.ml.feature_store import FeatureStore, FeatureVector
+    from app.ml.registry.model_registry import ModelRegistry
+    from app.ml.training import TrainingPipeline
+
+    store = FeatureStore()
+    store.add_batch(
+        [
+            FeatureVector("sensor.kitchen", {"temperature": 19.0}),
+            FeatureVector("sensor.kitchen", {"temperature": 20.0}),
+        ]
+    )
+    registry = ModelRegistry(tmp_path)
+    registry.register(TrainingPipeline(store).run("home-model"))
+
+    with TestClient(app) as client:
+        app.state.registry = registry
+        response = client.post(
+            "/ml/evaluate",
+            json={
+                "modelId": "home-model",
+                "samples": [
+                    {
+                        "sensor_id": "sensor.kitchen",
+                        "values": {"temperature": 21.0},
+                    }
+                ],
+            },
+        )
+
+    assert response.status_code == 200
+    metrics = {metric["name"]: metric["value"] for metric in response.json()["metrics"]}
+    assert metrics == {"mae": 1.0, "rmse": 1.0, "coverage": 1.0}

tests/automations/test_store.py

@@ -0,0 +1,53 @@
+from pathlib import Path
+
+import pytest
+
+from app.automations.models import (
+    AutomationProposal,
+    NumericStateTrigger,
+    ProposalStatus,
+    ServiceAction,
+)
+from app.automations.store import AutomationStore
+
+
+def proposal() -> AutomationProposal:
+    return AutomationProposal(
+        alias="Wohnzimmer bei Kälte heizen",
+        description="Aktiviert den Heizmodus unter 18 Grad.",
+        trigger=NumericStateTrigger(entity_id="sensor.living_room_temperature", below=18.0),
+        action=ServiceAction(
+            service="climate.set_temperature",
+            entity_id="climate.living_room",
+            data={"temperature": 21.0},
+        ),
+    )
+
+
+def test_store_persists_approval_and_exports_yaml(tmp_path: Path) -> None:
+    store = AutomationStore(tmp_path)
+    created = store.create(proposal())
+    approved = store.decide(created.proposal_id, ProposalStatus.APPROVED, 1)
+    yaml = AutomationStore(tmp_path).export_yaml(created.proposal_id)
+    assert approved.status is ProposalStatus.APPROVED
+    assert approved.revision == 2
+    assert "platform: numeric_state" in yaml
+    assert "service: climate.set_temperature" in yaml
+    assert "temperature: 21.0" in yaml
+
+
+def test_store_requires_approval_and_current_revision(tmp_path: Path) -> None:
+    store = AutomationStore(tmp_path)
+    created = store.create(proposal())
+    with pytest.raises(ValueError, match="freigegebene"):
+        store.export_yaml(created.proposal_id)
+    with pytest.raises(ValueError, match="Revision"):
+        store.decide(created.proposal_id, ProposalStatus.APPROVED, 2)
+
+
+def test_store_allows_only_one_decision(tmp_path: Path) -> None:
+    store = AutomationStore(tmp_path)
+    created = store.create(proposal())
+    store.decide(created.proposal_id, ProposalStatus.REJECTED, 1)
+    with pytest.raises(ValueError, match="bereits entschieden"):
+        store.decide(created.proposal_id, ProposalStatus.APPROVED, 2)

tests/ha/test_discovery.py

@@ -0,0 +1,76 @@
+from __future__ import annotations
+
+import pytest
+
+from app.ha.discovery import EntityRole, classify_entity, discover_entities
+from app.ha.models import HaEntitySummary
+
+
+@pytest.mark.parametrize(
+    ("entity", "role", "learnable"),
+    [
+        (
+            HaEntitySummary(
+                entity_id="sensor.temperature",
+                domain="sensor",
+                device_class="temperature",
+                state_class="measurement",
+                unit_of_measurement="°C",
+            ),
+            EntityRole.MEASUREMENT,
+            True,
+        ),
+        (
+            HaEntitySummary(
+                entity_id="binary_sensor.motion",
+                domain="binary_sensor",
+                device_class="motion",
+            ),
+            EntityRole.BINARY_CONTEXT,
+            True,
+        ),
+        (
+            HaEntitySummary(entity_id="person.simon", domain="person"),
+            EntityRole.CONTEXT,
+            True,
+        ),
+        (
+            HaEntitySummary(entity_id="light.living_room", domain="light"),
+            EntityRole.ACTUATOR,
+            False,
+        ),
+        (
+            HaEntitySummary(entity_id="camera.driveway", domain="camera"),
+            EntityRole.UNSUPPORTED,
+            False,
+        ),
+    ],
+)
+def test_classify_entity(
+    entity: HaEntitySummary,
+    role: EntityRole,
+    learnable: bool,
+) -> None:
+    result = classify_entity(entity)
+    assert result.role is role
+    assert result.learnable is learnable
+
+
+def test_discovery_filters_domain_and_learnable() -> None:
+    entities = [
+        HaEntitySummary(
+            entity_id="sensor.temperature",
+            domain="sensor",
+            device_class="temperature",
+        ),
+        HaEntitySummary(entity_id="sensor.status", domain="sensor"),
+        HaEntitySummary(
+            entity_id="binary_sensor.motion",
+            domain="binary_sensor",
+            device_class="motion",
+        ),
+    ]
+
+    result = discover_entities(entities, domains={" SENSOR "}, learnable=True)
+
+    assert [item.entity_id for item in result] == ["sensor.temperature"]

tests/ha/test_ha_client.py

@@ -1,5 +1,6 @@
 from __future__ import annotations
 
+from datetime import datetime, timezone
 from unittest.mock import Mock
 
 import pytest
@@ -69,3 +70,79 @@ def test_list_entities_rejects_non_list_payload() -> None:
     client = _client_with_response(_response(payload={"entity_id": "sensor.temperature"}))
     with pytest.raises(HaUnexpectedPayloadError):
         client.list_entities()
+
+
+def test_get_history_calls_home_assistant_history_api() -> None:
+    response = _response(payload=[[{"entity_id": "sensor.temperature", "state": "21.0"}]])
+    client = _client_with_response(response)
+    start = datetime(2026, 6, 1, tzinfo=timezone.utc)
+    end = datetime(2026, 6, 2, tzinfo=timezone.utc)
+
+    payload = client.get_history(["sensor.temperature"], start, end)
+
+    assert payload == [[{"entity_id": "sensor.temperature", "state": "21.0"}]]
+    client._session.get.assert_called_once()  # type: ignore[attr-defined]
+    call = client._session.get.call_args  # type: ignore[attr-defined]
+    assert "/api/history/period/2026-06-01T00:00:00+00:00" in call.args[0]
+    assert call.kwargs["params"]["filter_entity_id"] == "sensor.temperature"
+    assert call.kwargs["params"]["end_time"] == "2026-06-02T00:00:00+00:00"
+
+
+def test_list_entity_metadata_calls_template_api() -> None:
+    response = _response()
+    response.text = (
+        '[{"entity_id":"sensor.temperature","area_name":"Kueche","device_name":"Thermometer"}]'
+    )
+    client = HaClient(HaClientSettings(url="http://ha.local", token="test-token"))
+    client._session.post = Mock(return_value=response)  # type: ignore[method-assign]
+
+    metadata = client.list_entity_metadata(["sensor.temperature"])
+
+    assert metadata == {
+        "sensor.temperature": {
+            "area_id": None,
+            "area_name": "Kueche",
+            "device_id": None,
+            "device_name": "Thermometer",
+        }
+    }
+
+
+@pytest.mark.parametrize(
+    ("entity_ids", "start", "end"),
+    [
+        (
+            [],
+            datetime(2026, 6, 1, tzinfo=timezone.utc),
+            datetime(2026, 6, 2, tzinfo=timezone.utc),
+        ),
+        (
+            ["sensor.temperature"],
+            datetime(2026, 6, 1),
+            datetime(2026, 6, 2, tzinfo=timezone.utc),
+        ),
+        (
+            ["sensor.temperature"],
+            datetime(2026, 6, 2, tzinfo=timezone.utc),
+            datetime(2026, 6, 1, tzinfo=timezone.utc),
+        ),
+        (
+            ["invalid entity"],
+            datetime(2026, 6, 1, tzinfo=timezone.utc),
+            datetime(2026, 6, 2, tzinfo=timezone.utc),
+        ),
+        (
+            ["sensor.temperature"],
+            datetime(2026, 5, 1, tzinfo=timezone.utc),
+            datetime(2026, 6, 2, tzinfo=timezone.utc),
+        ),
+    ],
+)
+def test_get_history_validates_request(
+    entity_ids: list[str],
+    start: datetime,
+    end: datetime,
+) -> None:
+    client = HaClient(HaClientSettings(url="http://ha.local", token="test-token"))
+    with pytest.raises(ValueError):
+        client.get_history(entity_ids, start, end)

tests/ha/test_ha_reader.py

@@ -1,5 +1,7 @@
 from __future__ import annotations
 
+from datetime import datetime, timezone
+
 from app.ha.client import HaClient, HaClientSettings
 from app.ha.reader import HaReader
 
@@ -26,6 +28,32 @@ class FakeHaClient(HaClient):
             },
         ]
 
+    def get_history(
+        self,
+        entity_ids: list[str],
+        start_time: datetime,
+        end_time: datetime,
+    ) -> list[object]:
+        return [
+            [
+                {
+                    "entity_id": entity_ids[0],
+                    "state": "21.5",
+                    "last_changed": start_time.isoformat(),
+                }
+            ]
+        ]
+
+    def list_entity_metadata(self, entity_ids: list[str]) -> dict[str, dict[str, str | None]]:
+        return {
+            "sensor.temperature": {
+                "area_id": "kitchen",
+                "area_name": "Kueche",
+                "device_id": "device-1",
+                "device_name": "Thermometer",
+            }
+        }
+
 
 def test_ha_reader_returns_summaries() -> None:
     reader = HaReader(FakeHaClient())
@@ -35,3 +63,27 @@ def test_ha_reader_returns_summaries() -> None:
     assert domains == {"sensor", "light"}
     sensor = next(item for item in summaries if item.entity_id == "sensor.temperature")
     assert sensor.unit_of_measurement == "°C"
+    assert sensor.area_name == "Kueche"
+    assert sensor.device_name == "Thermometer"
+
+
+def test_ha_reader_discovers_learnable_sensors() -> None:
+    reader = HaReader(FakeHaClient())
+
+    discovered = reader.discover(learnable=True)
+
+    assert [entity.entity_id for entity in discovered] == ["sensor.temperature"]
+
+
+def test_ha_reader_normalizes_history() -> None:
+    reader = HaReader(FakeHaClient())
+    start = datetime(2026, 6, 1, tzinfo=timezone.utc)
+
+    history = reader.read_history(
+        ["sensor.temperature"],
+        start,
+        datetime(2026, 6, 2, tzinfo=timezone.utc),
+    )
+
+    assert history[0].entity_id == "sensor.temperature"
+    assert history[0].points[0].value == 21.5

tests/ha/test_history.py

@@ -0,0 +1,92 @@
+from __future__ import annotations
+
+from datetime import datetime, timezone
+
+import pytest
+
+from app.ha.exceptions import HaUnexpectedPayloadError
+from app.ha.history import normalize_history_payload
+
+
+def test_normalize_history_payload_groups_and_sorts_numeric_states() -> None:
+    payload = [
+        [
+            {
+                "entity_id": "sensor.temperature",
+                "state": "22.5",
+                "last_changed": "2026-06-01T12:15:00+00:00",
+            },
+            {
+                "state": "21.0",
+                "last_changed": "2026-06-01T12:00:00Z",
+            },
+        ],
+        [
+            {
+                "entity_id": "sensor.humidity",
+                "state": 45,
+                "last_updated": "2026-06-01T12:00:00+00:00",
+            }
+        ],
+    ]
+
+    result = normalize_history_payload(payload)
+
+    assert [series.entity_id for series in result] == [
+        "sensor.humidity",
+        "sensor.temperature",
+    ]
+    temperature = result[1]
+    assert [point.value for point in temperature.points] == [21.0, 22.5]
+    assert temperature.points[0].timestamp == datetime(
+        2026, 6, 1, 12, 0, tzinfo=timezone.utc
+    )
+
+
+def test_normalize_history_payload_skips_non_numeric_and_non_finite_states() -> None:
+    payload = [
+        [
+            {
+                "entity_id": "sensor.temperature",
+                "state": state,
+                "last_changed": "2026-06-01T12:00:00+00:00",
+            }
+            for state in ("unknown", "unavailable", "nan", "inf", "-inf", True, None)
+        ]
+    ]
+
+    assert normalize_history_payload(payload) == []
+
+
+@pytest.mark.parametrize(
+    "payload",
+    [
+        {},
+        [{}],
+        [["invalid"]],
+        [[{"entity_id": "invalid", "state": "21", "last_changed": "2026-06-01"}]],
+        [[{"entity_id": "sensor.a", "state": "21", "last_changed": "invalid"}]],
+        [[{"state": "21", "last_changed": "2026-06-01T12:00:00+00:00"}]],
+        [
+            [
+                {
+                    "entity_id": "sensor.a",
+                    "state": "21",
+                    "last_changed": "2026-06-01T12:00:00+00:00",
+                },
+                {
+                    "entity_id": "sensor.b",
+                    "state": "22",
+                    "last_changed": "2026-06-01T12:01:00+00:00",
+                },
+            ]
+        ],
+    ],
+)
+def test_normalize_history_payload_rejects_malformed_structure(payload: object) -> None:
+    with pytest.raises(HaUnexpectedPayloadError):
+        normalize_history_payload(payload)
+
+
+def test_normalize_history_payload_accepts_empty_series() -> None:
+    assert normalize_history_payload([[]]) == []

tests/ml/test_evaluation.py

@@ -24,14 +24,15 @@ def test_evaluate_returns_report_with_metrics() -> None:
     report = evaluator.evaluate(
         "artifact_v1",
         [
-            "artifact_v1:sensor.kitchen:{'temperature': 21.0}",
+            _vector("sensor.kitchen", 21.0),
-            "artifact_v1:sensor.bedroom:{'temperature': 18.5}",
+            _vector("sensor.bedroom", 18.5),
         ],
     )
     assert report.artifact_id == "artifact_v1"
     assert report.sample_size == 2
-    assert {metric.name for metric in report.metrics} == {"coverage", "unknown_rate"}
+    assert {metric.name for metric in report.metrics} == {"mae", "rmse", "coverage"}
     assert next(metric.value for metric in report.metrics if metric.name == "coverage") == 1.0
+    assert next(metric.value for metric in report.metrics if metric.name == "mae") == 0.0
 
 
 def test_evaluate_without_training_raises_value_error() -> None:
@@ -40,16 +41,16 @@ def test_evaluate_without_training_raises_value_error() -> None:
         evaluator.evaluate("artifact_v1", [])
 
 
-def test_coverage_is_bounded_and_requires_exact_sensor_match() -> None:
+def test_coverage_counts_only_supported_sensor_features() -> None:
     evaluator = evaluator_factory()
     report = evaluator.evaluate(
         "artifact_v1",
         [
-            "artifact_v1:sensor.kitchen:{'note': 'sensor.bedroom'}",
+            _vector("sensor.kitchen", 21.0),
-            "artifact_v1:sensor.kitchen_extra:{}",
+            FeatureVector(sensor_id="sensor.kitchen", values={"humidity": 50.0}),
-            "malformed",
+            _vector("sensor.kitchen_extra", 20.0),
         ],
     )
 
     metrics = {metric.name: metric.value for metric in report.metrics}
-    assert metrics == {"coverage": pytest.approx(1 / 3), "unknown_rate": pytest.approx(2 / 3)}
+    assert metrics["coverage"] == pytest.approx(1 / 3)

tests/ml/test_explanation.py

@@ -0,0 +1,34 @@
+from __future__ import annotations
+
+from app.ml.explanation import explain_feature
+from app.ml.training import FeatureModel
+
+
+def _model(slope: float) -> FeatureModel:
+    return FeatureModel(
+        sample_count=4,
+        mean=20.0,
+        standard_deviation=1.0,
+        minimum=18.0,
+        maximum=22.0,
+        slope=slope,
+        intercept=18.5,
+    )
+
+
+def test_explain_feature_describes_rising_forecast() -> None:
+    explanation = explain_feature("temperature", 21.0, 21.5, _model(0.5))
+
+    assert explanation.direction == "steigend"
+    assert explanation.change == 0.5
+    assert explanation.historical_range == (18.0, 22.0)
+    assert "4 Messwerte" in explanation.summary
+    assert "Trend +0.500" in explanation.summary
+
+
+def test_explain_feature_describes_stable_and_falling_forecasts() -> None:
+    stable = explain_feature("humidity", 50.0, 50.0, _model(0.0))
+    falling = explain_feature("temperature", 21.0, 20.5, _model(-0.5))
+
+    assert stable.direction == "stabil"
+    assert falling.direction == "fallend"

tests/ml/test_model_registry.py

@@ -19,6 +19,24 @@ def test_registry_loads_persisted_artifacts_after_restart(tmp_path: Path) -> Non
     assert restarted.load_artifact("model-v1") == artifact
 
 
+def test_registry_persists_statistical_parameters(tmp_path: Path) -> None:
+    from app.ml.feature_store import FeatureStore, FeatureVector
+    from app.ml.training import TrainingPipeline
+
+    store = FeatureStore()
+    store.add_batch(
+        [
+            FeatureVector("sensor.kitchen", {"temperature": 19.0}),
+            FeatureVector("sensor.kitchen", {"temperature": 20.0}),
+        ]
+    )
+    artifact = TrainingPipeline(store).run("model-v1")
+
+    ModelRegistry(tmp_path).register(artifact)
+
+    assert ModelRegistry(tmp_path).load_artifact("model-v1") == artifact
+
+
 def test_registry_replaces_persisted_artifact_after_restart(tmp_path: Path) -> None:
     registry = ModelRegistry(tmp_path)
     registry.register(TrainedArtifact("model-v1", ("sensor.kitchen",)))

tests/ml/test_predictor.py

@@ -13,16 +13,31 @@ def _vector(sensor_id: str, temperature: float, label: str | None = None) -> Fea
 
 def predictor() -> Predictor:
     store = FeatureStore()
-    store.add_batch([_vector("sensor.kitchen", 19.0), _vector("sensor.bedroom", 18.5)])
+    store.add_batch(
+        [
+            _vector("sensor.kitchen", 19.0),
+            _vector("sensor.kitchen", 20.0),
+            _vector("sensor.bedroom", 18.5),
+        ]
+    )
     pipeline = TrainingPipeline(store)
     pipeline.run("artifact_v1")
     return Predictor(pipeline)
 
 
-def test_predict_returns_expected_format() -> None:
+def test_predict_returns_statistical_forecast() -> None:
     p = predictor()
     result = p.predict("artifact_v1", _vector("sensor.kitchen", 21.0))
-    assert result == "artifact_v1:sensor.kitchen:{'temperature': 21.0}"
+    assert result.artifact_id == "artifact_v1"
+    assert result.sensor_id == "sensor.kitchen"
+    assert result.predictions == {"temperature": 22.0}
+    assert 0.0 < result.confidence <= 1.0
+    assert result.model_type == "statistical_baseline"
+    explanation = result.explanations["temperature"]
+    assert explanation.direction == "steigend"
+    assert explanation.current_value == 21.0
+    assert explanation.predicted_value == 22.0
+    assert explanation.sample_count == 2
 
 
 def test_predict_rejects_unknown_sensor() -> None:

tests/ml/test_training.py

@@ -27,6 +27,11 @@ def test_run_returns_trained_artifact() -> None:
     artifact = pipeline.run("artifact_v1")
     assert artifact.artifact_id == "artifact_v1"
     assert artifact.supported_sensors == ("sensor.bedroom", "sensor.kitchen")
+    kitchen = artifact.feature_models["sensor.kitchen"]["temperature"]
+    assert kitchen.sample_count == 2
+    assert kitchen.mean == 19.5
+    assert kitchen.slope == 1.0
+    assert kitchen.forecast() == 21.0
 
 
 def test_run_without_data_raises_value_error() -> None:

tests/ml/test_training_evaluation.py

@@ -16,18 +16,18 @@ def test_end_to_end_training_then_evaluation() -> None:
     artifact = pipeline.run("artifact_v1")
 
     evaluator = Evaluator(pipeline)
-    predictions = [
+    samples = [
-        "artifact_v1:sensor.kitchen:{'temperature': 21.0}",
+        _vector("sensor.kitchen", 21.0),
-        "artifact_v1:sensor.bedroom:{'temperature': 18.5}",
+        _vector("sensor.bedroom", 18.5),
     ]
-    report = evaluator.evaluate(artifact.artifact_id, predictions)
+    report = evaluator.evaluate(artifact.artifact_id, samples)
     assert isinstance(report, EvalReport)
-    assert report.sample_size == len(predictions)
+    assert report.sample_size == len(samples)
     assert any(metric.name == "coverage" for metric in report.metrics)
 
 
 def test_metric_helpers_are_serializable() -> None:
-    metric = Metric(name="coverage", value=0.85, threshold=0.8)
+    metric = Metric(name="mae", value=0.85, threshold=1.0)
-    assert metric.name == "coverage"
+    assert metric.name == "mae"
     assert metric.value == 0.85
-    assert metric.threshold == 0.8
+    assert metric.threshold == 1.0

tests/test_config.py

@@ -9,10 +9,22 @@ def test_load_settings_reads_documented_environment(monkeypatch: MonkeyPatch) ->
     monkeypatch.setenv("SILLYHOME_HA_URL", "http://ha.local:8123")
     monkeypatch.setenv("SILLYHOME_HA_TOKEN", "secret")
     monkeypatch.setenv("SILLYHOME_MODEL_STORE", "/tmp/models")
+    monkeypatch.setenv("SILLYHOME_AUTOMATION_STORE", "/tmp/automations")
+    monkeypatch.setenv("SILLYHOME_ACTUATOR_STORE", "/tmp/actuators")
+    monkeypatch.setenv("SILLYHOME_HISTORY_DAYS", "7")
+    monkeypatch.setenv("SILLYHOME_MIN_TRAINING_POINTS", "12")
+    monkeypatch.setenv("SILLYHOME_RETRAIN_STALE_HOURS", "48")
+    monkeypatch.setenv("SILLYHOME_RECONCILE_INTERVAL_SECONDS", "600")
 
     settings = load_settings()
 
     assert settings.ha_url == "http://ha.local:8123"
     assert settings.ha_token == "secret"
     assert settings.model_store == "/tmp/models"
+    assert settings.automation_store == "/tmp/automations"
+    assert settings.actuator_store == "/tmp/actuators"
+    assert settings.history_days == 7
+    assert settings.min_training_points == 12
+    assert settings.retrain_stale_hours == 48
+    assert settings.reconcile_interval_seconds == 600
     assert settings.ha_configured

tests/test_dashboard.py

@@ -0,0 +1,12 @@
+from fastapi.testclient import TestClient
+
+from app.main import app
+
+
+def test_dashboard_is_served_at_root() -> None:
+    with TestClient(app) as client:
+        response = client.get("/")
+
+    assert response.status_code == 200
+    assert "SillyHome Next" in response.text
+    assert "Automation-Entwurf" in response.text