Qualifier : corrections R2/R6 + router + modèle DB + wizard frontend

Corrections moteur (TDD) :
- R2 : within_mandate → record_in_observatory=True (Observatoire des décisions)
- R6 : >50 personnes → collective recommandé, pas obligatoire (confidence=recommended)
- R3 supprimée : affected_count=1 hors périmètre de l'outil
- R9-R12 renommés G1-G4 (garde-fous internes)
- 23 tests, 213/213 verts

Étape 1 — Router /api/v1/qualify :
- POST / → qualify() avec config depuis DB ou defaults
- GET /protocol → protocole actif
- POST /protocol → créer/remplacer (auth requise)

Étape 2 — Modèle QualificationProtocol :
- Table qualification_protocols (seuils configurables via admin)
- Migration Alembic + seed du protocole par défaut

Étape 3 — Wizard frontend decisions/new.vue :
- Étape 1 : formulaire de qualification (mandat, affected_count, structurant, contexte)
- Étape 2 : résultat (type, raisons, modalités, observatoire, on-chain)
- Étape 3 : formulaire de décision (titre, description, protocole si collectif)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

backend/app/models/__init__.py

@@ -5,6 +5,7 @@ from app.models.decision import Decision, DecisionStep
 from app.models.vote import VoteSession, Vote
 from app.models.mandate import Mandate, MandateStep
 from app.models.protocol import VotingProtocol, FormulaConfig
+from app.models.qualification import QualificationProtocol
 from app.models.sanctuary import SanctuaryEntry
 from app.models.cache import BlockchainCache
 
@@ -16,6 +17,7 @@ __all__ = [
     "VoteSession", "Vote",
     "Mandate", "MandateStep",
     "VotingProtocol", "FormulaConfig",
+    "QualificationProtocol",
     "SanctuaryEntry",
     "BlockchainCache",
 ]

backend/app/models/qualification.py

@@ -0,0 +1,38 @@
+import json
+import uuid
+from datetime import datetime
+
+from sqlalchemy import Boolean, DateTime, Integer, String, Text, Uuid, func
+from sqlalchemy.orm import Mapped, mapped_column
+
+from app.database import Base
+
+
+class QualificationProtocol(Base):
+    """Active configuration for the decision qualification engine.
+
+    Thresholds stored here override the engine defaults and can be updated
+    through the admin interface (meta-governance).
+    Only one record should be active at a time (is_active=True).
+    """
+    __tablename__ = "qualification_protocols"
+
+    id: Mapped[uuid.UUID] = mapped_column(Uuid, primary_key=True, default=uuid.uuid4)
+    name: Mapped[str] = mapped_column(String(128), nullable=False)
+    description: Mapped[str | None] = mapped_column(Text)
+
+    small_group_max: Mapped[int] = mapped_column(Integer, default=5)
+    collective_wot_min: Mapped[int] = mapped_column(Integer, default=50)
+
+    # JSON array of modality slugs, e.g. '["vote_wot","vote_smith","election"]'
+    default_modalities_json: Mapped[str] = mapped_column(
+        Text,
+        default='["vote_wot","vote_smith","consultation_avis","election"]',
+    )
+
+    is_active: Mapped[bool] = mapped_column(Boolean, default=True)
+    created_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), server_default=func.now())
+
+    @property
+    def default_modalities(self) -> list[str]:
+        return json.loads(self.default_modalities_json)