feat(stage7): claim clustering & contradiction candidates — semantic grouping, numeric normalization, pairwise analysis

- ClaimClusterModel: LLM-basierte semantische Gruppierung von Claims
- ClaimRelationModel: SUPPORTS, CONTRADICTS, DUPLICATE, UNCERTAIN pairwise relations
- ClaimNLUModel: numerische Normalisierung (%, Währungen, deutsche/englische Wörter)
- stage7_normalize_numerics.py: Regex-basiert mit 200+ deutschen/englischen Zahlenwörtern
- stage7_clustering.py: LLM-Clustering + pairwise claim-relation analysis
- API: POST cluster-claims, GET clusters, GET claim-relations
- 79 tests: numerische Normalisierung, LLM-Parsing, Clustering, Relationen, Edge-Cases
- Dedup: claims mit gleichen numerischen Werten werden zusammengefasst

src/nsct/storage/models.py

@@ -273,4 +273,110 @@ class ResearchReportModel(Base):
 
     __table_args__ = (
         Index("ix_research_reports_research_id", "research_id"),
+    )
+
+
+# ---------------------------------------------------------------------------
+# Stage 7 — Claim Clustering & Contradiction Candidates
+# ---------------------------------------------------------------------------
+
+
+class ClaimRelationType(str, enum.Enum):
+    """Relationship types between individual claims (Stage 7)."""
+
+    SUPPORTS = "supports"
+    CONTRADICTS = "contradicts"
+    DUPLICATE = "duplicate"
+    UNCERTAIN = "uncertain"
+
+
+class ClaimClusterModel(Base):
+    """Semantisches Cluster von Claims innerhalb eines Research-Runs."""
+
+    __tablename__ = "claim_clusters"
+
+    id = Column(String(36), primary_key=True, default=lambda: str(uuid4()))
+    research_run_id = Column(String(36), nullable=False, default=lambda: str(uuid4()))
+    cluster_label = Column(Text, nullable=False)
+    representative_claim_id = Column(String(36), ForeignKey("claims.id"), nullable=True)
+    claim_count = Column(Integer, nullable=False, default=1)
+    created_at = Column(DateTime, nullable=False, default=datetime.utcnow)
+
+    # Relationships
+    claims = relationship("ClaimModel", secondary="claim_cluster_mapping", back_populates="clusters")
+    relations = relationship(
+        "ClaimRelationModel",
+        back_populates="cluster",
+        cascade="all, delete-orphan",
+    )
+
+    __table_args__ = (
+        Index("ix_claim_clusters_research_run_id", "research_run_id"),
+    )
+
+
+# Association table: claims ↔ clusters (many-to-many)
+_claim_cluster_mapping = Base()  # noqa: F811  -- dummy for type resolution
+
+claim_cluster_mapping = Base()
+claim_cluster_mapping.__tablename__ = "claim_cluster_mapping"
+claim_cluster_mapping.id = Column(String(36), primary_key=True, default=lambda: str(uuid4()))
+claim_cluster_mapping.cluster_id = Column(String(36), ForeignKey("claim_clusters.id"), nullable=False)
+claim_cluster_mapping.claim_id = Column(String(36), ForeignKey("claims.id"), nullable=False)
+claim_cluster_mapping.position = Column(Integer, nullable=False, default=0)
+
+__table_args__ = (
+    Index("ix_claim_cluster_mapping_cluster_id", "cluster_id"),
+    Index("ix_claim_cluster_mapping_claim_id", "claim_id"),
+    Index("uq_claim_cluster_mapping_cluster_claim", "cluster_id", "claim_id", unique=True),
+)
+
+
+class ClaimRelationModel(Base):
+    """Pairwise Beziehung zwischen zwei Claims innerhalb eines Clusters."""
+
+    __tablename__ = "claim_relations"
+
+    id = Column(String(36), primary_key=True, default=lambda: str(uuid4()))
+    source_claim_id = Column(String(36), ForeignKey("claims.id"), nullable=False)
+    target_claim_id = Column(String(36), ForeignKey("claims.id"), nullable=False)
+    relation_type = Column(Enum(ClaimRelationType), nullable=False)
+    confidence = Column(Float, nullable=False, default=0.5)
+    reason = Column(Text, nullable=True)
+    cluster_id = Column(String(36), ForeignKey("claim_clusters.id"), nullable=True)
+    created_at = Column(DateTime, nullable=False, default=datetime.utcnow)
+
+    # Relationships
+    cluster = relationship("ClaimClusterModel", back_populates="relations")
+    source_claim = relationship("ClaimModel", foreign_keys=[source_claim_id])
+    target_claim = relationship("ClaimModel", foreign_keys=[target_claim_id])
+
+    __table_args__ = (
+        Index("ix_claim_relations_source_claim_id", "source_claim_id"),
+        Index("ix_claim_relations_target_claim_id", "target_claim_id"),
+        Index("ix_claim_relations_cluster_id", "cluster_id"),
+        Index(
+            "uq_claim_relations_pair",
+            "source_claim_id",
+            "target_claim_id",
+            "relation_type",
+            unique=True,
+        ),
+    )
+
+
+class ClaimNLUModel(Base):
+    """Numerisch normalisierte Ausdrücke aus Claims (Stage 7 — NLU)."""
+
+    __tablename__ = "claim_nlu_numeric"
+
+    id = Column(String(36), primary_key=True, default=lambda: str(uuid4()))
+    claim_id = Column(String(36), ForeignKey("claims.id"), nullable=False)
+    normalized_numeric_value = Column(String(64), nullable=True)
+    original_numeric_text = Column(Text, nullable=False)
+    unit = Column(String(16), nullable=True)
+    created_at = Column(DateTime, nullable=False, default=datetime.utcnow)
+
+    __table_args__ = (
+        Index("ix_claim_nlu_numeric_claim_id", "claim_id"),
     )