from schemas.models import ConsensusRequest, ConsensusResult


class ConsensusEngine:
    """Calculates weighted composite consensus EV across 7 key intelligence vectors."""

    @staticmethod
    def calculate_consensus(req: ConsensusRequest) -> ConsensusResult:
        weights = {
            "evidence": 0.25,
            "agreement": 0.20,
            "source_quality": 0.15,
            "historical_accuracy": 0.15,
            "memory_similarity": 0.10,
            "model_confidence": 0.10,
            "risk_penalty": 0.05,
        }

        weighted_score = (
            (req.evidence_confidence * weights["evidence"])
            + (req.agreement_score * weights["agreement"])
            + (req.source_quality * weights["source_quality"])
            + (req.historical_accuracy * weights["historical_accuracy"])
            + (req.memory_similarity * weights["memory_similarity"])
            + (req.model_confidence * weights["model_confidence"])
            - (req.mission_risk * weights["risk_penalty"])
        )

        composite = round(max(0.0, min(100.0, weighted_score)), 2)

        if composite >= 85.0:
            rec = "HIGH_CONFIDENCE_EXECUTE"
            tier = "TIER_1_OPTIMAL"
            risk_desc = "Low operational risk; verified across independent evidence channels."
        elif composite >= 65.0:
            rec = "PROCEED_WITH_VERIFICATION"
            tier = "TIER_2_MODERATE"
            risk_desc = "Moderate confidence; minor conflicts or unverified secondary claims."
        else:
            rec = "REQUIRES_HUMAN_REVIEW_OR_DEEPER_SEARCH"
            tier = "TIER_3_ELEVATED_RISK"
            risk_desc = "Elevated risk; high conflict score or low source authority."

        return ConsensusResult(
            mission_id=req.mission_id,
            topic=req.topic,
            composite_consensus_score=composite,
            decision_recommendation=rec,
            risk_assessment=risk_desc,
            confidence_tier=tier,
        )