"""Per-dimension evidence trust policy for taxonomy consumers."""

from __future__ import annotations

import hashlib
import json
from typing import Any


SOURCE_TRUST_POLICY: list[dict[str, Any]] = [
    {
        "dimension": "genre_family",
        "trusted": ["groundtruth high/medium", "CLAP high score with MAEST/metadata support"],
        "candidate_only": ["CLAP alone", "MAEST Discogs neighbor"],
        "rule": "Use CLAP for Turkish-native genre only when corroborated; use MAEST as a Western crosswalk.",
    },
    {
        "dimension": "genre_subtype",
        "trusted": ["groundtruth high/medium", "deterministic title cue"],
        "candidate_only": ["CLAP subtype prompts", "MAEST hints"],
        "rule": "Do not present subtype as definitive from zero-shot audio alone.",
    },
    {
        "dimension": "instrumentation",
        "trusted": ["groundtruth credits", "stem-window high confidence"],
        "candidate_only": ["CLAP instrument tags", "stem-window medium/low confidence"],
        "rule": "Mention specific instruments only when GT/stem evidence supports them.",
    },
    {
        "dimension": "vocal_configuration",
        "trusted": ["groundtruth/review labels", "clear stem vocal presence"],
        "candidate_only": ["CLAP vocal tags"],
        "rule": "Use CLAP vocal configuration as a starting hypothesis for review, not as gold.",
    },
    {
        "dimension": "vocal_technique",
        "trusted": ["groundtruth/review labels"],
        "candidate_only": ["CLAP technique prompts"],
        "rule": "Do not assert hançere, gazel, or taverna technique unless reviewed or source-supported.",
    },
    {
        "dimension": "mood",
        "trusted": ["lyrics/themes", "review labels"],
        "candidate_only": ["CLAP mood tags"],
        "rule": "Let lyrics dominate mood when available; CLAP mood is weak cultural evidence.",
    },
    {
        "dimension": "arrangement_aesthetic",
        "trusted": ["groundtruth/review labels", "source-supported recording identity"],
        "candidate_only": ["CLAP genre/instrument mixture"],
        "rule": "Never infer production arrangement solely from CLAP.",
    },
    {
        "dimension": "recording_context",
        "trusted": ["groundtruth/review labels", "audible/source evidence"],
        "candidate_only": ["CLAP/stem context hints"],
        "rule": "Live, broadcast, lo-fi, and studio context need explicit evidence.",
    },
]


def source_policy_hash() -> str:
    payload = json.dumps(SOURCE_TRUST_POLICY, sort_keys=True, ensure_ascii=False)
    return hashlib.sha256(payload.encode("utf-8")).hexdigest()[:12]


def format_source_policy_for_prompt() -> str:
    rows: list[str] = []
    for item in SOURCE_TRUST_POLICY:
        rows.append(
            "- {dimension}: trusted={trusted}; candidate_only={candidate_only}; rule={rule}".format(
                dimension=item["dimension"],
                trusted=", ".join(item["trusted"]),
                candidate_only=", ".join(item["candidate_only"]),
                rule=item["rule"],
            )
        )
    return "\n".join(rows)