"""解析模型返回为演化候选草稿（纯函数，无 IO）。

模型可能返回纯 JSON、markdown 代码块或夹杂解释文本；本模块容错提取首个 JSON
数组，并归一化为演化候选草稿。``evidence`` 序号由 application 映射为记忆 ID。

缺省字段取**安全悲观默认**：``level=L4``（必须人工确认）、``auto_apply_allowed=False``，
避免模型漏字段时误自动应用高风险演化。
"""

import json
import re
from dataclasses import dataclass, field
from typing import Any


@dataclass
class EvolutionCandidateDraft:
    type: str
    title: str
    proposal: str
    level: str = "L4"
    scope: str = "character_scope"
    source: str = "big_cycle"
    delta: dict[str, Any] = field(default_factory=dict)
    confidence: float = 0.5
    benefit_score: float = 0.5
    risk_score: float = 0.5
    drift_score: float = 0.0
    auto_apply_allowed: bool = False
    evidence: list[int] = field(default_factory=list)


def parse_evolution_candidates(raw: str) -> list[EvolutionCandidateDraft]:
    drafts: list[EvolutionCandidateDraft] = []
    for item in _extract_json_array(raw):
        if not isinstance(item, dict):
            continue
        type_ = str(item.get("type", "")).strip()
        title = str(item.get("title", "")).strip()
        proposal = str(item.get("proposal", "")).strip()
        if not type_ or not title or not proposal:
            continue
        drafts.append(
            EvolutionCandidateDraft(
                type=type_,
                title=title,
                proposal=proposal,
                level=_level(item.get("level")),
                scope=_scope(item.get("scope")),
                source=str(item.get("source", "big_cycle")).strip() or "big_cycle",
                delta=_dict(item.get("delta")),
                confidence=_clamp_float(item.get("confidence"), 0.5),
                benefit_score=_clamp_float(item.get("benefit_score"), 0.5),
                risk_score=_clamp_float(item.get("risk_score"), 0.5),
                drift_score=_clamp_float(item.get("drift_score"), 0.0),
                auto_apply_allowed=bool(item.get("auto_apply_allowed", False)),
                evidence=_int_list(item.get("evidence")),
            )
        )
    return drafts


def _extract_json_array(raw: str) -> list[Any]:
    text = raw.strip()
    matched = re.search(r"```(?:json)?\s*(\[.*?\])\s*```", text, re.DOTALL)
    if matched:
        text = matched.group(1)
    else:
        start, end = text.find("["), text.rfind("]")
        if start != -1 and end > start:
            text = text[start : end + 1]
    try:
        data = json.loads(text)
    except json.JSONDecodeError:
        return []
    return data if isinstance(data, list) else []


_VALID_LEVELS = {"L0", "L1", "L2", "L3", "L4"}
_VALID_SCOPES = {"character_scope", "user_scope", "task_scope", "global_scope"}


def _level(value: Any) -> str:
    text = str(value or "").strip().upper()
    return text if text in _VALID_LEVELS else "L4"


def _scope(value: Any) -> str:
    text = str(value or "").strip()
    return text if text in _VALID_SCOPES else "character_scope"


def _dict(value: Any) -> dict[str, Any]:
    return value if isinstance(value, dict) else {}


def _clamp_float(value: Any, default: float) -> float:
    try:
        num = float(value)
    except (TypeError, ValueError):
        return default
    return max(0.0, min(1.0, num))


def _int_list(value: Any) -> list[int]:
    if not isinstance(value, list):
        return []
    result: list[int] = []
    for item in value:
        try:
            result.append(int(item))
        except (TypeError, ValueError):
            continue
    return result