"""validate_llm_determinism_pack_v1.py — spec/58: H008_LLM_DETERMINISM_AUDIT

Validates that final_context_for_llm_v5.yaml contains all required
pre-computed sections and that no section demands arithmetic from the LLM.

formula_id: VALIDATE_LLM_DETERMINISM_PACK_V1
contract:   spec/58_llm_determinism_contract.yaml
"""
from __future__ import annotations

import json
import re
import sys
from pathlib import Path

import yaml

ROOT = Path(__file__).resolve().parents[1]

DEFAULT_CONTEXT = ROOT / "Temp" / "final_context_for_llm_v5.yaml"
OUTPUT_PATH     = ROOT / "Temp" / "llm_determinism_pack_v1.json"

# Required sections from spec/58
REQUIRED_SECTIONS = [
    "01_metadata_and_manifest_alias",
    "02_portfolio_health",
    "03_hard_blockers",
    "04_sell_priority_table",
    "05_buy_hold_sell_action_table",
    "06_cash_and_risk_budget",
    "07_shadow_ledger_visible_items",
    "08_data_missing_items",
    "09_market_regime_summary_precomputed",
    "10_education_notes_preapproved",
    "11_forbidden_phrases_and_no_math_rules",
]

# Patterns that indicate arithmetic instructions to LLM
ARITHMETIC_INSTRUCTION_PATTERNS = [
    r"계산\s*하시오",
    r"계산\s*해\s*주",
    r"더해\s*서",
    r"나누어",
    r"빼\s*면",
    r"평균\s*구하",
    r"합계\s*구하",
    r"계산.*결과를\s*출력",
    r"LLM.*계산",
    r"\bcompute\b.*\bprice\b",
    r"\bcalculate\b.*\bquantity\b",
]

# Numeric fields that must be pre-filled (not left for LLM)
REQUIRED_NUMERIC_FIELDS = [
    "total_asset_krw",
    "cash_ratio_pct",
    "goal_achievement_pct",
    "available_cash_krw",
    "max_allowed_mdd_pct",
]


def _load_yaml(path: Path) -> dict:
    if not path.exists():
        return {"_missing": True, "_path": str(path)}
    try:
        obj = yaml.safe_load(path.read_text(encoding="utf-8"))
        return obj if isinstance(obj, dict) else {"_empty": True}
    except Exception as e:
        return {"_error": str(e), "_path": str(path)}


def _check_required_sections(context: dict) -> tuple[list[str], list[str]]:
    """Return (found_sections, missing_sections)."""
    context_text = str(context)
    found, missing = [], []
    for sec in REQUIRED_SECTIONS:
        if sec in context_text:
            found.append(sec)
        else:
            missing.append(sec)
    return found, missing


def _count_arithmetic_instructions(context: dict) -> tuple[int, list[str]]:
    context_text = json.dumps(context, ensure_ascii=False)
    findings = []
    for pattern in ARITHMETIC_INSTRUCTION_PATTERNS:
        if re.search(pattern, context_text, re.IGNORECASE):
            findings.append(pattern)
    return len(findings), findings


def _check_numeric_fields_precomputed(context: dict) -> tuple[float, list[str]]:
    """Check that required numeric fields have actual values (not placeholders)."""
    context_text = json.dumps(context, ensure_ascii=False)
    unfilled = []
    for field in REQUIRED_NUMERIC_FIELDS:
        # Look for field = null / field = "" / field = "DATA_MISSING"
        null_pattern = rf'"{field}"\s*:\s*(null|""|"DATA_MISSING")'
        if re.search(null_pattern, context_text):
            unfilled.append(field)
        elif field not in context_text:
            unfilled.append(field)
    filled = len(REQUIRED_NUMERIC_FIELDS) - len(unfilled)
    coverage_pct = 100.0 * filled / len(REQUIRED_NUMERIC_FIELDS) if REQUIRED_NUMERIC_FIELDS else 100.0
    return coverage_pct, unfilled


def _check_llm_numeric_generation(context: dict) -> int:
    """Count fields that ask LLM to generate a number."""
    context_text = json.dumps(context, ensure_ascii=False)
    generation_patterns = [
        r"최종\s*수량\s*산출",
        r"손절가\s*계산",
        r"익절가\s*계산",
        r"LLM.*숫자.*생성",
    ]
    count = 0
    for p in generation_patterns:
        count += len(re.findall(p, context_text, re.IGNORECASE))
    return count


def run(context_path: Path) -> dict:
    context = _load_yaml(context_path)

    if context.get("_missing"):
        result = {
            "gate": "SKIP",
            "reason": f"context file missing: {context_path}",
            "missing_sections": [],
            "arithmetic_instruction_count": 0,
            "precomputed_field_coverage_pct": 0.0,
            "llm_numeric_generation_count": 0,
            "contract": "spec/58_llm_determinism_contract.yaml",
        }
        OUTPUT_PATH.parent.mkdir(parents=True, exist_ok=True)
        OUTPUT_PATH.write_text(json.dumps(result, ensure_ascii=False, indent=2))
        return result

    found_sections, missing_sections = _check_required_sections(context)
    arith_count, arith_patterns     = _count_arithmetic_instructions(context)
    coverage_pct, unfilled_fields   = _check_numeric_fields_precomputed(context)
    llm_gen_count                   = _check_llm_numeric_generation(context)

    gate = "PASS"
    if missing_sections or arith_count > 0 or llm_gen_count > 0:
        gate = "FAIL"
    elif coverage_pct < 80.0:
        gate = "WARN"

    result = {
        "gate": gate,
        "found_sections": found_sections,
        "missing_sections": missing_sections,
        "arithmetic_instruction_count": arith_count,
        "arithmetic_instruction_patterns": arith_patterns,
        "precomputed_field_coverage_pct": round(coverage_pct, 2),
        "unfilled_required_fields": unfilled_fields,
        "llm_numeric_generation_count": llm_gen_count,
        "sections_required": len(REQUIRED_SECTIONS),
        "sections_found": len(found_sections),
        "contract": "spec/58_llm_determinism_contract.yaml",
    }

    OUTPUT_PATH.parent.mkdir(parents=True, exist_ok=True)
    OUTPUT_PATH.write_text(json.dumps(result, ensure_ascii=False, indent=2))
    return result


def main() -> None:
    import argparse
    parser = argparse.ArgumentParser(description="H008 LLM Determinism Pack Validator")
    parser.add_argument("--context", default=str(DEFAULT_CONTEXT))
    args = parser.parse_args()

    result = run(Path(args.context))
    gate = result.get("gate", "FAIL")
    print(f"[H008_LLM_DETERMINISM_PACK] gate={gate} "
          f"sections={result.get('sections_found', 0)}/{result.get('sections_required', 0)} "
          f"arithmetic={result.get('arithmetic_instruction_count', 0)} "
          f"field_coverage={result.get('precomputed_field_coverage_pct', 0):.1f}%")
    if gate == "FAIL":
        print("  Missing sections:", result.get("missing_sections"))
        print("  Arithmetic patterns:", result.get("arithmetic_instruction_patterns"))
        sys.exit(1)


if __name__ == "__main__":
    main()