import json
import sys
from pathlib import Path


KEY_CONTRACT_TERMS = {
    "qpi": [
        "problem_owner",
        "time_scale",
        "dominant_scarcity",
        "evidence_gap",
        "classification",
        "no_call"
    ],
    "intellectual_archaeology": [
        "should_call",
        "recommended_max_depth",
        "stop_reason",
        "no_deeper_reason",
        "validation_needed",
        "action_implication"
    ]
}


def read_json(path):
    return json.loads(path.read_text(encoding="utf-8"))


def card_path_for_model(root, model_id):
    return root / "cards" / f"{model_id}.md"


def check_model_card(root, model_path):
    model = read_json(model_path)
    model_id = model["model_id"]
    card_path = card_path_for_model(root, model_id)
    errors = []
    manual_review = []

    if not card_path.exists():
        return [f"missing card {card_path.relative_to(root).as_posix()}"], []

    text = card_path.read_text(encoding="utf-8")
    hard_values = [
        model_id,
        model["model_name"],
        model["model_type"],
        model["pipeline_position"],
        model["regression_status"],
        model["version"],
        model["last_updated"],
        model.get("status", "")
    ]
    stability_level = model.get("stability_profile", {}).get("stability_level")
    if stability_level:
        hard_values.append(stability_level)

    for value in hard_values:
        if value and value not in text:
            errors.append(f"{model_id} card missing hard value {value}")

    for source_id in model.get("source_articles", []):
        if source_id not in text:
            errors.append(f"{model_id} card missing source article {source_id}")

    for excerpt_id in model.get("source_evidence", []):
        if excerpt_id not in text:
            errors.append(f"{model_id} card missing source evidence {excerpt_id}")

    for term in KEY_CONTRACT_TERMS.get(model_id, []):
        if term not in text:
            errors.append(f"{model_id} card missing key contract term {term}")

    manual_review.append(f"{model_id}: long Chinese definitions, examples, and risk notes require human semantic review.")
    return errors, manual_review


def write_report(root, errors, manual_review):
    report_path = root / "reports" / "model_card_sync_report_v0.2.md"
    lines = [
        "# Model Card Sync Report v0.2",
        "",
        f"Status: `{'PASS' if not errors else 'FAIL'}`",
        "",
    ]
    if errors:
        lines.append("## Errors")
        lines.append("")
        lines.extend(f"- {error}" for error in errors)
        lines.append("")
    else:
        lines.append("## Hard-Field Result")
        lines.append("")
        lines.append("- Model IDs, names, types, pipeline positions, source IDs, evidence IDs, stability fields, regression status, version, and last_updated are present in Markdown cards.")
        lines.append("- Key output contract and depth-control terms are present.")
        lines.append("")

    lines.append("## Manual Review Items")
    lines.append("")
    lines.extend(f"- {item}" for item in manual_review)
    report_path.write_text("\n".join(lines) + "\n", encoding="utf-8")
    return report_path


def check_sync(root):
    root = Path(root)
    errors = []
    manual_review = []
    for model_path in sorted((root / "models").glob("*.model.json")):
        model_errors, model_manual_review = check_model_card(root, model_path)
        errors.extend(model_errors)
        manual_review.extend(model_manual_review)
    report_path = write_report(root, errors, manual_review)
    return errors, report_path


def main():
    root = Path(__file__).resolve().parents[1]
    errors, report_path = check_sync(root)
    print(f"model/card sync report written to {report_path}")
    if errors:
        for error in errors:
            print(f"ERROR: {error}")
        return 1
    print("model/card sync passed")
    return 0


if __name__ == "__main__":
    sys.exit(main())