from __future__ import annotations

import json
from pathlib import Path

from config.client import extract_entity_sets, utc_now_iso
from config.settings import LOGS_DIR


IMPORTANT_KEYWORDS = [
    "документ",
    "справочник",
    "регистр",
    "плансчетов",
    "хозрасчетный",
    "контрагенты",
    "договоры",
    "организации",
    "банковскиесчета",
    "document",
    "counterparty",
    "account",
    "posting",
    "contract",
]


def mark_priority(name: str) -> tuple[bool, list[str]]:
    lowered = name.lower().replace("_", "").replace(" ", "")
    matched = [kw for kw in IMPORTANT_KEYWORDS if kw in lowered]
    return bool(matched), matched


def main() -> int:
    metadata_path = LOGS_DIR / "metadata.xml"
    if not metadata_path.exists():
        print(
            "[error] logs/metadata.xml not found. "
            "Run `python -m odata_probe.fetch_metadata` first."
        )
        return 1

    metadata_xml = metadata_path.read_text(encoding="utf-8")
    entity_sets = extract_entity_sets(metadata_xml)

    annotated: list[dict[str, object]] = []
    for item in entity_sets:
        priority, matched_keywords = mark_priority(item["name"])
        annotated.append(
            {
                **item,
                "priority": priority,
                "matched_keywords": matched_keywords,
            }
        )

    output = {
        "generated_at": utc_now_iso(),
        "total": len(annotated),
        "priority_total": sum(1 for item in annotated if item["priority"]),
        "entity_sets": annotated,
    }

    output_json = LOGS_DIR / "entity_sets_annotated.json"
    output_json.write_text(
        json.dumps(output, ensure_ascii=False, indent=2),
        encoding="utf-8",
    )

    output_txt = LOGS_DIR / "entity_sets_annotated.txt"
    lines = []
    for item in annotated:
        flag = "*" if item["priority"] else " "
        matched = ", ".join(item["matched_keywords"]) if item["matched_keywords"] else "-"
        lines.append(f"{flag} {item['name']} | {item['entity_type']} | matched: {matched}")
    output_txt.write_text("\n".join(lines), encoding="utf-8")

    print(f"[ok] total entity sets: {output['total']}")
    print(f"[ok] priority entity sets: {output['priority_total']}")
    print(f"[ok] saved: {output_json}")
    print(f"[ok] saved: {output_txt}")
    return 0


if __name__ == "__main__":
    raise SystemExit(main())