from datetime import datetime, timezone
from functools import lru_cache
from pathlib import Path

import pytest
from sqlalchemy import create_engine
from sqlalchemy.orm import Session

from central.app.datasets.nfe_moc_tables import parse_moc_tables
from central.app.datasets.service import DatasetService
from central.app.models import Base

SOURCE = Path(__file__).parents[2] / "sources" / "raw" / "bootstrap_v1" / "moc7-anexo-i-leiaute-e-rv.pdf"
pytestmark = pytest.mark.skipif(not SOURCE.exists(), reason="official MOC PDF not present locally")


@lru_cache(maxsize=1)
def _payload():
    # Parsing this 153-page PDF is expensive (~40s); every test in this module
    # shares one parse instead of re-running it per test.
    return parse_moc_tables(SOURCE.read_bytes())


def test_parses_all_five_domains_with_expected_minimum_coverage():
    payload = _payload()
    from collections import Counter
    counts = Counter(e.domain for e in payload.entries)
    assert counts["ORIGEM_MERCADORIA"] == 9
    assert counts["CST_ICMS"] >= 11
    assert counts["CSOSN"] >= 10
    assert counts["CST_PIS"] >= 29
    assert counts["CST_COFINS"] >= 29


def test_every_entry_has_provenance_and_no_invented_text():
    payload = _payload()
    for entry in payload.entries:
        assert entry.source_section.startswith("MOC7 p.")
        assert entry.description.strip()
        assert entry.code.strip()


def test_origem_codes_are_0_through_8():
    payload = _payload()
    codes = sorted(e.code for e in payload.entries if e.domain == "ORIGEM_MERCADORIA")
    assert codes == [str(n) for n in range(9)]


def test_cst_icms_includes_split_group_40_41_50():
    payload = _payload()
    by_code = {e.code: e for e in payload.entries if e.domain == "CST_ICMS"}
    assert "isenta" in by_code["40"].description.lower()
    assert "tributada" in by_code["41"].description.lower()
    assert "suspens" in by_code["50"].description.lower()


def test_reimport_is_idempotent():
    payload = _payload()
    engine = create_engine("sqlite://")
    Base.metadata.create_all(engine)
    with Session(engine) as session:
        service = DatasetService(session)
        first = service.import_moc_tables(payload, "file://moc7.pdf", datetime.now(timezone.utc))
        second = service.import_moc_tables(payload, "file://moc7.pdf", datetime.now(timezone.utc))
        assert first.id == second.id
        assert second.status == "ACTIVE"


def test_lookup_smoke_for_each_domain():
    payload = _payload()
    engine = create_engine("sqlite://")
    Base.metadata.create_all(engine)
    with Session(engine) as session:
        service = DatasetService(session)
        service.import_moc_tables(payload, "file://moc7.pdf", datetime.now(timezone.utc))

        assert len(service.fiscal_codes("ORIGEM_MERCADORIA")) == 9
        assert service.fiscal_codes("ORIGEM_MERCADORIA", code="0")[0].code == "0"
        assert service.fiscal_codes("CST_ICMS", code="00")
        assert service.fiscal_codes("CSOSN", code="101")
        assert service.fiscal_codes("CST_PIS", code="01")
        assert service.fiscal_codes("CST_COFINS", code="01")
        assert service.fiscal_codes("CST_ICMS", code="99") == []


def test_dataset_versions_includes_nfe_moc_tables_and_changes_hash_on_reimport():
    engine = create_engine("sqlite://")
    Base.metadata.create_all(engine)
    with Session(engine) as session:
        service = DatasetService(session)
        assert service.dataset_versions()["nfe_moc_tables"] is None
        service.import_moc_tables(_payload(), "file://moc7.pdf", datetime.now(timezone.utc))
        assert service.dataset_versions()["nfe_moc_tables"] is not None