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 import bootstrap_v1_extra as parsers
from central.app.datasets.service import DatasetService
from central.app.models import Base

ROOT = Path(__file__).parents[2] / "sources" / "raw" / "bootstrap_v1"
FILES = {
    "cfop": ROOT / "IT_2023_002_v2_10_Tabela_CFOP_Descrição.xlsx",
    "fcp": ROOT / "Tabela_Alíquotas_FCP_por_UF_2025-02_07.xlsx",
    "anp": ROOT / "Tabela_Codigo_Produto_ANP-20241118.xlsx",
    "ncm_utrib": ROOT / "IT2024.001v2.40 - Tabela NCM 2022 com Utrib_Comércio Exterior_vigência 01.10.26.xlsx",
    "biofuel": ROOT / "IT 2025.004 v.1.20 - Tabela de Índice.xlsx",
    "credpres": ROOT / "cCredPres_2026-06-22.xlsx",
}
pytestmark = pytest.mark.skipif(not all(p.exists() for p in FILES.values()), reason="bootstrap_v1 extra source files not present locally")


@lru_cache(maxsize=None)
def _payload(name):
    parse_fn = {"cfop": parsers.parse_cfop, "fcp": parsers.parse_fcp, "anp": parsers.parse_anp,
                "ncm_utrib": parsers.parse_ncm_utrib, "biofuel": parsers.parse_rtc_biofuel_index,
                "credpres": parsers.parse_rtc_cred_pres}[name]
    return parse_fn(FILES[name].read_bytes())


def _engine():
    engine = create_engine("sqlite://")
    Base.metadata.create_all(engine)
    return engine


def test_cfop_parses_full_table_with_official_fields():
    payload = _payload("cfop")
    assert len(payload.entries) >= 500
    sample = payload.entries[0]
    assert set(sample) >= {"code", "title", "description", "effective_from", "ind_nfe", "ind_comb"}


def test_fcp_covers_all_states_and_df():
    payload = _payload("fcp")
    ufs = {e["uf"] for e in payload.entries}
    assert len(ufs) == 27  # 26 states + DF
    assert "BA" in ufs


def test_anp_has_no_invented_descriptions():
    payload = _payload("anp")
    assert len(payload.entries) >= 900
    assert all(set(e) == {"cprod_anp"} for e in payload.entries)  # source truly has no description column


def test_ncm_utrib_covers_thousands_of_ncms():
    payload = _payload("ncm_utrib")
    assert len(payload.entries) >= 10000
    assert all(len(e["ncm"]) == 8 for e in payload.entries[:50])


def test_biofuel_index_separates_current_from_historical():
    payload = _payload("biofuel")
    current = [e for e in payload.entries if not e["is_historical"]]
    historical = [e for e in payload.entries if e["is_historical"]]
    assert current and historical
    assert all(e["p_bio_percent"] is not None for e in current)


def test_credpres_preserves_rich_official_fields_not_just_code_description():
    payload = _payload("credpres")
    sample = payload.entries[0]
    assert sample["legal_basis"]
    assert "d_ini_vig_cbs" in sample and "aliquota_cbs_text" in sample


@pytest.mark.parametrize("domain,method,lookup", [
    ("cfop", "import_cfop", lambda s: s.cfop_lookup(code="1101")),
    ("fcp", "import_fcp", lambda s: [s.fcp_for_uf("BA")]),
    ("anp", "import_anp", lambda s: [s.anp_code_exists("110203073")]),
    ("ncm_utrib", "import_ncm_utrib", lambda s: [s.ncm_utrib_for_ncm("01012100")]),
    ("biofuel", "import_rtc_biofuel_index", lambda s: s.rtc_biofuel_for_anp("320102001")),
    ("credpres", "import_rtc_cred_pres", lambda s: s.rtc_cred_pres(code="1")),
])
def test_import_idempotent_and_lookup_smoke(domain, method, lookup):
    engine = _engine()
    with Session(engine) as session:
        service = DatasetService(session)
        payload = _payload(domain)
        first = getattr(service, method)(payload, f"file://{domain}", datetime.now(timezone.utc))
        second = getattr(service, method)(payload, f"file://{domain}", datetime.now(timezone.utc))
        assert first.id == second.id
        assert second.status == "ACTIVE"
        assert second.record_count == len(payload.entries)

        result = lookup(service)
        assert result
        assert all(item is not None and item is not False for item in result) or result == [True]


def test_dataset_versions_includes_all_new_kinds():
    engine = _engine()
    with Session(engine) as session:
        service = DatasetService(session)
        versions = service.dataset_versions()
        for key in ("cfop", "fcp", "anp", "ncm_utrib", "rtc_biofuel_index", "rtc_ccredpres"):
            assert key in versions
            assert versions[key] is None
        service.import_cfop(_payload("cfop"), "file://cfop", datetime.now(timezone.utc))
        assert service.dataset_versions()["cfop"] is not None