"""ILOSTAT connector — offline tests on a trimmed DF_UNE_2EAP_SEX_AGE_RT CSV (CAN, FRA, SOM, X01), plus live smoke."""
from __future__ import annotations

from datetime import date

import pytest

from countryatlas.connectors.ilo import ILOConnector
from countryatlas.models import IndicatorSourceSpec
from tests.connectors._helpers import install_fake_get, raw_from_file


@pytest.fixture
def fake_get(monkeypatch):
    return install_fake_get(monkeypatch)


SPEC =IndicatorSourceSpec(indicator_id="unemployment-rate", connector="ilo", dataset="DF_UNE_2EAP_SEX_AGE_RT",
                           code=".A..SEX_T.AGE_YTHADULT_YGE15", frequency="A", priority=4)


def _raw(fixtures_dir, **meta):
    return raw_from_file("ilo", SPEC.dataset, SPEC.code, fixtures_dir / "ilo" / "DF_UNE_2EAP_SEX_AGE_RT.csv", "text/csv", **meta)


def test_normalize_drops_aggregates_flags_projections_and_imputations(fixtures_dir):
    conn = ILOConnector()
    rows = conn.normalize(_raw(fixtures_dir, dataflow_meta={"release_year": 2025, "name": "… Nov. 2025"}), SPEC)
    assert {r.country_id for r in rows} == {"CAN", "FRA", "SOM"}  # X01 aggregate dropped
    can = {r.year: r for r in rows if r.country_id == "CAN"}
    assert can[2024].is_estimate is False and can[2024].is_forecast is False  # OBS_STATUS = R
    assert can[2025].is_estimate is True and can[2025].is_forecast is False  # nowcast (blank status), release year
    assert can[2026].is_forecast is True and can[2027].is_forecast is True
    assert can[2025].value == pytest.approx(6.907)  # = WB SL.UEM.TOTL.ZS 2025
    som = {r.year: r for r in rows if r.country_id == "SOM"}
    assert som[2024].is_estimate is True and som[2024].metadata["low"] == pytest.approx(4.737)
    assert all(r.frequency == "A" and r.period == date(r.year, 1, 1) and r.unit == "% of labour force" for r in rows)
    assert conn.validate(rows).quarantine_dataset is False


def test_release_year_fallback_is_current_year(fixtures_dir):
    from datetime import UTC, datetime

    rows = ILOConnector().normalize(_raw(fixtures_dir), SPEC)  # no dataflow_meta → years > now.year are forecasts
    y = datetime.now(UTC).year
    assert all(r.is_forecast == (r.year > y) for r in rows)


def test_fetch_builds_verified_url_and_parses_release_year(fixtures_dir, fake_get):
    conn = ILOConnector()
    csv_body = (fixtures_dir / "ilo" / "DF_UNE_2EAP_SEX_AGE_RT.csv").read_bytes()
    flow = (fixtures_dir / "ilo" / "dataflow.json").read_bytes()

    def router(url, params):
        if "/dataflow/ILO/" in url:
            return flow, "application/vnd.sdmx.structure+json"
        return csv_body, "text/csv"

    calls = fake_get(conn, router)
    p = conn.fetch(SPEC)
    assert calls[0]["url"] == "https://sdmx.ilo.org/rest/dataflow/ILO/DF_UNE_2EAP_SEX_AGE_RT"
    assert calls[1]["url"] == "https://sdmx.ilo.org/rest/data/ILO,DF_UNE_2EAP_SEX_AGE_RT/.A..SEX_T.AGE_YTHADULT_YGE15"
    assert calls[1]["params"] == {"format": "csv"}
    assert p.meta["dataflow_meta"]["release_year"] == 2025
    assert "Nov. 2025" in p.meta["dataflow_meta"]["name"]
    rows = conn.normalize(p, SPEC)
    assert {r.year for r in rows if r.is_forecast} == {2026, 2027}


@pytest.mark.live
def test_live_youth_unemployment():
    conn = ILOConnector()
    spec = SPEC.model_copy(update={"indicator_id": "youth-unemployment-rate", "code": ".A..SEX_T.AGE_YTHADULT_Y15-24",
                                   "params": {"startPeriod": "2023"}})
    rows = conn.normalize(conn.fetch(spec), spec)
    assert len({r.country_id for r in rows}) >= 150 and "CAN" in {r.country_id for r in rows}
    assert conn.validate(rows).quarantine_dataset is False