"""WHO GHO connector — offline tests on a trimmed recorded payload (M_Est_tob_curr, CAN + USA + decoys) + live smoke."""
from __future__ import annotations

from datetime import UTC, datetime

import orjson
import pytest

from countryatlas.connectors.who import WHOConnector
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="smoking-prevalence", connector="who", dataset="GHO", code="M_Est_tob_curr",
                           params={"Dim1": "SEX_BTSX"}, priority=2)


def _raw(fixtures_dir):
    return raw_from_file("who", "GHO", "M_Est_tob_curr", fixtures_dir / "who" / "M_Est_tob_curr.json", "application/json",
                         indicator_meta={"name": "Estimate of current tobacco use prevalence (%)"})


def test_build_filter_includes_country_type_and_dims():
    f = WHOConnector.build_filter({"Dim1": "SEX_BTSX", "Dim2": "AGEGROUP_YEARSALL"})
    assert f == "SpatialDimType eq 'COUNTRY' and Dim1 eq 'SEX_BTSX' and Dim2 eq 'AGEGROUP_YEARSALL'"
    assert WHOConnector.build_filter(None) == "SpatialDimType eq 'COUNTRY'"


def test_normalize_maps_iso3_filters_dims_and_flags_projections(fixtures_dir):
    conn = WHOConnector()
    rows = conn.normalize(_raw(fixtures_dir), SPEC)
    # decoys: REGION row (AMR), SEX_MLE row, unknown ISO3 'ZZZ' are all dropped
    assert {r.country_id for r in rows} == {"CAN", "USA"}
    assert all(r.frequency == "A" and r.period.month == 1 and r.period.day == 1 for r in rows)
    assert all(r.unit == "% of adults" for r in rows)
    can = {r.year: r for r in rows if r.country_id == "CAN"}
    assert can[2030].is_forecast is True and can[2030].value == pytest.approx(8.3)
    assert can[2024].is_forecast is False
    assert all(r.year <= datetime.now(UTC).year or r.is_forecast for r in rows)
    # confidence bounds go to metadata; the row `Date` feeds source_updated_at
    assert "low" in rows[0].metadata and "high" in rows[0].metadata
    assert rows[0].source_updated_at is not None and rows[0].source_updated_at.year >= 2024
    # no duplicate keys
    keys = {(r.country_id, r.period) for r in rows}
    assert len(keys) == len(rows)
    assert conn.validate(rows).quarantine_dataset is False


def test_normalize_applies_scalar_transform(fixtures_dir):
    spec = SPEC.model_copy(update={"transform": "x/10", "indicator_id": "nurses-per-1000"})
    rows = WHOConnector().normalize(_raw(fixtures_dir), spec)
    can = {r.year: r for r in rows if r.country_id == "CAN"}
    assert can[2030].value == pytest.approx(0.83)
    assert can[2030].unit == "per 1,000 people"


def test_fetch_uses_filter_and_metadata(fixtures_dir, fake_get):
    conn = WHOConnector()
    data = (fixtures_dir / "who" / "M_Est_tob_curr.json").read_bytes()
    meta = (fixtures_dir / "who" / "Indicator_M_Est_tob_curr.json").read_bytes()

    def router(url, params):
        if url.endswith("/Indicator"):
            return meta, "application/json"
        return data, "application/json"

    calls = fake_get(conn, router)
    payloads = conn.fetch(SPEC)
    assert len(payloads) == 1 and payloads[0].pages == 1
    assert calls[0]["url"].endswith("/Indicator") and "M_Est_tob_curr" in calls[0]["params"]["$filter"]
    assert calls[1]["url"].endswith("/M_Est_tob_curr")
    assert calls[1]["params"]["$filter"] == "SpatialDimType eq 'COUNTRY' and Dim1 eq 'SEX_BTSX'"
    assert payloads[0].meta["indicator_meta"]["name"].startswith("Estimate of current tobacco")
    assert orjson.loads(payloads[0].body)["value"]


@pytest.mark.live
def test_live_suicide_rate_has_no_duplicates():
    conn = WHOConnector()
    spec = IndicatorSourceSpec(indicator_id="suicide-rate", connector="who", dataset="GHO", code="SDGSUICIDE",
                               params={"Dim1": "SEX_BTSX", "Dim2": "AGEGROUP_YEARSALL"})
    rows = conn.normalize(conn.fetch(spec), spec)
    assert len(rows) > 3000 and "CAN" in {r.country_id for r in rows}
    assert conn.validate(rows).quarantine_dataset is False