"""Тесты ingest_reservations (#1062) — offline, без живой сети и реальной БД.

Мокаем: pravo_gov66_client (search/get_pdf_url/fetch_pdf) + extract_from_pdf + SessionLocal.
Проверяем: фильтрацию заголовков, дедупликацию doc_id, upsert-параметры, graceful на ошибках.
"""

from __future__ import annotations

from contextlib import contextmanager
from datetime import date
from typing import Any

from app.services.scrapers.page_reservation_parser import ReservationRecord
from app.workers.tasks import reservation_ingest

# ── Двойники ──────────────────────────────────────────────────────────────────


class _FakeDB:
    """Session-двойник: пишет execute-вызовы, begin_nested как no-op CM."""

    def __init__(self) -> None:
        self.executed: list[tuple[Any, dict[str, Any] | None]] = []
        self.commit_count = 0
        self.closed = False

    @contextmanager
    def begin_nested(self) -> Any:
        yield

    def execute(self, sql: Any, params: dict[str, Any] | None = None) -> Any:
        self.executed.append((sql, params))
        return None

    def commit(self) -> None:
        self.commit_count += 1

    def rollback(self) -> None:
        pass

    def close(self) -> None:
        self.closed = True


# ── Фикстуры ──────────────────────────────────────────────────────────────────

_SAMPLE_RECORDS = [
    ReservationRecord(
        cad_num="66:41:0101001:1",
        reservation_kind="изъятие",
        act_number="509-ПП",
        act_date=date(2024, 3, 12),
        purpose="реконструкция ул. Татищева",
        raw_excerpt="...участок 66:41:0101001:1...",
    ),
    ReservationRecord(
        cad_num="66:41:0101001:2",
        reservation_kind="изъятие",
        act_number="509-ПП",
        act_date=date(2024, 3, 12),
        purpose="реконструкция ул. Татищева",
        raw_excerpt=None,
    ),
]

_RELEVANT_HITS = [
    {
        "doc_id": "35496",
        "title": "Об изъятии земельных участков Екатеринбург",
        "url": "https://www.pravo.gov66.ru/35496/",
    },
    {
        "doc_id": "34812",
        "title": "О резервировании земельных участков Свердловской области",
        "url": "https://www.pravo.gov66.ru/34812/",
    },
]

_IRRELEVANT_HIT = {
    "doc_id": "99999",
    "title": "О бюджете на 2024 год",
    "url": "https://www.pravo.gov66.ru/99999/",
}

_PDF_BYTES = b"%PDF-1.4 fake content"


# ── Helpers ───────────────────────────────────────────────────────────────────


def _patch_enumerate(
    monkeypatch: Any,
    db: _FakeDB,
    hits: list[dict[str, Any]],
    pdf_url: str | None = "https://www.pravo.gov66.ru/media/pravo/509-ПП_abc.pdf",
    pdf_bytes: bytes = _PDF_BYTES,
    records: list[ReservationRecord] = _SAMPLE_RECORDS,
) -> None:
    """Патчит все зависимости ingest_reservations для enumerate-режима."""
    monkeypatch.setattr(reservation_ingest, "SessionLocal", lambda: db)
    monkeypatch.setattr(reservation_ingest, "search_documents", lambda q, **kw: hits)
    monkeypatch.setattr(reservation_ingest, "get_document_pdf_url", lambda doc_id: pdf_url)
    monkeypatch.setattr(reservation_ingest, "fetch_pdf", lambda url: pdf_bytes)
    monkeypatch.setattr(reservation_ingest, "extract_from_pdf", lambda b: records)


# ── Тесты фильтрации заголовков ───────────────────────────────────────────────


class TestIsRelevantTitle:
    def test_relevant_izyatie_ekb(self) -> None:
        assert reservation_ingest._is_relevant_title(
            "Об изъятии земельных участков в г. Екатеринбурге"
        )

    def test_relevant_rezervirovanie_sverdlov(self) -> None:
        assert reservation_ingest._is_relevant_title(
            "О резервировании земельных участков Свердловской области"
        )

    def test_irrelevant_no_geo(self) -> None:
        assert not reservation_ingest._is_relevant_title(
            "Об изъятии земельных участков в г. Нижний Тагил"
        )

    def test_irrelevant_no_operation(self) -> None:
        assert not reservation_ingest._is_relevant_title(
            "О бюджете Свердловской области на 2024 год"
        )

    def test_irrelevant_empty(self) -> None:
        assert not reservation_ingest._is_relevant_title("")


# ── Тесты enumerate-режима ────────────────────────────────────────────────────


class TestIngestEnumerate:
    def test_returns_docs_and_parcels(self, monkeypatch: Any) -> None:
        db = _FakeDB()
        _patch_enumerate(monkeypatch, db, hits=_RELEVANT_HITS)
        result = reservation_ingest.ingest_reservations()
        assert result["docs"] >= 1
        assert result["parcels"] >= 1

    def test_filters_irrelevant_titles(self, monkeypatch: Any) -> None:
        """Документ с нерелевантным заголовком не должен попасть в ingest."""
        db = _FakeDB()
        irrelevant_only = [_IRRELEVANT_HIT]
        _patch_enumerate(monkeypatch, db, hits=irrelevant_only)
        result = reservation_ingest.ingest_reservations()
        assert result["docs"] == 0
        assert result["parcels"] == 0
        assert db.executed == []

    def test_deduplicates_doc_id_across_queries(self, monkeypatch: Any) -> None:
        """Один и тот же doc_id из двух запросов инжестируется один раз."""
        db = _FakeDB()
        same_hit = _RELEVANT_HITS[:1]  # один хит

        call_count = [0]

        def _search_twice(q: str, **kw: Any) -> list[dict[str, Any]]:
            call_count[0] += 1
            return same_hit  # оба запроса возвращают один и тот же doc_id

        monkeypatch.setattr(reservation_ingest, "SessionLocal", lambda: db)
        monkeypatch.setattr(reservation_ingest, "search_documents", _search_twice)
        monkeypatch.setattr(
            reservation_ingest,
            "get_document_pdf_url",
            lambda doc_id: "https://www.pravo.gov66.ru/media/pravo/test.pdf",
        )
        monkeypatch.setattr(reservation_ingest, "fetch_pdf", lambda url: _PDF_BYTES)
        monkeypatch.setattr(reservation_ingest, "extract_from_pdf", lambda b: _SAMPLE_RECORDS[:1])

        result = reservation_ingest.ingest_reservations()
        # 2 запроса, но 1 уникальный doc_id → fetched=1
        assert result["docs"] == 1

    def test_skips_doc_when_no_pdf_url(self, monkeypatch: Any) -> None:
        db = _FakeDB()
        _patch_enumerate(monkeypatch, db, hits=_RELEVANT_HITS[:1], pdf_url=None)
        result = reservation_ingest.ingest_reservations()
        assert result["docs"] == 0
        assert result["parcels"] == 0

    def test_skips_doc_when_pdf_fetch_fails(self, monkeypatch: Any) -> None:
        db = _FakeDB()
        monkeypatch.setattr(reservation_ingest, "SessionLocal", lambda: db)
        monkeypatch.setattr(
            reservation_ingest, "search_documents", lambda q, **kw: _RELEVANT_HITS[:1]
        )
        monkeypatch.setattr(
            reservation_ingest,
            "get_document_pdf_url",
            lambda doc_id: "https://www.pravo.gov66.ru/media/pravo/test.pdf",
        )
        monkeypatch.setattr(
            reservation_ingest, "fetch_pdf", lambda url: (_ for _ in ()).throw(OSError("net err"))
        )
        monkeypatch.setattr(reservation_ingest, "extract_from_pdf", lambda b: _SAMPLE_RECORDS[:1])
        result = reservation_ingest.ingest_reservations()
        assert result["docs"] == 0

    def test_upsert_params_contain_cad_num(self, monkeypatch: Any) -> None:
        db = _FakeDB()
        _patch_enumerate(monkeypatch, db, hits=_RELEVANT_HITS[:1], records=_SAMPLE_RECORDS[:1])
        reservation_ingest.ingest_reservations()
        assert db.executed, "должен быть хотя бы один execute"
        _, params = db.executed[0]
        assert params is not None
        assert params["cad_num"] == "66:41:0101001:1"
        assert params["source"] == "pravo_gov66"

    def test_db_closed_after_run(self, monkeypatch: Any) -> None:
        db = _FakeDB()
        _patch_enumerate(monkeypatch, db, hits=[])
        reservation_ingest.ingest_reservations()
        assert db.closed

    def test_skips_empty_records(self, monkeypatch: Any) -> None:
        """PDF без кад-номеров → ни одного execute."""
        db = _FakeDB()
        _patch_enumerate(monkeypatch, db, hits=_RELEVANT_HITS[:1], records=[])
        result = reservation_ingest.ingest_reservations()
        assert result["parcels"] == 0
        assert db.executed == []


# ── Тесты explicit-docs (back-compat) ─────────────────────────────────────────


class TestIngestExplicitDocs:
    def test_explicit_docs_bypass_enumerate(self, monkeypatch: Any) -> None:
        """docs=... → enumerate не вызывается, search_documents не дёргается."""
        db = _FakeDB()
        monkeypatch.setattr(reservation_ingest, "SessionLocal", lambda: db)

        def _should_not_call(*a: Any, **kw: Any) -> list[dict[str, Any]]:
            raise AssertionError("enumerate не должен вызываться")

        monkeypatch.setattr(reservation_ingest, "search_documents", _should_not_call)
        monkeypatch.setattr(reservation_ingest, "fetch_pdf", lambda url: _PDF_BYTES)
        monkeypatch.setattr(reservation_ingest, "extract_from_pdf", lambda b: _SAMPLE_RECORDS[:1])

        explicit = [
            {
                "url": "https://www.pravo.gov66.ru/media/pravo/509-ПП_abc.pdf",
                "act_number": "509-ПП",
                "basis_act": "Постановление № 509-ПП",
                "purpose": "реконструкция",
                "kind": "изъятие",
            }
        ]
        result = reservation_ingest.ingest_reservations(docs=explicit)
        assert result["docs"] == 1
        assert result["parcels"] == 1

    def test_explicit_docs_upsert_uses_doc_kind(self, monkeypatch: Any) -> None:
        """Если rec.reservation_kind пуст, берётся kind из doc."""
        db = _FakeDB()
        monkeypatch.setattr(reservation_ingest, "SessionLocal", lambda: db)
        monkeypatch.setattr(reservation_ingest, "fetch_pdf", lambda url: _PDF_BYTES)

        no_kind_record = ReservationRecord(
            cad_num="66:41:0000001:1",
            reservation_kind="",  # пустой — должен взяться из doc
            act_number=None,
            act_date=None,
            purpose=None,
            raw_excerpt=None,
        )
        monkeypatch.setattr(reservation_ingest, "extract_from_pdf", lambda b: [no_kind_record])

        explicit = [
            {
                "url": "https://x.ru/test.pdf",
                "act_number": "1-ПП",
                "basis_act": "Постановление",
                "purpose": None,
                "kind": "резервирование",
            }
        ]
        reservation_ingest.ingest_reservations(docs=explicit)
        assert db.executed
        _, params = db.executed[0]
        assert params is not None
        assert params["reservation_kind"] == "резервирование"