gendesign/tradein-mvp/backend/tests/test_avito_newbuilding_sweep.py

"""Avito newbuilding (novostroyka) citywide sweep.

Тесты (без сети, без реального DB):
1. _build_newbuilding_url — novostroyka-slug в пути, s=104, page param, без geo.
2. _parse_html на novostroyka-SERP fixture → listing_segment='novostroyki' +
   newbuilding_id заполнен (DOM-маркер застройщика + zhk-якорь).
3. fetch_newbuildings break-on-empty + dedup (mirror fetch_city_wide).
4. NewbuildingSweepCounters — defaults + to_dict.

Legacy `run_avito_newbuilding_sweep` orchestration wiring (fetch_newbuildings +
save_listings + mark_done) удалён вместе с `app.services.scrape_pipeline` (#2397
Part E1) — эквивалентная regression-coverage теперь в
`test_scraper_kit_pipeline_parity2.py::test_avito_newbuilding_sweep` (kit
`run_avito_newbuilding_sweep`).
"""

from __future__ import annotations

import os
from pathlib import Path
from unittest.mock import AsyncMock, patch

import pytest

os.environ.setdefault("DATABASE_URL", "postgresql+psycopg://test:test@localhost:5432/test")

from app.services.scrapers.avito import NOVOSTROYKA_SLUG, AvitoScraper
from app.services.scrapers.base import ScrapedLot

FIXTURE = Path(__file__).parent / "fixtures" / "avito_serp_novostroyka.html"


def _make_lot(source_id: str) -> ScrapedLot:
    return ScrapedLot(
        source="avito",
        source_url=f"https://www.avito.ru/ekaterinburg/kvartiry/{source_id}",
        source_id=source_id,
        price_rub=6_000_000,
        listing_segment="novostroyki",
    )


# ── 1. _build_newbuilding_url ───────────────────────────────────────────────


def test_build_newbuilding_url_contains_slug_and_sort() -> None:
    s = AvitoScraper()
    url = s._build_newbuilding_url(page=1)
    assert NOVOSTROYKA_SLUG in url
    assert "novostroyka-" in url
    assert "geoCoords" not in url
    assert "radius" not in url
    assert "s=104" in url
    assert "p=1" in url
    assert "ekaterinburg/kvartiry/prodam/" in url


def test_build_newbuilding_url_page_param() -> None:
    s = AvitoScraper()
    url = s._build_newbuilding_url(page=2)
    assert "novostroyka-" in url
    assert "p=2" in url


def test_build_newbuilding_url_differs_from_citywide() -> None:
    s = AvitoScraper()
    assert s._build_newbuilding_url(page=1) != s._build_citywide_url(page=1)


# ── 2. _parse_html classifies novostroyka cards ─────────────────────────────


def test_parse_novostroyka_fixture_segments_and_newbuilding_id() -> None:
    html = FIXTURE.read_text(encoding="utf-8")
    s = AvitoScraper()
    lots = s._parse_html(html, source_url_base=s._build_newbuilding_url(page=1))

    assert len(lots) == 2
    for lot in lots:
        assert lot.listing_segment == "novostroyki"
        assert lot.newbuilding_id is not None
        assert lot.newbuilding_url is not None

    ids = {lot.newbuilding_id for lot in lots}
    assert "meridian-ekaterinburg" in ids
    assert "severnyy-kvartal-ekaterinburg" in ids


# ── 3. fetch_newbuildings break-on-empty + dedup ────────────────────────────


@pytest.mark.asyncio
async def test_fetch_newbuildings_break_on_empty() -> None:
    s = AvitoScraper()
    pages_data: list[list[ScrapedLot]] = [[_make_lot("A"), _make_lot("B")], []]
    call_n = 0

    async def mock_html(url: str, page: int) -> str:
        assert "novostroyka-" in url
        return f"<html>{page}</html>"

    def mock_parse(html: str, source_url_base: str) -> list[ScrapedLot]:
        nonlocal call_n
        idx = call_n
        call_n += 1
        return pages_data[idx] if idx < len(pages_data) else []

    with patch.object(s, "_fetch_serp_html", side_effect=mock_html):
        with patch.object(s, "_parse_html", side_effect=mock_parse):
            with patch.object(s, "sleep_between_requests", new=AsyncMock()):
                result = await s.fetch_newbuildings(pages=10, delay_override_sec=0)

    assert len(result) == 2
    assert call_n == 2  # page1 + page2(empty) → stop


@pytest.mark.asyncio
async def test_fetch_newbuildings_dedup_by_source_id() -> None:
    s = AvitoScraper()
    dup = _make_lot("SAME")
    pages_data: list[list[ScrapedLot]] = [[dup], [dup], []]
    call_n = 0

    async def mock_html(url: str, page: int) -> str:
        return f"<html>{page}</html>"

    def mock_parse(html: str, source_url_base: str) -> list[ScrapedLot]:
        nonlocal call_n
        idx = call_n
        call_n += 1
        return pages_data[idx] if idx < len(pages_data) else []

    with patch.object(s, "_fetch_serp_html", side_effect=mock_html):
        with patch.object(s, "_parse_html", side_effect=mock_parse):
            with patch.object(s, "sleep_between_requests", new=AsyncMock()):
                result = await s.fetch_newbuildings(pages=10, delay_override_sec=0)

    assert len(result) == 1
    assert result[0].source_id == "SAME"


def test_fetch_city_wide_url_unchanged() -> None:
    """fetch_city_wide остаётся citywide-URL — рефактор не сменил его поведение."""
    s = AvitoScraper()
    assert "novostroyka-" not in s._build_citywide_url(page=1)
    assert "prodam-ASgBAgICAUSSA8YQ" in s._build_citywide_url(page=1)


# ── 4. NewbuildingSweepCounters ─────────────────────────────────────────────


def test_newbuilding_counters_defaults_and_to_dict() -> None:
    from scraper_kit.orchestration.pipeline import NewbuildingSweepCounters

    c = NewbuildingSweepCounters()
    assert c.lots_fetched == 0
    assert c.lots_inserted == 0
    assert c.lots_updated == 0
    assert c.errors_count == 0
    d = c.to_dict()
    assert set(d.keys()) == {"lots_fetched", "lots_inserted", "lots_updated", "errors_count"}