"""Airport pages: the URL shape, the rendered body and the structured data. The first hub page type (build plan §6, B4). The page is keyed by the IATA code on the `shops` row; the words after it are for humans and may go stale. Everything the page shows is counted in the database, so these tests build one AirportDetail by hand and pin what the renderer must say about it: the H1, the counts, the product cards with their prices and links, CollectionPage about an Airport with an ItemList, breadcrumbs, and a canonical that ignores filters. No database. """ import json import pathlib import re from datetime import UTC, datetime import pytest from app.models.hubs import AirportDetail, AirportShop from app.models.schemas import CategoryCount, ProductSummary, ShopPrice, TopAward from app.services import place_guides, seo, urls CHECKED = datetime(2026, 8, 25, 23, 6, tzinfo=UTC) SHELL = ( "x" '' '
' ) def summary(**over) -> ProductSummary: base = dict( id=1156, name="Santa Teresa 1796 Solera Rum 1L", brand="Santa Teresa", category="Rum", quantity_ml=1000, abv=40.0, is_exclusive=False, thumb_url="https://img.example/1156.jpg", shop_count=3, best_shop="London Heathrow", best_shop_iata="LHR", shop_labels=["LHR", "JFK", "EZE"], top_prices=[ShopPrice(label="LHR", usd=38.0, in_stock=True), ShopPrice(label="JFK", usd=45.0, in_stock=None), ShopPrice(label="EZE", usd=55.5, in_stock=True)], top_award=TopAward(competition_slug="nyisc", competition="NYISC", medal="Double Gold", year=2024), cheapest_usd=38.0, dearest_usd=55.5, award_count=1, ) base.update(over) return ProductSummary(**base) def airport(**over) -> AirportDetail: base = dict( iata="LHR", path="/airports/heathrow-lhr-london", name="London Heathrow", city="London", country="United Kingdom", currency="GBP", product_variants=266, last_collected_at=CHECKED, shops=[AirportShop(code="LHR", retailer_name="Avolta", product_variants=266, last_collected_at=CHECKED)], comparable=108, cheapest_here=67, exclusives=8, categories=[CategoryCount(category="Whisky", count=67), CategoryCount(category="Gin & ", count=24)], savings=[summary()], exclusive_items=[summary(id=7, name="Talisker Dark Storm 1L", is_exclusive=True, award_count=0, top_award=None, shop_count=1, top_prices=[], thumb_url=None, cheapest_usd=60.0, dearest_usd=60.0)], total=266, limit=24, offset=0, items=[summary(), summary(id=8, name="Bell's Blended Scotch 1l", award_count=0, top_award=None)], ) base.update(over) return AirportDetail(**base) class TestAirportUrls: def test_declared_airport_leads_with_its_common_name(self): """Mark's review of 7 Sep: "heathrow" is overwhelmingly what is searched, so the name leads, then the code, then the city; JFK is the one airport known by its code. The pages had shipped code-first four days earlier; nothing was indexed, so the cost was a rename plus redirects.""" assert urls.airport_path("LHR", "London") == "/airports/heathrow-lhr-london" assert urls.airport_path("JFK", "New York", "New York JFK") == "/airports/jfk-new-york" # The city is dropped where the name already says it. assert urls.airport_path("HKG", "Hong Kong", "Hong Kong International") == "/airports/hong-kong-hkg" # Paris is CDG and Orly together, named for Paris. assert urls.airport_path("CDG", "Paris", "Paris Charles de Gaulle and Orly") == "/airports/paris-cdg" assert urls.airport_path("DUB", "Dublin", "Dublin & Cork") == "/airports/dublin-dub" # Case of the code does not matter. assert urls.airport_path("lhr", "London") == "/airports/heathrow-lhr-london" def test_undeclared_airport_keeps_the_old_shape_so_nothing_404s(self): assert urls.airport_path("XYZ", "Nowhere") == "/airports/xyz-nowhere" assert urls.airport_path("XYZ", None, "Some Name") == "/airports/xyz-some-name" assert urls.airport_path("QQQ", None, None) == "/airports/qqq" def test_every_shape_ever_linked_resolves_to_its_airport(self): """Every earlier shape must still land, the bare code, the pre-review `lhr-london`, a stale new-shape slug, because the route 301s from whatever resolves.""" for slug in ( "heathrow-lhr-london", "lhr", "LHR", "LHR-Heathrow", "lhr-london", "lhr-london-airport", "heathrow-lhr-terminal-5", "london-lhr", ): assert urls.parse_airport_slug(slug) == "LHR", slug assert urls.parse_airport_slug("jfk-new-york") == "JFK" assert urls.parse_airport_slug("jfk") == "JFK" # Undeclared old shape. assert urls.parse_airport_slug("xyz-nowhere") == "XYZ" for slug in ("nope", "12a-x", "lh", "lhrx", "", "athens", "london"): assert urls.parse_airport_slug(slug) is None, slug def test_a_three_letter_word_is_only_a_code_when_it_leads_or_is_declared(self): """San Salvador's slug starts with a three-letter word that is also San Diego's code; it must resolve to SAL, exactly and when stale, never to an airport that was never in the table.""" assert urls.parse_airport_slug("san-salvador-sal") == "SAL" assert urls.parse_airport_slug("san-salvador-sal-stale") == "SAL" def test_no_declared_slug_collides_with_another_declared_code(self): for iata, slug in urls.AIRPORT_SLUGS.items(): assert re.fullmatch(r"[A-Z]{3}", iata) assert re.fullmatch(r"[a-z0-9]+(-[a-z0-9]+)*", slug) assert urls.parse_airport_slug(slug) == iata head = slug.split("-")[0] # Only a true collision risk when the leading word is itself a # declared code (SAL's slug leads with "san", San Diego's code, # but SAN is undeclared, so parse_airport_slug never sees it as # a match; see test_a_three_letter_word_is_only_a_code...). if len(head) == 3 and head.upper() in urls.AIRPORT_SLUGS: assert head.upper() == iata, f"{slug} leads with a code that is not its own" assert iata.lower() in slug.split("-") assert len(set(urls.AIRPORT_SLUGS.values())) == len(urls.AIRPORT_SLUGS) def test_query_omits_defaults_so_the_clean_page_has_none(self): assert urls.airport_query(None, "featured", 1) == "" assert urls.airport_query("Gin & Tonic", "featured", 1) == "?category=Gin%20%26%20Tonic" assert urls.airport_query(None, "price", 3) == "?sort=price&page=3" def test_feature_flags_ride_between_the_category_and_the_sort_and_imply_the_full_list(self): """G5: the storefront's chips as query parameters named as /api/products names them; any one of them can only mean the full list, like a category.""" assert urls.airport_query("Whisky", "price", 2, "all", {"awarded_only": True}) == "?tab=all&category=Whisky&awarded_only=true&sort=price&page=2" assert urls.airport_query(None, "featured", 1, "all", urls.feature_flags(multi_only=True, exclusives_only=True)) == "?tab=all&multi_only=true&exclusives_only=true" assert urls.airport_query(None, "featured", 1, "all", {}) == "?tab=all" assert urls.airport_tab(None, None, "featured", 1, {"multi_only": True}) == "all" assert urls.airport_tab(None, None, "featured", 1, {}) == "value" assert urls.airport_tab("exclusives", None, "featured", 1, {"awarded_only": True}) == "exclusives" assert urls.read_flag("true") and urls.read_flag("1") and urls.read_flag(" Yes ") and not urls.read_flag("bogus") and not urls.read_flag(None) assert urls.feature_flags(multi_only=False, awarded_only=True, nonsense=True) == {"awarded_only": True} def test_the_ts_mirror_names_the_same_feature_parameters(self): src = (pathlib.Path(__file__).resolve().parents[1] / "web" / "src" / "lib" / "urls.ts").read_text() block = src[src.index("export const AIRPORT_FEATURE_PARAMS"):src.index("];", src.index("export const AIRPORT_FEATURE_PARAMS"))] assert re.findall(r'"(multi_only|awarded_only|exclusives_only)"', block) == list(urls.FEATURE_PARAMS) def test_the_comparison_preset_address_and_its_mirror(self): """G6: `/savings?from=`, read once by the savings page as its view.""" assert urls.savings_from_path("lhr") == "/savings?from=LHR" src = (pathlib.Path(__file__).resolve().parents[1] / "web" / "src" / "lib" / "urls.ts").read_text() assert "return `/savings?from=${iata.toUpperCase()}`;" in src assert 'params.get("from")' in src def test_seo_module_still_exports_the_helpers(self): """Routes and older tests import the URL helpers from seo; the move to urls.py must not break that.""" assert seo.product_path(1, "x") == "/products/x-1" and seo.airport_path("LHR", "London") == "/airports/heathrow-lhr-london" class TestAirportSlugTable: def test_ts_mirror_holds_the_same_table(self): """The two tables must agree or server canonicals fight client links. Read the SPA's own file the way TestRouteInventory reads App.tsx.""" src = (pathlib.Path(__file__).resolve().parents[1] / "web" / "src" / "lib" / "urls.ts").read_text() start = src.index("export const AIRPORT_SLUGS") block = src[start:src.index("};", start)] entries = re.findall(r'^\s+([A-Z]{3}): "([a-z0-9-]+)",$', block, re.M) ts_table = dict(entries) assert ts_table assert ts_table == urls.AIRPORT_SLUGS class TestAirportHead: def test_title_description_and_canonical(self): head = seo.head_for_airport(airport()) assert head.title == "London Heathrow (LHR) duty-free prices | Duty Free Professor" assert head.description == ( "266 duty-free products at London Heathrow (LHR), checked 25 Aug 2026. " "108 sold at other airports too: see where London is the better buy." ) assert head.canonical_path == "/airports/heathrow-lhr-london" assert head.last_modified == CHECKED assert head.seed_key == "__DFP_AIRPORT__" and head.seed["iata"] == "LHR" def test_a_filtered_or_paged_view_keeps_the_clean_canonical(self): head = seo.head_for_airport(airport(category="Whisky", sort="price", offset=24)) assert head.canonical_path == "/airports/heathrow-lhr-london" assert not head.noindex def test_collection_page_about_an_airport_with_an_item_list(self): page = seo.head_for_airport(airport()).apply(SHELL, "https://s.example") blocks = [json.loads(m) for m in re.findall(r'', page)] collection = next(b for b in blocks if b["@type"] == "CollectionPage") assert collection["@id"] == "https://s.example/airports/heathrow-lhr-london#page" assert collection["url"] == "https://s.example/airports/heathrow-lhr-london" assert collection["isPartOf"] == {"@id": "https://s.example/#website"} about = collection["about"] assert about["@type"] == "Airport" and about["iataCode"] == "LHR" and about["name"] == "London Heathrow" # Only what the shops row holds: no street, no coordinates. assert about["address"] == {"@type": "PostalAddress", "addressLocality": "London", "addressCountry": "United Kingdom"} assert "geo" not in about and "streetAddress" not in about["address"] items = collection["mainEntity"] assert items["@type"] == "ItemList" and items["numberOfItems"] == 266 assert items["itemListElement"][0] == { "@type": "ListItem", "position": 1, "url": "https://s.example/products/santa-teresa-1796-solera-rum-1l-1156", "name": "Santa Teresa 1796 Solera Rum 1L", } crumbs = next(b for b in blocks if b["@type"] == "BreadcrumbList") assert crumbs["itemListElement"][0]["item"] == "https://s.example/airports" assert crumbs["itemListElement"][1] == {"@type": "ListItem", "position": 2, "name": "London Heathrow"} def test_positions_continue_across_pages(self): head = seo.head_for_airport(airport(offset=24)) positions = [e["position"] for e in head.jsonld[0]["mainEntity"]["itemListElement"]] assert positions == [25, 26] def test_an_airport_without_a_city_or_country_has_no_address(self): head = seo.head_for_airport(airport(city=None, country=None, path="/airports/lhr")) assert "address" not in head.jsonld[0]["about"] class TestAirportBody: def test_the_page_reads_without_javascript(self): body = seo.airport_body(airport()) assert '

Duty free at London Heathrow

' in body assert "266 duty-free products priced by Avolta at London Heathrow (LHR), checked 25 Aug 2026." in body for label, value in (("Products priced", "266"), ("Sold elsewhere too", "108"), ("Cheapest here", "67"), ("Travel exclusives", "8")): assert f"
{label}
{value}
" in body, label assert '
Shops we read
Avolta (266 products)
' in body assert '
Prices checked
25 Aug 2026
' in body assert 'href="/airports">Airports we price' in body def test_the_three_shelves_are_tabs_and_the_page_opens_on_the_savings(self): """The page used to stack four grids of cards; it now shows one shelf at a time, the shelf in the URL, so a shopper scrolls one list, not four.""" body = seo.airport_body(airport()) assert '