d17ac794fd
* Centralize filesystem paths in a single module Add paths.py, which resolves the repo root once and derives the cache, accounts DB, logs, templates, frontend and bundled-city-data locations from it. Replace the 13 per-module `dirname(__file__)/..` anchors with references to it, so a module's location no longer determines where the app reads its data. Env overrides (accounts DB, VAPID, IndexNow) are unchanged; every resolved path is byte-identical to before. Groundwork for moving modules into packages without re-pointing paths. Claude-Session: https://claude.ai/code/session_01XXxmNFy9cZ6Gh8Y9thZn62 * Split the backend into domain packages Group the flat backend modules into packages that mirror their concerns: data/ climate, grading, scoring, grid, places, cities, city_events, store web/ app, views, homepage, content, schemas notifications/ notify, digest, push, mailer, discord, discord_interactions, discord_link accounts/ models, users, api_accounts, db core/ metrics, singleton, audit Intra-project imports are rewritten to the package-qualified form. The entry scripts (indexnow, warm_cities, migrate, gen_cities, gen_flavor) and paths.py stay at the backend/ root, and backend/app.py becomes a shim re-exporting web.app:app so the launch target stays `app:app` — run.sh, the systemd units, and CI need no change. Verified: full suite (318) passes, `uvicorn app:app` boots and serves the home/SEO/static/API surfaces, and every root script imports clean. Claude-Session: https://claude.ai/code/session_01XXxmNFy9cZ6Gh8Y9thZn62
559 lines
26 KiB
Python
559 lines
26 KiB
Python
"""Tests for the crawlable SEO content pages: the city set, robots/sitemap, and
|
||
that the server-rendered pages contain the stats in the HTML (with the weather
|
||
layer faked, like test_api.py)."""
|
||
import math
|
||
import re
|
||
|
||
import pytest
|
||
from fastapi.testclient import TestClient
|
||
|
||
from web import app as appmod
|
||
from data import cities
|
||
from data import climate
|
||
from data import grading
|
||
|
||
# BASE defaults to /thermograph in tests (THERMOGRAPH_BASE unset).
|
||
B = "/thermograph"
|
||
SLUG = "london-england-gb" # present in cities.json
|
||
|
||
|
||
@pytest.fixture
|
||
def client(monkeypatch, history, recent):
|
||
monkeypatch.setattr(climate, "load_cached_history", lambda cell: history.clone())
|
||
monkeypatch.setattr(climate, "get_recent_forecast", lambda cell: recent.clone())
|
||
monkeypatch.setattr(climate, "get_history", lambda cell: (history.clone(), {"cached": True}))
|
||
return TestClient(appmod.app)
|
||
|
||
|
||
def test_city_set_slugs_unique_and_lookup():
|
||
slugs = cities.all_slugs()
|
||
assert len(slugs) == len(set(slugs)) >= 100
|
||
assert cities.get(SLUG) is not None
|
||
assert cities.get("does-not-exist") is None
|
||
|
||
|
||
def test_robots_txt(client):
|
||
r = client.get(f"{B}/robots.txt")
|
||
assert r.status_code == 200
|
||
assert "Sitemap:" in r.text and "/sitemap.xml" in r.text
|
||
assert "Disallow: /thermograph/api/" in r.text
|
||
|
||
|
||
def test_sitemap_lists_city_urls(client):
|
||
r = client.get(f"{B}/sitemap.xml")
|
||
assert r.status_code == 200
|
||
assert "<urlset" in r.text
|
||
assert f"/climate/{SLUG}</loc>" in r.text
|
||
assert f"/climate/{SLUG}/july</loc>" in r.text
|
||
assert f"/climate/{SLUG}/records</loc>" in r.text
|
||
|
||
|
||
def test_city_page_renders_stats_in_html(client):
|
||
r = client.get(f"{B}/climate/{SLUG}")
|
||
assert r.status_code == 200
|
||
b = r.text
|
||
assert "London" in b and "climate" in b.lower()
|
||
assert 'rel="canonical"' in b and f"/climate/{SLUG}" in b
|
||
assert '"@type":"Dataset"' in b
|
||
assert "average temperatures by month" in b.lower()
|
||
# a real number rendered in the normals/today text (London is GB, so °C)
|
||
assert re.search(r"-?\d+°C</span>", b)
|
||
|
||
|
||
def _jsonld_breadcrumbs(html: str) -> list[dict]:
|
||
import json, re
|
||
m = re.search(r'<script type="application/ld\+json">(.*?)</script>', html, re.S)
|
||
assert m, "no JSON-LD block on page"
|
||
graph = json.loads(m.group(1))["@graph"]
|
||
return [n for n in graph if n["@type"] == "BreadcrumbList"]
|
||
|
||
|
||
@pytest.mark.parametrize("path", [f"/climate/{SLUG}", f"/climate/{SLUG}/records"])
|
||
def test_breadcrumb_jsonld_items_valid_for_google(client, path):
|
||
# Google: every ListItem except the last must carry an ``item`` URL, so
|
||
# unlinked intermediate crumbs (the country) must not appear in JSON-LD.
|
||
(bc,) = _jsonld_breadcrumbs(client.get(B + path).text)
|
||
els = bc["itemListElement"]
|
||
assert [e["position"] for e in els] == list(range(1, len(els) + 1))
|
||
for e in els[:-1]:
|
||
assert e["item"].startswith("http"), f"crumb {e['name']!r} missing item"
|
||
|
||
|
||
def test_city_404(client):
|
||
assert client.get(f"{B}/climate/nope-not-a-city").status_code == 404
|
||
|
||
|
||
def test_curated_events_have_valid_slugs():
|
||
from data import city_events
|
||
slugs = set(cities.all_slugs())
|
||
assert len(city_events.EVENTS) >= 20
|
||
assert all(s in slugs for s in city_events.EVENTS), \
|
||
[s for s in city_events.EVENTS if s not in slugs]
|
||
|
||
|
||
def test_curated_event_renders(client):
|
||
b = client.get(f"{B}/climate/new-orleans-louisiana-us").text
|
||
assert "Notable weather in New Orleans" in b
|
||
assert "Hurricane Katrina" in b
|
||
|
||
|
||
def test_uncurated_city_has_no_event_section(client):
|
||
b = client.get(f"{B}/climate/shanghai-cn").text
|
||
assert "city-event" not in b # falls back to the flavor blurb
|
||
|
||
|
||
def test_city_travel_cta_prefills_compare(client):
|
||
b = client.get(f"{B}/climate/{SLUG}").text
|
||
assert "Thinking of visiting" in b
|
||
assert "/compare#loc=" in b # the city is pre-loaded on the compare page
|
||
|
||
|
||
def test_city_blurb_renders_with_attribution(client, monkeypatch):
|
||
monkeypatch.setattr(cities, "flavor",
|
||
lambda slug: {"extract": "Testville is a lovely place to test.",
|
||
"url": "https://en.wikipedia.org/wiki/Testville", "title": "Testville"})
|
||
b = client.get(f"{B}/climate/{SLUG}").text
|
||
assert "Testville is a lovely place to test." in b
|
||
assert "via Wikipedia" in b # CC BY-SA attribution link
|
||
|
||
|
||
def test_month_and_records(client):
|
||
assert client.get(f"{B}/climate/{SLUG}/july").status_code == 200
|
||
assert client.get(f"{B}/climate/{SLUG}/records").status_code == 200
|
||
assert client.get(f"{B}/climate/{SLUG}/notamonth").status_code == 404
|
||
|
||
|
||
def test_month_records_cover_every_metric_high_and_low(client):
|
||
# The month page's records show a record high AND low for each metric present, not
|
||
# just warmest/coldest temperature — one card per metric with both extremes. (The
|
||
# test fixture carries tmax/tmin/precip; real cities add feels/humidity/wind/gust.)
|
||
b = client.get(f"{B}/climate/{SLUG}/july").text
|
||
assert "July records" in b
|
||
for label in ("High", "Low", "Precip"):
|
||
assert f'class="rc-metric">{label}<' in b
|
||
# Every card carries a labelled high and low row (temp metrics use Highest/Lowest).
|
||
n = b.count('rc-row rc-high')
|
||
assert n >= 3 and b.count('rc-row rc-low') == n
|
||
assert "Highest" in b and "Lowest" in b
|
||
# Precip's extremes are the wettest/driest whole month by total accumulation.
|
||
assert "Wettest" in b and "Driest" in b
|
||
|
||
|
||
def test_records_page_has_monthly_and_seasonal(client):
|
||
b = client.get(f"{B}/climate/{SLUG}/records").text
|
||
# Monthly records: all 12 months, each linking to its month page.
|
||
assert "Records by month" in b
|
||
for month in ("January", "July", "December"):
|
||
assert month in b
|
||
assert f"/climate/{SLUG}/january" in b
|
||
# Seasonal records: labelled for London's Northern Hemisphere, with month spans.
|
||
assert "Records by season" in b
|
||
assert "Northern Hemisphere" in b
|
||
assert 'Winter <span class="season-span">(Dec–Feb)</span>' in b
|
||
assert 'Summer <span class="season-span">(Jun–Aug)</span>' in b
|
||
# All-time table still present, plus structured data and real values.
|
||
assert "All-time records" in b
|
||
assert '"@type":"Dataset"' in b
|
||
assert re.search(r"-?\d+°C</span>", b)
|
||
# The Dataset structured data still advertises the month/season coverage.
|
||
assert "monthly" in b.lower() and "seasonal" in b.lower()
|
||
|
||
|
||
def test_records_seasons_flip_in_southern_hemisphere(client):
|
||
# São Paulo is south of the equator, so Dec–Feb is summer, not winter.
|
||
b = client.get(f"{B}/climate/sao-paulo-br/records").text
|
||
assert "Southern Hemisphere" in b
|
||
assert 'Summer <span class="season-span">(Dec–Feb)</span>' in b
|
||
assert 'Winter <span class="season-span">(Jun–Aug)</span>' in b
|
||
|
||
|
||
def test_climate_pages_are_colour_coded(client):
|
||
# Temperatures wear the diverging cold→hot palette (a heat map, not a plain grid),
|
||
# and the city page carries the monthly temperature-range strip.
|
||
city = client.get(f"{B}/climate/{SLUG}").text
|
||
assert "t-cell t-" in city # tinted temperature cells in the normals table
|
||
assert "range-fill" in city and "--c1:var(--" in city # the monthly range strip's gradient bars
|
||
recs = client.get(f"{B}/climate/{SLUG}/records").text
|
||
assert "t-inline t-" in recs # tinted record chips
|
||
month = client.get(f"{B}/climate/{SLUG}/july").text
|
||
assert "t-inline t-" in month # tinted hero numbers + records
|
||
|
||
|
||
def test_records_show_both_extremes_per_metric(client):
|
||
# Each metric column shows BOTH its record warmest (▲) and coldest (▼) — so the
|
||
# daytime high carries its record-low high, and the overnight low its record-high low.
|
||
b = client.get(f"{B}/climate/{SLUG}/records").text
|
||
assert "Daytime high" in b and "Overnight low" in b
|
||
assert "▲" in b and "▼" in b
|
||
# 12 months × 2 metric columns × 2 extremes = 48 chips just in the monthly table.
|
||
assert b.count('class="rec-ext"') >= 48
|
||
|
||
|
||
def test_hub_glossary_about(client):
|
||
assert client.get(f"{B}/climate").status_code == 200
|
||
assert client.get(f"{B}/glossary").status_code == 200
|
||
assert client.get(f"{B}/glossary/percentile").status_code == 200
|
||
assert client.get(f"{B}/glossary/not-a-term").status_code == 404
|
||
assert client.get(f"{B}/about").status_code == 200
|
||
|
||
|
||
def test_footer_has_no_leaked_jinja_comment(client):
|
||
# A comment whose prose contained a literal "#}" closed early and leaked
|
||
# "wrapper. #}" into the footer. Guard every base-template page against a
|
||
# stray Jinja comment delimiter reaching the rendered HTML.
|
||
import re
|
||
for path in (f"{B}/about", f"{B}/climate/{SLUG}", f"{B}/climate/{SLUG}/records"):
|
||
b = client.get(path).text
|
||
footer = re.search(r"<footer.*?</footer>", b, re.S)
|
||
assert footer, f"no footer on {path}"
|
||
assert "#}" not in footer.group(0) and "{#" not in footer.group(0), path
|
||
|
||
|
||
def test_title_name_is_short_and_unique():
|
||
# Titles lead with the bare city name so the payload ("… in July") survives
|
||
# SERP truncation — display_name()'s region+country is what pushed it off the
|
||
# end. The exception is a name that recurs: those must still be told apart,
|
||
# and every rendered title must stay unique across the whole city set.
|
||
all_cities = cities.all_cities()
|
||
titles = [cities.title_name(c) for c in all_cities]
|
||
assert len(titles) == len(set(titles)), "two cities share a title"
|
||
# The overwhelming majority pay no disambiguation cost at all.
|
||
assert sum(1 for t in titles if "," not in t) > 0.9 * len(titles)
|
||
|
||
by_slug = {c["slug"]: cities.title_name(c) for c in all_cities}
|
||
assert by_slug["seattle-washington-us"] == "Seattle"
|
||
# Same name, different countries -> country code is enough.
|
||
assert by_slug["london-england-gb"] == "London, GB"
|
||
assert by_slug["london-ontario-ca"] == "London, CA"
|
||
# Same name AND same country -> the country code would collide, so fall back
|
||
# to admin1. Without this, both Columbuses would render "Columbus, US".
|
||
assert by_slug["columbus-ohio-us"] == "Columbus, Ohio"
|
||
assert by_slug["columbus-georgia-us"] == "Columbus, Georgia"
|
||
|
||
|
||
def test_page_titles_lead_with_city_and_payload(client):
|
||
for path, must_start, payload in (
|
||
(f"{B}/climate/{SLUG}", "London, GB climate", "how unusual it is now"),
|
||
(f"{B}/climate/{SLUG}/july", "London, GB in July", "normal weather"),
|
||
(f"{B}/climate/{SLUG}/records", "London, GB weather records", "hottest & coldest"),
|
||
):
|
||
b = client.get(path).text
|
||
title = re.search(r"<title>(.*?)</title>", b, re.S).group(1)
|
||
title = title.replace("&", "&")
|
||
assert title.startswith(must_start), title
|
||
assert payload in title, title
|
||
# The region+country the old template spent its width on is gone.
|
||
assert "England, United Kingdom" not in title
|
||
# Whatever else happens, the city and the page's subject are inside the
|
||
# first ~40 chars, so a SERP cut can no longer remove the payload.
|
||
assert len(must_start) <= 40
|
||
# og:title mirrors <title> via base.html.j2's self.title().
|
||
og = re.search(r'<meta property="og:title" content="(.*?)"', b, re.S).group(1)
|
||
assert og == re.search(r"<title>(.*?)</title>", b, re.S).group(1)
|
||
|
||
|
||
def test_meta_descriptions_lead_with_real_numbers(client):
|
||
for path in (f"{B}/climate/{SLUG}", f"{B}/climate/{SLUG}/july",
|
||
f"{B}/climate/{SLUG}/records"):
|
||
b = client.get(path).text
|
||
desc = re.search(r'<meta name="description" content="(.*?)"', b, re.S).group(1)
|
||
assert len(desc) <= 155, f"{len(desc)}: {desc}"
|
||
# A real temperature, not a generic "averages, records and rainfall" —
|
||
# in the city's own unit, since a meta description is never converted
|
||
# client-side. London is GB, so these read °C.
|
||
assert re.search(r"-?\d+°C", desc), desc
|
||
assert desc.endswith("years of local history."), desc
|
||
# The records description names the dates the extremes were set.
|
||
rec = client.get(f"{B}/climate/{SLUG}/records").text
|
||
desc = re.search(r'<meta name="description" content="(.*?)"', rec, re.S).group(1)
|
||
assert re.search(r"\([A-Z][a-z]{2} \d{4}\)", desc), desc
|
||
|
||
|
||
_TEMP_SPAN = re.compile(r'<span class="temp" data-temp-f="(-?[\d.]+)"([^>]*)>(-?\d+)°([CF]?)</span>')
|
||
|
||
|
||
def _temp_spans(html):
|
||
"""(fahrenheit_attr, shown_number, shown_letter) for every rendered temp span.
|
||
|
||
Reading the spans rather than the raw document matters: base.html.j2 mentions
|
||
"°F/°C" in an HTML comment, so a bare `"°F" in html` assertion passes on a page
|
||
whose every temperature is Celsius.
|
||
"""
|
||
return [(float(f), int(n), letter) for f, _attrs, n, letter in _TEMP_SPAN.findall(html)]
|
||
|
||
|
||
def test_climate_pages_carry_convertible_temps(client):
|
||
# Every temperature is a client-convertible span; the old inline "(NN°C)" dual
|
||
# form is gone (the toggle now supplies the other unit).
|
||
for path in (f"{B}/climate/{SLUG}", f"{B}/climate/{SLUG}/july", f"{B}/climate/{SLUG}/records"):
|
||
b = client.get(path).text
|
||
spans = _temp_spans(b)
|
||
assert spans, f"no temperature spans on {path}"
|
||
assert re.search(r"\d+°F \(-?\d+°C\)", b) is None # no leftover dual-unit strings
|
||
|
||
|
||
def test_units_default_to_the_citys_country(client):
|
||
# The search snippet shows whatever the server rendered, and impressions skew
|
||
# to °C countries — so a UK city page must say °C in the HTML itself, with no
|
||
# JS and no stored preference. data-temp-f stays Fahrenheit either way: it is
|
||
# what climate.js converts from.
|
||
gb = client.get(f"{B}/climate/{SLUG}").text # London, GB
|
||
us = client.get(f"{B}/climate/seattle-washington-us").text
|
||
|
||
gb_spans, us_spans = _temp_spans(gb), _temp_spans(us)
|
||
assert gb_spans and us_spans
|
||
assert {letter for _f, _n, letter in gb_spans if letter} == {"C"}
|
||
assert {letter for _f, _n, letter in us_spans if letter} == {"F"}
|
||
|
||
# The attribute is still Fahrenheit, and really does differ from the Celsius
|
||
# text — this is what breaks if someone "helpfully" converts the attribute too.
|
||
assert any(math.floor(f + 0.5) != n for f, n, _l in gb_spans), "data-temp-f was converted"
|
||
# On a °F page the attribute and the text agree by definition. Rounded half-up,
|
||
# not with Python's round(): the page renders the number JS would.
|
||
assert all(math.floor(f + 0.5) == n for f, n, _l in us_spans)
|
||
|
||
assert 'data-unit-default="C"' in gb
|
||
assert 'data-unit-default="F"' in us
|
||
|
||
|
||
_PRECIP_SPAN = re.compile(r'<span class="precip" data-precip-in="([\d.]+)">([\d.]+) (in|mm)</span>')
|
||
|
||
|
||
def test_precip_and_wind_follow_the_citys_unit(client):
|
||
# One toggle drives every measure: a °C page shows mm and km/h, not °C mixed
|
||
# with inches and mph. The data-* attribute stays imperial in both, so
|
||
# climate.js converts from the same source of truth it does for temperature.
|
||
gb = client.get(f"{B}/climate/{SLUG}/records").text # London, GB
|
||
us = client.get(f"{B}/climate/seattle-washington-us/records").text
|
||
|
||
gb_p, us_p = _PRECIP_SPAN.findall(gb), _PRECIP_SPAN.findall(us)
|
||
assert gb_p and us_p
|
||
|
||
assert {u for _raw, _shown, u in gb_p} == {"mm"}
|
||
assert {u for _raw, _shown, u in us_p} == {"in"}
|
||
|
||
# The conversion is real, not just a relabel.
|
||
for raw, shown, _u in gb_p:
|
||
assert abs(float(shown) - float(raw) * 25.4) < 0.51, (raw, shown)
|
||
# On an imperial page the attribute and the text agree.
|
||
for raw, shown, _u in us_p:
|
||
assert abs(float(shown) - float(raw)) < 0.005, (raw, shown)
|
||
|
||
|
||
def test_wind_and_humidity_render_per_unit_system():
|
||
# The synthetic history carries only tmax/tmin/precip, so wind and humidity
|
||
# never reach a rendered page in tests — exercise the renderers directly
|
||
# rather than assert on a page that structurally cannot show them.
|
||
from web import content
|
||
with content._unit_scope("F"):
|
||
assert content._wind_text(10) == "10 mph"
|
||
assert '<span class="wind" data-wind-mph="10.0">10 mph</span>' == content._wind(10)
|
||
assert content._fmt("humid", 7.75) == "7.8 g/m³"
|
||
with content._unit_scope("C"):
|
||
# 10 mph -> 16.09 km/h, rounded half-up like JS.
|
||
assert content._wind_text(10) == "16 km/h"
|
||
# data-wind-mph stays imperial: it is what climate.js converts from.
|
||
assert 'data-wind-mph="10.0"' in content._wind(10)
|
||
# Absolute humidity has no imperial equivalent anyone would recognise, so
|
||
# it reads the same in both systems.
|
||
assert content._fmt("humid", 7.75) == "7.8 g/m³"
|
||
|
||
|
||
def test_precip_precision_differs_by_unit():
|
||
# Rainfall is reported in whole millimetres; a tenth of a mm is below what
|
||
# the source resolves. Inches keep two decimals.
|
||
from web import content
|
||
with content._unit_scope("F"):
|
||
assert content._precip_text(0.04) == "0.04 in"
|
||
with content._unit_scope("C"):
|
||
assert content._precip_text(0.04) == "1 mm"
|
||
assert content._precip_text(1.81) == "46 mm"
|
||
|
||
|
||
def test_measure_conversion_constants_match_the_frontend():
|
||
# Same drift risk as the country list: two hand-maintained copies.
|
||
import os
|
||
from web import content
|
||
units_js = os.path.join(os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__)))),
|
||
os.pardir, "frontend", "units.js")
|
||
with open(units_js, encoding="utf-8") as f:
|
||
src = f.read()
|
||
mm = float(re.search(r"MM_PER_IN\s*=\s*([\d.]+)", src).group(1))
|
||
kmh = float(re.search(r"KMH_PER_MPH\s*=\s*([\d.]+)", src).group(1))
|
||
assert mm == content.MM_PER_IN
|
||
assert kmh == content.KMH_PER_MPH
|
||
|
||
|
||
def test_unit_default_absent_without_a_city(client):
|
||
# No city -> no attribute -> units.js keeps today's locale behaviour. Pinning
|
||
# these to "F" would silently regress the interactive pages.
|
||
for path in (f"{B}/", f"{B}/climate", f"{B}/about", f"{B}/glossary"):
|
||
assert "data-unit-default" not in client.get(path).text, path
|
||
|
||
|
||
def test_unit_does_not_leak_between_requests(client):
|
||
# The page handlers are sync `def`, so Starlette runs them on a recycled
|
||
# threadpool thread: a unit left set would bleed into the next request served
|
||
# by that thread. Interleave and re-check rather than trusting the reset.
|
||
assert 'data-unit-default="C"' in client.get(f"{B}/climate/{SLUG}").text
|
||
about = client.get(f"{B}/about").text
|
||
assert "data-unit-default" not in about
|
||
# about.html.j2 renders an illustrative temperature; it must stay °F.
|
||
assert {l for _f, _n, l in _temp_spans(about) if l} == {"F"}
|
||
|
||
us = client.get(f"{B}/climate/seattle-washington-us").text
|
||
assert {l for _f, _n, l in _temp_spans(us) if l} == {"F"}
|
||
gb_again = client.get(f"{B}/climate/{SLUG}").text
|
||
assert {l for _f, _n, l in _temp_spans(gb_again) if l} == {"C"}
|
||
|
||
|
||
def test_f_country_list_matches_the_frontend():
|
||
# Two hand-maintained copies of the same 14 codes; nothing else stops them
|
||
# drifting apart, and drift means the server and the client disagree about
|
||
# what a first-time visitor should see.
|
||
import os
|
||
from web import content
|
||
units_js = os.path.join(os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__)))),
|
||
os.pardir, "frontend", "units.js")
|
||
with open(units_js, encoding="utf-8") as f:
|
||
src = f.read()
|
||
literal = re.search(r"F_REGIONS = new Set\(\[(.*?)\]\)", src, re.S).group(1)
|
||
assert set(re.findall(r'"([A-Z]{2})"', literal)) == set(content.F_COUNTRIES)
|
||
|
||
|
||
def test_celsius_rounding_matches_js():
|
||
# Python's round() goes to even, JS's Math.round goes up. Where they disagree
|
||
# the server prints one number and climate.js repaints another — a flicker for
|
||
# the visitor whose unit already matched.
|
||
from web import content
|
||
assert content._round_half_up(0.5) == 1 # round(0.5) would give 0
|
||
assert content._round_half_up(1.5) == 2
|
||
assert content._round_half_up(2.5) == 3 # round(2.5) would give 2
|
||
assert content._round_half_up(-0.5) == 0 # Math.round(-0.5) === 0
|
||
# 31.1°F -> -0.5°C exactly: the case that actually shows up on a cold page.
|
||
assert content._c(31.1) == 0
|
||
|
||
|
||
def test_city_page_etag_is_stable(client):
|
||
# The unit comes from the URL's city, never from a request header, so the
|
||
# response stays a pure function of the URL — no Vary, and the ETag still works.
|
||
r1 = client.get(f"{B}/climate/{SLUG}")
|
||
r2 = client.get(f"{B}/climate/{SLUG}")
|
||
assert r1.headers["etag"] == r2.headers["etag"]
|
||
r3 = client.get(f"{B}/climate/{SLUG}", headers={"If-None-Match": r1.headers["etag"]})
|
||
assert r3.status_code == 304
|
||
|
||
|
||
def test_seo_pages_load_unit_and_account_scripts(client):
|
||
# Header parity with the interactive pages: the °F/°C toggle, account/bell, and
|
||
# the temperature converter are all loaded.
|
||
b = client.get(f"{B}/climate/{SLUG}").text
|
||
for src in (f"{B}/units.js", f"{B}/account.js", f"{B}/climate.js"):
|
||
assert f'src="{src}"' in b
|
||
|
||
|
||
def test_indexnow_key_file_served(client):
|
||
import indexnow
|
||
k = indexnow.key()
|
||
r = client.get(f"{B}/{k}.txt")
|
||
assert r.status_code == 200
|
||
assert r.text.strip() == k
|
||
assert r.headers["content-type"].startswith("text/plain")
|
||
|
||
|
||
def test_sitemap_lastmod_is_stable(client):
|
||
import datetime
|
||
import re
|
||
body = client.get(f"{B}/sitemap.xml").text
|
||
mods = set(re.findall(r"<lastmod>([^<]+)</lastmod>", body))
|
||
assert len(mods) == 1 # one build date, not a per-request churn
|
||
datetime.date.fromisoformat(next(iter(mods))) # a valid ISO date
|
||
|
||
|
||
def test_indexnow_submit_all_builds_payload(monkeypatch):
|
||
import indexnow
|
||
|
||
calls = []
|
||
|
||
class _Resp:
|
||
status_code = 200
|
||
text = "ok"
|
||
|
||
monkeypatch.setattr(indexnow.httpx, "post",
|
||
lambda url, json=None, **kw: (calls.append((url, json)), _Resp())[1])
|
||
res = indexnow.submit_all("https://thermograph.org")
|
||
assert res["submitted"] == res["total"] > 100
|
||
first = calls[0][1]
|
||
assert first["host"] == "thermograph.org"
|
||
assert first["keyLocation"] == f"https://thermograph.org/{indexnow.key()}.txt"
|
||
assert first["urlList"][0] == "https://thermograph.org/"
|
||
assert all(len(c[1]["urlList"]) <= 10000 for c in calls) # batched under the IndexNow cap
|
||
|
||
|
||
def test_indexnow_if_changed_state(monkeypatch, tmp_path):
|
||
import indexnow
|
||
monkeypatch.setattr(indexnow, "_STATE_PATH", str(tmp_path / "state.txt"))
|
||
sig = indexnow.url_signature()
|
||
assert len(sig) == 40 and sig == indexnow.url_signature() # stable sha1 of the URL set
|
||
assert indexnow._read_state() == "" # first deploy → would submit
|
||
indexnow._write_state(sig)
|
||
assert indexnow._read_state() == sig # unchanged → deploy would skip
|
||
|
||
|
||
def test_search_verification_meta(client, monkeypatch):
|
||
monkeypatch.setenv("THERMOGRAPH_GOOGLE_VERIFY", "gtok123")
|
||
monkeypatch.setenv("THERMOGRAPH_BING_VERIFY", "btok456")
|
||
seo = client.get(f"{B}/climate/{SLUG}").text # Jinja SEO page
|
||
home = client.get(f"{B}/").text # static homepage via _page()
|
||
for b in (seo, home):
|
||
assert '<meta name="google-site-verification" content="gtok123">' in b
|
||
assert '<meta name="msvalidate.01" content="btok456">' in b
|
||
|
||
|
||
# The brand lockup is the site-wide "go home" affordance. The static tool pages
|
||
# and the Jinja pages build their headers separately, so this is easy to add to
|
||
# one and forget on the other — assert every page carries it.
|
||
BRAND_PAGES = ["/", "/calendar", "/day", "/compare", "/legend", "/alerts",
|
||
"/about", "/privacy", "/climate", "/glossary"]
|
||
|
||
|
||
@pytest.mark.parametrize("path", BRAND_PAGES)
|
||
def test_brand_lockup_links_home(client, path):
|
||
html = client.get(f"{B}{path}").text
|
||
brand = html.split('<div class="brand">', 1)[1].split("</header>", 1)[0]
|
||
head = brand.split("</h1>")[0] if "<h1" in brand else brand.split("</p>")[0]
|
||
# The link must wrap the whole lockup, so clicking the mark works too, not
|
||
# just the word.
|
||
assert "<a href=" in head, f"{path}: brand is not a link"
|
||
href = head.split('<a href="', 1)[1].split('"', 1)[0]
|
||
assert href in ("./", f"{B}/"), f"{path}: brand links to {href!r}, not home"
|
||
assert '<span class="logo">' in head.split("<a href=", 1)[1], \
|
||
f"{path}: the mark sits outside the home link"
|
||
|
||
|
||
# Percentiles are shown on the Day page, the calendar tooltip, the chart, the
|
||
# city pages and the homepage strip. They all have to say the same thing about
|
||
# the same reading, so the rule lives in grading.pct_ordinal() and every surface
|
||
# defers to it (the frontend mirrors it as pctOrd in shared.js).
|
||
def test_pct_ordinal_rule():
|
||
assert grading.pct_ordinal(66.4) == "66th"
|
||
assert grading.pct_ordinal(1.2) == "1st"
|
||
assert grading.pct_ordinal(22) == "22nd"
|
||
assert grading.pct_ordinal(3) == "3rd"
|
||
assert grading.pct_ordinal(11) == "11th"
|
||
# Floored into 1..99: an empirical percentile is a rank against the sample,
|
||
# so "100th percentile" reads as a measurement error.
|
||
assert grading.pct_ordinal(99.6) == "99th"
|
||
assert grading.pct_ordinal(100) == "99th"
|
||
assert grading.pct_ordinal(0) == "1st"
|
||
assert grading.pct_ordinal(None) == "—"
|
||
|
||
|
||
def test_city_page_percentiles_are_real_ordinals(client):
|
||
"""The city page used to render `{{ c.percentile }}th pct` against a float,
|
||
producing "97.1th pct" and "1th pct" on every one of the ~1000 city pages."""
|
||
html = client.get(f"{B}/climate/{SLUG}").text
|
||
assert "th pct" in html or "st pct" in html or "nd pct" in html or "rd pct" in html
|
||
# No float ordinals, no impossible ones, no "1th"/"2th"/"3th".
|
||
for bad in re.findall(r"\d+\.\d+(?:st|nd|rd|th)|100th|\b0th|\b\d*[123]th", html):
|
||
raise AssertionError(f"malformed percentile ordinal: {bad!r}")
|