election-game/backend/data/pipeline/tests/test_races2024.py

128 lines
5.0 KiB
Python
Raw Normal View History

import json
import pytest
from election_data.races2024 import (
OUT_LEGISLATOR,
OUT_PRESIDENT,
district_towns,
legislator_cap,
president_cap,
)
def test_president_cap_matches_published_2020_figure() -> None:
# 2019-07-31 人口 23,592,598 → 中選會公告 2020 總統選舉上限 430,297,000
assert president_cap(23_592_598) == 430_297_000
def test_legislator_cap_formula() -> None:
assert legislator_cap(14_005) == 10_295_000
@pytest.mark.skipif(not (OUT_PRESIDENT / "00000.json").exists(), reason="需要先執行 election-data races2024")
def test_president_totals_match_cec() -> None:
data = json.loads((OUT_PRESIDENT / "00000.json").read_text(encoding="utf-8"))
total = sum(sum(v["votes"].values()) for v in data["byVillage"].values())
assert total == data["cecValid"] == 13_947_506
assert [c["name"] for c in data["candidates"]] == ["柯文哲", "賴清德", "侯友宜"]
assert all(c["finance"] for c in data["candidates"])
@pytest.mark.skipif(not OUT_LEGISLATOR.exists(), reason="需要先執行 election-data races2024")
def test_legislator_districts() -> None:
# calibration.json 是校準結果,不是選區
files = sorted(OUT_LEGISLATOR.glob("[0-9]*-[0-9]*.json"))
assert len(files) == 73
seen: set[str] = set()
for path in files:
data = json.loads(path.read_text(encoding="utf-8"))
villages = set(data["byVillage"])
assert not villages & seen, path.name # 每個里只屬於一個選區
seen |= villages
numbers = {f"#{c['number']}" for c in data["candidates"]}
for v in data["byVillage"].values():
assert set(v["votes"]) <= numbers
def test_district_towns_marks_partial_towns() -> None:
sites = {
"63000110001": "臺北市士林區",
"63000110002": "臺北市士林區",
"63000120001": "臺北市北投區",
}
towns = district_towns({"63000110001", "63000120001"}, sites, "臺北市")
assert towns == [
{"code": "63000110", "name": "士林區", "partial": True},
{"code": "63000120", "name": "北投區", "partial": False},
]
@pytest.mark.skipif(
not (OUT_LEGISLATOR / "63000-01.json").exists(), reason="需要先執行 election-data races2024"
)
def test_taipei_first_district_towns() -> None:
data = json.loads((OUT_LEGISLATOR / "63000-01.json").read_text(encoding="utf-8"))
assert [(t["name"], t["partial"]) for t in data["towns"]] == [("士林區", True), ("北投區", False)]
class _FakeIndex:
def __init__(self, edges: dict[str, list[str]]) -> None:
self._edges = edges
def neighbors(self, code: str) -> list[str]:
return self._edges.get(code, [])
def _shape(code: str):
from types import SimpleNamespace
return SimpleNamespace(code=code, county_code=code[:5])
def test_assign_orphans_prefers_same_town_and_spreads_outward() -> None:
from election_data.races2024 import assign_orphans
district_of = {"63000110001": "A", "63000120001": "B", "63000120002": "B"}
edges = {
# 新里 003 同時鄰接士林(A)與北投(B),同區的鄰居優先
"63000120003": ["63000110001", "63000120001"],
# 新里 004 只鄰接另一個新里,要等 003 先分好
"63000120004": ["63000120003"],
}
shapes = [_shape(c) for c in [*district_of, "63000120003", "63000120004"]]
assert assign_orphans(district_of, shapes, _FakeIndex(edges)) == {
"63000120003": "B",
"63000120004": "B",
}
@pytest.mark.skipif(not OUT_LEGISLATOR.exists(), reason="需要先執行 election-data races2024")
def test_every_pack_village_is_in_a_district() -> None:
by_county: dict[str, set[str]] = {}
for path in OUT_LEGISLATOR.glob("[0-9]*-[0-9]*.json"):
data = json.loads(path.read_text(encoding="utf-8"))
by_county.setdefault(data["countyCode"], set()).update(data["byVillage"])
packs = OUT_LEGISLATOR.parent.parent / "packs"
for county, villages in by_county.items():
pack = packs / f"{county}-before-2024-01-13" / "villages.json"
if not pack.exists():
continue
codes = {v["code"] for v in json.loads(pack.read_text(encoding="utf-8"))["villages"]}
assert codes <= villages, (county, sorted(codes - villages)[:5])
def test_remap_population_matches_renamed_offices_by_village_name() -> None:
from election_data.population import VillagePopulation
from election_data.races2024 import remap_population
def vp(code: str, site: str, village: str, total: int) -> VillagePopulation:
return VillagePopulation(code, site, village, total, (0,) * 101)
old = [
vp("64000051001", "高雄市三民一", "鼎金里", 100), # 2019 年三民區分兩個戶政所
vp("64000010001", "高雄市鹽埕區", "藍橋里", 50),
]
now = [vp("64000050007", "高雄市三民區", "鼎金里", 0), vp("64000010001", "高雄市鹽埕區", "藍橋里", 0)]
assert remap_population(old, now) == {"64000050007": 100, "64000010001": 50}