feat(tools): replace xxm fan import with bilibili csv

This commit is contained in:
kairee committed 2026-11-01 02:23:09 +08:00
1 parent 58df38814a
commit c4c9c0b038
4 files changed
+392 -469

No files matched your search

+132
View File
@@ -0,0 +1,132 @@
# -*- coding: utf-8 -*-
"""粉丝历史 CSV 回填测试(离线): uv run python tests/test_fans_csv.py"""
import os
import sys
import tempfile
from pathlib import Path
sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
sys.stdout.reconfigure(encoding="utf-8")
from reaction_board.database import Database # noqa: E402
from reaction_board.repository import Repository # noqa: E402
from tools.import_fans_csv import ( # noqa: E402
import_fans_series,
parse_fans_csv,
)
CSV_TEXT = """\ufeff时间,粉丝,7日变化,1日变化
2026-01-03 10:00,1003,0,1
2026-01-01 09:00,1001,0,0
2026-01-02 08:30,1002,0,1
2026-01-02 23:30,1005,0,4
2026-01-04,1007,0,1
"""
def test_parse_csv():
series = parse_fans_csv(CSV_TEXT)
assert series["first_date"] == "2026-01-01"
assert series["last_date"] == "2026-01-04"
assert [point["date"] for point in series["points"]] == [
"2026-01-01", "2026-01-02", "2026-01-03", "2026-01-04",
]
# 同一天两行,保留时间较晚的一条
assert [point["follower"] for point in series["points"]] == [1001, 1005, 1003, 1007]
assert series["unique_days"] == 4
assert series["total_rows"] == 5
assert series["gaps"] == 0
assert any("多行" in warning for warning in series["warnings"])
for bad in ("", "时间,粉丝\n2026-01-01\n", "date,follower\nx,1\n", "时间,粉丝\n"):
try:
parse_fans_csv(bad)
except ValueError:
pass
else:
raise AssertionError(f"非法 CSV 应报错: {bad!r}")
def test_parse_rejects_bad_values():
for bad in (
"时间,粉丝\n2026-01-01,nope\n",
"时间,粉丝\nnot-a-date,100\n",
"时间,粉丝\n2026-01-01,-1\n",
):
try:
parse_fans_csv(bad)
except ValueError:
pass
else:
raise AssertionError(f"非法行应报错: {bad!r}")
def test_import_default_force_and_replace():
series = parse_fans_csv(CSV_TEXT)
with tempfile.TemporaryDirectory() as tmp:
db = Database(Path(tmp) / "app.db")
db.initialize()
try:
repo = Repository(db)
repo.upsert_up_stat("2026-01-01", 37754047, follower=1, following=9, ok=True, captured_at=1)
repo.upsert_up_stat("2026-01-02", 37754047, follower=None, following=None, ok=False, captured_at=1)
repo.upsert_up_stat("2026-01-03", 99999, follower=1, following=9, ok=True, captured_at=1)
dry = import_fans_series(db, series, dry_run=True)
assert dry["inserted"] == 2
assert dry["updated"] == 1
assert dry["skipped"] == 1
assert repo.get_up_stat("2026-01-02", 37754047)["ok"] == 0
summary = import_fans_series(db, series)
assert summary["inserted"] == 2
assert summary["updated"] == 1
assert summary["skipped"] == 1
assert repo.get_up_stat("2026-01-01", 37754047)["follower"] == 1
assert repo.get_up_stat("2026-01-02", 37754047)["follower"] == 1005
assert repo.get_up_stat("2026-01-02", 37754047)["ok"] == 1
assert repo.get_up_stat("2026-01-04", 37754047)["follower"] == 1007
# 另一个 mid 不受影响
assert repo.get_up_stat("2026-01-03", 99999)["follower"] == 1
forced = import_fans_series(db, series, force=True)
assert forced["updated"] == 4
assert repo.get_up_stat("2026-01-01", 37754047)["follower"] == 1001
replaced = import_fans_series(db, series, replace=True)
assert replaced["deleted"] == 4
assert replaced["inserted"] == 4
assert replaced["updated"] == 0
assert repo.get_up_stat("2026-01-01", 37754047)["follower"] == 1001
assert repo.get_up_stat("2026-01-03", 99999)["follower"] == 1
finally:
db.close()
def test_replace_dry_run_counts_only():
series = parse_fans_csv(CSV_TEXT)
with tempfile.TemporaryDirectory() as tmp:
db = Database(Path(tmp) / "app.db")
db.initialize()
try:
repo = Repository(db)
repo.upsert_up_stat("2026-01-01", 37754047, follower=1, following=None, ok=True, captured_at=1)
result = import_fans_series(db, series, replace=True, dry_run=True)
assert result["deleted"] == 1
assert result["inserted"] == 4
assert repo.get_up_stat("2026-01-01", 37754047)["follower"] == 1
finally:
db.close()
def main_test():
test_parse_csv()
test_parse_rejects_bad_values()
test_import_default_force_and_replace()
test_replace_dry_run_counts_only()
print("OK: 粉丝 CSV 解析/导入/替换/幂等测试全部通过")
if __name__ == "__main__":
main_test()
-174
View File
@@ -1,174 +0,0 @@
# -*- coding: utf-8 -*-
"""粉丝历史回填测试(离线): uv run python tests/test_fans_history.py"""
import json
import os
import sys
import tempfile
from datetime import date, timedelta
from pathlib import Path
sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
sys.stdout.reconfigure(encoding="utf-8")
from reaction_board.database import Database # noqa: E402
from reaction_board.repository import Repository # noqa: E402
from tools.import_fans_history import ( # noqa: E402
build_follower_series,
import_follower_series,
parse_data_js,
)
def make_payload(*, days=10, gap_day=None, mismatch=False):
start = date(2026, 1, 1)
dates = [start + timedelta(days=index) for index in range(days)]
anchor = dates[-1]
# 从 900 起每天 +10,最后一天为 900 + (days-1)*10
values = {day: 900 + index * 10 for index, day in enumerate(dates)}
corr = []
for index, day in enumerate(dates):
delta = 0 if index == 0 else values[day] - values[dates[index - 1]]
corr.append([day.isoformat(), 0, delta])
if gap_day is not None:
corr = [row for row in corr if row[0] != gap_day]
week = [
[day.strftime("%Y/%m/%d"), values[day], 0]
for day in dates[-3:]
]
if mismatch:
week[-1][1] += 7
payload = {
"meta": {"updated": anchor.isoformat()},
"followers": {
"咻咻满": {
"total": values[anchor],
"week": week,
}
},
"corr": corr,
}
return payload, values
def test_parse_data_js():
payload, _ = make_payload()
text = "/* demo */\nwindow.XXM = " + json.dumps(payload, ensure_ascii=False) + ";\n"
parsed = parse_data_js(text)
assert parsed["meta"]["updated"] == "2026-01-10"
for bad in ("", "<html></html>", "window.XXM = {}"):
try:
parse_data_js(bad)
except ValueError:
pass
else:
raise AssertionError(f"非法输入应报错: {bad!r}")
def test_build_follower_series():
payload, values = make_payload(days=10)
series = build_follower_series(payload, days=10)
assert series["anchor_follower"] == values[date(2026, 1, 10)]
assert series["first_date"] == "2026-01-01"
assert series["last_date"] == "2026-01-10"
assert series["points"][0] == {"date": "2026-01-01", "follower": 900}
assert series["points"][-1] == {"date": "2026-01-10", "follower": 990}
assert series["warnings"] == []
assert series["stopped_at_gap"] is False
limited = build_follower_series(payload, days=3)
assert [point["date"] for point in limited["points"]] == [
"2026-01-08", "2026-01-09", "2026-01-10",
]
try:
build_follower_series(payload, days=0)
except ValueError:
pass
else:
raise AssertionError("days=0 应报错")
def test_gap_and_mismatch():
payload, _ = make_payload(days=10, gap_day="2026-01-05")
series = build_follower_series(payload, days=10)
assert series["stopped_at_gap"] is True
assert series["first_date"] == "2026-01-06"
assert series["last_date"] == "2026-01-10"
assert any("不连续" in warning for warning in series["warnings"])
payload, _ = make_payload(days=10, mismatch=True)
try:
build_follower_series(payload, days=10)
except ValueError as exc:
assert "不一致" in str(exc)
else:
raise AssertionError("week 与 corr 不一致时应报错")
def test_import_and_force():
payload, values = make_payload(days=10)
series = build_follower_series(payload, days=10)
with tempfile.TemporaryDirectory() as tmp:
db = Database(Path(tmp) / "app.db")
db.initialize()
try:
repo = Repository(db)
repo.upsert_up_stat(
"2026-01-08", 37754047, follower=1, following=5,
ok=True, captured_at=1,
)
repo.upsert_up_stat(
"2026-01-09", 37754047, follower=None, following=None,
ok=False, captured_at=1,
)
repo.upsert_up_stat(
"2026-01-10", 37754047, follower=2, following=None,
ok=True, captured_at=1,
)
dry = import_follower_series(
db, series, mid=37754047, dry_run=True
)
assert dry["inserted"] == 7
assert dry["updated"] == 1
assert dry["skipped"] == 2
assert repo.get_up_stat("2026-01-09", 37754047)["ok"] == 0
summary = import_follower_series(db, series, mid=37754047)
assert summary["inserted"] == 7
assert summary["updated"] == 1
assert summary["skipped"] == 2
assert repo.get_up_stat("2026-01-01", 37754047)["follower"] == 900
row = repo.get_up_stat("2026-01-09", 37754047)
assert row["follower"] == values[date(2026, 1, 9)]
assert row["ok"] == 1
assert row["following"] is None
preserved = repo.get_up_stat("2026-01-08", 37754047)
assert preserved["follower"] == 1
assert preserved["following"] == 5
again = import_follower_series(db, series, mid=37754047)
assert again["inserted"] == 0
assert again["skipped"] == 10
forced = import_follower_series(
db, series, mid=37754047, force=True
)
assert forced["updated"] == 10
assert repo.get_up_stat("2026-01-08", 37754047)["follower"] == values[date(2026, 1, 8)]
finally:
db.close()
def main_test():
test_parse_data_js()
test_build_follower_series()
test_gap_and_mismatch()
test_import_and_force()
print("OK: 粉丝历史解析/回填/缺口/幂等测试全部通过")
if __name__ == "__main__":
main_test()