butubb
20e672834c
两件都是「作品栏里把这东西认出来」的延伸： * **账号级指标**：作品列表只会说「这条涨了多少赞」，说不了「这个人整个 账号的粉丝在涨还是在掉」。抖音的资料接口本来就有粉丝数/总获赞/作品数， 每轮顺手记一条快照（`monitor_creator_stat`，粒度 = 任务×博主×轮次， 和作品指标同形）。组头显示最近一条。 快照在「一条作品都没采到」的早退**之前**落：作品列表被风控挡住的那一轮， 正是「粉丝还在涨、但新作品没在发现」最该被看见的时刻。 * **作品备注**：博主备注回答「这个账号是谁」，这条回答「这条我要盯着」。 一个博主底下常常只有一两件值得盯的作品，所以不能合并成一条。键取 (platform, note_id)，跨任务共用一份。 两边都守住同一条口径：**不知道就是 null，不写成 0** —— 0 在趋势图上是一条 砸到底的线，和「还没采到」是两回事。
221 lines
7.8 KiB
Python
221 lines
7.8 KiB
Python
# -*- coding: utf-8 -*-
|
||
"""作品栏里的**博主**：备注（他到底是谁）与账号级指标（他现在多大）。
|
||
|
||
按 creator_hash 分组、显示 creator_name，两样都认不出人：一个是哈希，一个是平台昵称。
|
||
备注是人自己起的名字。账号级指标则是作品列表给不了的东西 —— 作品说的是"这条涨了多少赞"，
|
||
粉丝数说的是"这个人整个账号在涨还是在掉"。
|
||
"""
|
||
|
||
import httpx
|
||
import pytest
|
||
import pytest_asyncio
|
||
|
||
from api.main import app
|
||
from api.monitor import db as monitor_db
|
||
from api.monitor.models import (
|
||
MODE_CREATOR,
|
||
MonitorCreatorStat,
|
||
MonitorNote,
|
||
MonitorTask,
|
||
)
|
||
|
||
CREATOR_HASH = "hash-a"
|
||
NICKNAME = "张三"
|
||
|
||
|
||
@pytest_asyncio.fixture
|
||
async def client(tmp_path):
|
||
monitor_db.set_sqlite_path(tmp_path / "monitor.db")
|
||
await monitor_db.init_db()
|
||
transport = httpx.ASGITransport(app=app)
|
||
async with httpx.AsyncClient(transport=transport, base_url="http://test") as http_client:
|
||
yield http_client
|
||
await monitor_db.dispose_engine()
|
||
|
||
|
||
async def _seed(platform: str = "xhs", task_name: str = "任务") -> int:
|
||
"""一个任务 + 一条作品，博主固定用 CREATOR_HASH / NICKNAME。"""
|
||
async with monitor_db.get_session() as session:
|
||
task = MonitorTask(
|
||
name=task_name, platform=platform, mode=MODE_CREATOR, enabled=True,
|
||
interval_minutes=60, max_notes_count=20, enable_comments=False,
|
||
max_comments_count=50, run_timeout_seconds=3600,
|
||
notify_enabled=False, created_at=0, updated_at=0,
|
||
)
|
||
session.add(task)
|
||
await session.flush()
|
||
session.add(
|
||
MonitorNote(
|
||
task_id=task.id, note_id=f"{platform}-n1", title="作品",
|
||
note_url="", cover="", creator_hash=CREATOR_HASH,
|
||
creator_name=NICKNAME, source_kind="", published_at=None,
|
||
first_seen_run_id=1, first_seen_at=0,
|
||
last_seen_run_id=1, last_seen_at=0,
|
||
)
|
||
)
|
||
return task.id
|
||
|
||
|
||
async def _notes(client, platform: str = "xhs"):
|
||
return (await client.get("/api/monitor/notes", params={"platform": platform})).json()["notes"]
|
||
|
||
|
||
async def _set_alias(client, alias: str, platform: str = "xhs", creator_hash=CREATOR_HASH):
|
||
return await client.put(
|
||
f"/api/monitor/creators/{creator_hash}",
|
||
params={"platform": platform},
|
||
json={"alias": alias},
|
||
)
|
||
|
||
|
||
class TestCreatorAlias:
|
||
@pytest.mark.asyncio
|
||
async def test_notes_start_without_an_alias(self, client):
|
||
await _seed()
|
||
|
||
assert (await _notes(client))[0]["creator_alias"] == ""
|
||
|
||
@pytest.mark.asyncio
|
||
async def test_an_alias_comes_back_with_the_notes(self, client):
|
||
await _seed()
|
||
|
||
response = await _set_alias(client, "竞品A")
|
||
|
||
assert response.status_code == 200
|
||
note = (await _notes(client))[0]
|
||
assert note["creator_alias"] == "竞品A"
|
||
# 备注是**叠加**在昵称之上的，不是替换 —— 昵称仍然是有用的对照。
|
||
assert note["creator_name"] == NICKNAME
|
||
|
||
@pytest.mark.asyncio
|
||
async def test_an_alias_is_shared_across_tasks(self, client):
|
||
"""同一个博主出现在两个任务里，备注只该填一次。
|
||
|
||
creator_hash 对同一个 uid 是稳定的，所以键取 (platform, creator_hash) 而不是
|
||
按任务存 —— 否则每加一个任务都要重新认一遍人。
|
||
"""
|
||
await _seed(task_name="任务甲")
|
||
await _seed(task_name="任务乙")
|
||
await _set_alias(client, "竞品A")
|
||
|
||
for note in await _notes(client):
|
||
assert note["creator_alias"] == "竞品A"
|
||
|
||
@pytest.mark.asyncio
|
||
async def test_the_alias_does_not_leak_to_another_platform(self, client):
|
||
"""同一个哈希在另一个平台上是另一个（或同一个）人 —— 别串味。"""
|
||
await _seed(platform="xhs")
|
||
await _seed(platform="dy")
|
||
|
||
await _set_alias(client, "小红书那边的", platform="xhs")
|
||
|
||
assert (await _notes(client, "xhs"))[0]["creator_alias"] == "小红书那边的"
|
||
assert (await _notes(client, "dy"))[0]["creator_alias"] == ""
|
||
|
||
@pytest.mark.asyncio
|
||
async def test_an_empty_alias_clears_it(self, client):
|
||
await _seed()
|
||
await _set_alias(client, "竞品A")
|
||
|
||
await _set_alias(client, "")
|
||
|
||
assert (await _notes(client))[0]["creator_alias"] == ""
|
||
|
||
@pytest.mark.asyncio
|
||
async def test_an_alias_is_trimmed(self, client):
|
||
await _seed()
|
||
|
||
await _set_alias(client, " 竞品A ")
|
||
|
||
assert (await _notes(client))[0]["creator_alias"] == "竞品A"
|
||
|
||
|
||
async def _add_stat(
|
||
task_id: int,
|
||
run_id: int,
|
||
fans: int | None,
|
||
*,
|
||
total_favorited: int | None = 83000,
|
||
works: int | None = 42,
|
||
creator_hash: str = CREATOR_HASH,
|
||
captured_at: int = 1,
|
||
) -> None:
|
||
async with monitor_db.get_session() as session:
|
||
session.add(
|
||
MonitorCreatorStat(
|
||
task_id=task_id,
|
||
run_id=run_id,
|
||
creator_hash=creator_hash,
|
||
nickname=NICKNAME,
|
||
fans=fans,
|
||
total_favorited=total_favorited,
|
||
works_count=works,
|
||
following=7,
|
||
captured_at=captured_at,
|
||
)
|
||
)
|
||
|
||
|
||
class TestCreatorStats:
|
||
"""账号级指标跟着作品一起返回 —— 界面上是按博主归组的，为了一个组头再发一轮请求
|
||
没道理。"""
|
||
|
||
@pytest.mark.asyncio
|
||
async def test_without_snapshots_the_fields_are_null(self, client):
|
||
"""null 而不是 0：0 会显示成「粉丝 0」，而事实是"还没采到"。"""
|
||
await _seed()
|
||
|
||
note = (await _notes(client))[0]
|
||
|
||
assert note["creator_fans"] is None
|
||
assert note["creator_total_favorited"] is None
|
||
assert note["creator_works"] is None
|
||
|
||
@pytest.mark.asyncio
|
||
async def test_a_snapshot_shows_up_on_every_work_of_that_creator(self, client):
|
||
task_id = await _seed()
|
||
await _add_stat(task_id, run_id=1, fans=12000)
|
||
|
||
for note in await _notes(client):
|
||
assert note["creator_fans"] == 12000
|
||
assert note["creator_total_favorited"] == 83000
|
||
assert note["creator_works"] == 42
|
||
|
||
@pytest.mark.asyncio
|
||
async def test_the_latest_run_wins(self, client):
|
||
"""一轮一条，所以总会有好几条 —— 给界面的必须是最近那条。"""
|
||
task_id = await _seed()
|
||
await _add_stat(task_id, run_id=1, fans=12000)
|
||
await _add_stat(task_id, run_id=2, fans=12300)
|
||
|
||
assert (await _notes(client))[0]["creator_fans"] == 12300
|
||
|
||
@pytest.mark.asyncio
|
||
async def test_another_platforms_snapshot_does_not_leak(self, client):
|
||
"""两个平台上恰好同名同哈希的博主是两个人 —— 快照挂在任务上，不该串。"""
|
||
await _seed(platform="xhs")
|
||
dy_task_id = await _seed(platform="dy")
|
||
await _add_stat(dy_task_id, run_id=1, fans=999)
|
||
|
||
assert (await _notes(client, "xhs"))[0]["creator_fans"] is None
|
||
assert (await _notes(client, "dy"))[0]["creator_fans"] == 999
|
||
|
||
@pytest.mark.asyncio
|
||
async def test_an_unparsed_count_stays_null(self, client):
|
||
"""快照在，但某一项没解析出来 —— 那一项必须是 null，不能变成 0。
|
||
|
||
和上一条的区别：那条是"根本没有快照"，这条是"有快照、其中一项平台没给"。
|
||
界面上两种都该是「—」。
|
||
"""
|
||
task_id = await _seed()
|
||
await _add_stat(
|
||
task_id, run_id=1, fans=None, total_favorited=None, works=None
|
||
)
|
||
|
||
note = (await _notes(client))[0]
|
||
|
||
assert note["creator_fans"] is None
|
||
assert note["creator_works"] is None
|
||
# 快照本身是有的（有采集时间），只是值不知道 —— 前端要能分开这两件事。
|
||
assert note["creator_stats_at"] is not None
|