fix(P0-02): 积分榜改为追加快照(append-only) + available_at cutoff

去掉 uq_standings_league_season_team,改为 (league, season, team, available_at) 唯一; 每次采集 INSERT 新行(available_at=now),ON CONFLICT DO NOTHING,不覆盖旧行。 standings_slice(before):DISTINCT ON (team_id) WHERE available_at<=cutoff ORDER available_at DESC;before=None → cutoff=now()。公开 list_standings 取每队最新可用快照(子查询 max available_at)。迁移 0023 + 切片/路由/docs 同步;测试 test_p0_standings_cutoff(5/5)。 284 测试全绿。
2026-09-22 02:45:50 +08:00
parent 63caa6736c
commit 49d78136a1
8 changed files with 260 additions and 32 deletions
@@ -0,0 +1,138 @@
+"""P0-02 回归测试: 积分榜改为追加快照(append-only) + available_at cutoff。
+
+运行: pytest tests/test_p0_standings_cutoff.py -v
+(模型约束用 fake DB;cutoff 过滤语义用 fake session 验证参数传递。)
+"""
+import pytest
+from sqlalchemy import select
+
+from src.db.models import League, Standing, Team
+
+
+# ── fake DB(对齐现有测试约定) ────────────────────────────────────
+class _FakeResult:
+    def __init__(self, items): self._items = list(items)
+    def scalars(self):
+        class _S:
+            def __init__(self, items): self._items = items
+            def all(self): return list(self._items)
+        return _S(self._items)
+    def scalar_one_or_none(self):
+        return self._items[0] if self._items else None
+
+
+class _FakeDB:
+    captured: list = []
+
+    def __init__(self, league=None, standing_rows=None):
+        self._league = league
+        self._rows = standing_rows or []
+        _FakeDB.captured = []
+
+    def add(self, obj):
+        _FakeDB.captured.append(obj)
+
+    async def execute(self, stmt):
+        # 记录生成的 SQL(字符串化)供断言
+        _FakeDB.captured.append(str(stmt))
+        compiled = str(stmt)
+        if "league" in compiled.lower() and "standing" not in compiled.lower():
+            return _FakeResult([self._league] if self._league else [])
+        return _FakeResult(self._rows)
+
+    async def flush(self):
+        pass
+
+
+class _FakeTeam:
+    def __init__(self, tid, name):
+        self.id = tid
+        self.name = name
+        self.name_zh = None
+
+
+class _Header:
+    def __init__(self):
+        from src.llm.slices.common import MatchHeader
+        self._h = MatchHeader(
+            match_id=1, home_name="A", away_name="B", league_name="E0",
+            season="2026", match_date="2026-01-01", match_dt=None,
+            stage=None, home_team_id=10, away_team_id=20, league_id=1,
+        )
+
+    def __getattr__(self, name):
+        return getattr(self._h, name)
+
+
+class TestStandingsModel:
+    """standings 模型必须有 available_at + 新唯一约束。"""
+
+    def test_available_at_column(self):
+        cols = {c.name for c in Standing.__table__.columns}
+        assert "available_at" in cols
+
+    def test_unique_constraint_includes_available_at(self):
+        names = {c.name for c in Standing.__table__.constraints}
+        assert any("available" in n and n.startswith("uq_") for n in names), \
+            f"缺少含 available_at 的唯一约束,现有: {names}"
+
+    def test_old_unique_constraint_removed(self):
+        names = {c.name for c in Standing.__table__.constraints}
+        assert "uq_standings_league_season_team" not in names, \
+            "旧约束 uq_standings_league_season_team 应已被替换"
+
+
+class TestStandingsSliceCutoff:
+    """standings_slice 必须尊重 before(cutoff):before=None → now()。"""
+
+    def test_before_none_uses_now(self):
+        """before=None 时应将 cutoff 视为 now()(取最新可用快照)。"""
+        from src.llm.slices import standings as st_mod
+        from datetime import datetime, timezone
+
+        calls = {}
+        real_execute = None
+
+        class _DB:
+            def __init__(self): self._league = League(id=1, code="E0", name="E0")
+            def add(self, obj): pass
+            async def execute(self, stmt):
+                # 捕获 WHERE available_at <= ? 的参数
+                sql = str(stmt)
+                if "available_at" in sql:
+                    # 提取编译后的 params
+                    try:
+                        params = stmt.compile().params
+                        calls["cutoff"] = params.get("available_at_1")
+                    except Exception:
+                        pass
+                if "league" in sql.lower() and "standing" not in sql.lower():
+                    return _FakeResult([self._league])
+                return _FakeResult([])
+            async def flush(self): pass
+
+        async def run():
+            db = _DB()
+            before = None
+            await st_mod.standings_slice(_Header(), before=before, db=db)
+
+        import asyncio
+        asyncio.run(run())
+        # before=None 时应注入 now() 作为 cutoff
+        assert "cutoff" in calls, "未对 available_at 施加 cutoff 过滤"
+        assert calls["cutoff"] is not None
+
+
+class TestStandingsAppendOnly:
+    """采集应 INSERT 新行(带 available_at),不覆盖旧行。"""
+
+    def test_values_include_available_at(self):
+        """采集构造的 Standing 必须含 available_at 字段。"""
+        from src.data import bzzoiro_standings as bzs
+        # 检查函数源码是否包含 available_at(编译期守卫)
+        import inspect
+        src = inspect.getsource(bzs.ingest_bzzoiro_standings)
+        assert "available_at" in src, "采集函数必须设置 available_at"
+        # 不应再出现按 (league, season, team) 的 upsert 查询
+        assert "scalar_one_or_none" not in src or "Standing.league_id == league.id" not in src.replace("available_at", ""), \
+            "不应再按 (league, season, team) 做 upsert 查询"
@@ -194,8 +194,8 @@ async def test_r2_standings_actually_upserts(monkeypatch):
    assert first.zone == "Champions League"     # 优先取 label


-async def test_r2_standings_upsert_updates_existing(monkeypatch):
-    """行为测试: 已存在同 (league, season, team) 时应就地更新而非新增。"""
+async def test_r2_standings_append_new_row(monkeypatch):
+    """P0-02 行为测试: 每次采集 INSERT 新行(带 available_at),不更新旧行。"""
    import src.data.bzzoiro as bz
    from src.db.models import League, Standing

@@ -217,16 +217,20 @@ async def test_r2_standings_upsert_updates_existing(monkeypatch):
    existing = Standing(league_id=42, season="2025-2026", team_id=7, position=9)
    existing.points = 1

-    # 查询顺序: League → Team 预载(命中) → Standing 查询(命中已有行)
-    db = _FakeDb(results=[_FakeResult([league]), _FakeResult([team]), _FakeResult([existing])])
+    # 查询顺序: League(命中) → Team 预载(命中) → (P0-02 不再查询 Standing)
+    db = _FakeDb(results=[_FakeResult([league]), _FakeResult([team])])

    result = await bz.ingest_bzzoiro_standings(db, leagues=["EPL"])

    assert result["total_upserted"] == 1
-    assert existing.points == 30, "已有行应被就地更新"
+    # P0-02: 追加快照——新增 Standing 行,旧行不被修改
+    new_rows = [o for o in db.added if isinstance(o, Standing)]
+    assert len(new_rows) == 1, "P0-02 应新增一条 Standing 行"
+    assert new_rows[0].points == 30, "新行应承载新采集数据"
+    assert new_rows[0].available_at is not None, "新行必须含 available_at"
+    # 旧行未被修改(仍保持原值)
+    assert existing.points == 1, "P0-02 旧行不应被覆盖"
    assert result["leagues"]["EPL"]["teams_created"] == 0
-    # 不应新增 Standing(只有 league/team 层面的 add)
-    assert not [o for o in db.added if isinstance(o, Standing)]


 def test_r2_source_contains_real_upsert_loop():
@@ -237,7 +241,9 @@ def test_r2_source_contains_real_upsert_loop():
    assert "total_upserted" in src
    assert 'result["total_upserted"] +=' in src, "total_upserted 必须真的被累加"
    assert "Standing(" in src, "必须真的构造 Standing"
-    assert "select(Standing)" in src, "必须查询已有快照以决定 insert/update"
+    # P0-02: 追加快照——每次 INSERT 新行(带 available_at),不查询旧行做 upsert
+    assert "available_at" in src, "P0-02 采集必须设置 available_at"
+    assert "scalar_one_or_none" not in src, "P0-02 不应再按 (league, season, team) 做 upsert 查询"


 # ============================================================