feat(tools): add deterministic preview database seeder

tests/test_preview_seed.py

@@ -0,0 +1,290 @@
+# -*- coding: utf-8 -*-
+"""预览库生成器测试（离线）: uv run python tests/test_preview_seed.py"""
+
+import json
+import os
+import random
+import sqlite3
+import sys
+import tempfile
+from datetime import date
+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.analytics import METRICS  # noqa: E402
+from reaction_board.bilibili.models import FavResource  # noqa: E402
+from reaction_board.config import FolderConfig  # noqa: E402
+from reaction_board.database import Database  # noqa: E402
+from reaction_board.repository import Repository  # noqa: E402
+from tools.seed_preview_db import (  # noqa: E402
+    FAKE_BLACKLIST_MID,
+    FAKE_CANDIDATE_AIDS,
+    backfill_metrics,
+    seed_preview,
+    spread_first_seen,
+)
+
+LATEST = "2026-10-31"
+START = "2026-10-24"
+
+
+def resource(aid: int) -> FavResource:
+    return FavResource(
+        aid=aid,
+        bvid=f"BV{aid}",
+        title=f"视频 {aid}",
+        up_mid=1000 + (aid % 5),
+        up_name=f"UP {aid % 5}",
+        pubdate=1_790_000_000 - aid * 3600,
+        duration=7200,
+        cover=f"//i0.hdslb.com/{aid}.jpg",
+        attr=0,
+        fav_time=None,
+    )
+
+
+def build_source(tmp: str, *, count: int = 40) -> Path:
+    path = Path(tmp) / "app.db"
+    db = Database(path)
+    db.initialize()
+    try:
+        repo = Repository(db)
+        folder = FolderConfig(year=2026, media_id=3700023351, title="2026")
+        repo.ensure_folder(folder)
+        repo.seed_blacklist([12345])
+        for aid in range(1, count + 1):
+            repo.upsert_from_fav(resource(aid), folder.media_id, was_added=True)
+            repo.upsert_snapshot(
+                LATEST,
+                aid,
+                metrics={
+                    "view": 1000 + aid * 500,
+                    "like": 50 + aid * 10,
+                    "coin": 20 + aid * 5,
+                    "favorite": 30 + aid * 4,
+                    "share": 2 + aid,
+                    "reply": 10 + aid * 2,
+                    "danmaku": 15 + aid * 3,
+                },
+                ok=True,
+                err=None,
+                captured_at=1_793_445_517,
+            )
+        repo.upsert_up_stat(
+            LATEST, 37754047, follower=2_767_270, following=95, ok=True,
+            captured_at=1_793_445_517,
+        )
+    finally:
+        db.close()
+    return path
+
+
+def snapshot_rows(path: Path) -> list[dict]:
+    conn = sqlite3.connect(path)
+    conn.row_factory = sqlite3.Row
+    try:
+        return [dict(row) for row in conn.execute(
+            "SELECT * FROM snapshot ORDER BY aid, stat_date"
+        )]
+    finally:
+        conn.close()
+
+
+def test_backfill_metrics():
+    latest = {
+        "view": 10_000, "like": 500, "coin": 200, "favorite": 300,
+        "share": 20, "reply": 100, "danmaku": 150,
+    }
+    first = backfill_metrics(latest, 5, random.Random(7), multiplier=1.0)
+    second = backfill_metrics(latest, 5, random.Random(7), multiplier=1.0)
+    assert first == second, "同一随机种子必须生成相同序列"
+    assert len(first) == 5
+    previous = first[0]
+    for row in first[1:]:
+        for metric in METRICS:
+            assert row[metric] >= previous[metric], (metric, previous, row)
+        previous = row
+    for row in first:
+        for metric in METRICS:
+            if metric != "view":
+                assert row[metric] <= row["view"]
+    assert all(value >= 0 for row in first for value in row.values())
+
+    zero = backfill_metrics({}, 3, random.Random(1))
+    assert all(value == 0 for row in zero for value in row.values())
+
+    try:
+        backfill_metrics(latest, -1, random.Random(1))
+    except ValueError:
+        pass
+    else:
+        raise AssertionError("负数天数应抛 ValueError")
+
+
+def test_spread_first_seen():
+    aids = list(range(1, 101))
+    rng = random.Random(42)
+    result = spread_first_seen(aids, date.fromisoformat(START), date.fromisoformat(LATEST), rng)
+    assert set(result) == set(aids)
+    assert len(set(result.values())) > 1
+    for ts in result.values():
+        day = date.fromtimestamp(ts).isoformat()
+        assert START <= day <= LATEST
+
+
+def test_refuses_same_path_and_existing_target():
+    with tempfile.TemporaryDirectory() as tmp:
+        source = build_source(tmp)
+        try:
+            seed_preview(source, source, days=7)
+        except ValueError as exc:
+            assert "不能与源数据库相同" in str(exc)
+        else:
+            raise AssertionError("同路径应拒绝")
+
+        target = Path(tmp) / "preview.db"
+        seed_preview(source, target, days=7, config_source=None)
+        try:
+            seed_preview(source, target, days=7, config_source=None)
+        except FileExistsError:
+            pass
+        else:
+            raise AssertionError("已存在目标且未 --force 应拒绝")
+
+
+def test_seed_preview_end_to_end():
+    with tempfile.TemporaryDirectory() as tmp:
+        source = build_source(tmp)
+        target = Path(tmp) / "preview" / "app.db"
+        config_out = Path(tmp) / "preview" / "config.json"
+        summary = seed_preview(
+            source, target, days=7, seed=20261031, config_source=None,
+            config_out=config_out,
+        )
+        assert summary["ok"] is True
+        assert summary["snapshots_failed"] == 3
+        assert summary["date_from"] == START
+        assert summary["date_to"] == LATEST
+        assert summary["videos"] == 40
+
+        # 源库快照数量不变
+        assert len(snapshot_rows(source)) == 40
+        rows = snapshot_rows(target)
+        assert len(rows) > 40
+        dates = sorted({row["stat_date"] for row in rows})
+        assert dates == [
+            START, "2026-10-25", "2026-10-26", "2026-10-27",
+            "2026-10-28", "2026-10-29", "2026-10-30", LATEST,
+        ]
+        synthetic = [row for row in rows if row["stat_date"] != LATEST]
+        assert synthetic
+        assert all(row["src"] == "preview" for row in synthetic)
+        assert sum(1 for row in synthetic if row["ok"] == 0) == 3
+        assert any(row["ok"] == 0 for row in synthetic)
+
+        conn = sqlite3.connect(target)
+        conn.row_factory = sqlite3.Row
+        try:
+            assert conn.execute(
+                "SELECT v FROM kv WHERE k='preview:seeded'"
+            ).fetchone()[0] == "1"
+            first_seen = {
+                int(row["aid"]): int(row["first_seen"])
+                for row in conn.execute("SELECT aid, first_seen FROM video")
+            }
+            for row in synthetic:
+                day = date.fromtimestamp(first_seen[int(row["aid"])]).isoformat()
+                assert row["stat_date"] >= day, (row["aid"], row["stat_date"], day)
+            assert len(conn.execute(
+                "SELECT * FROM candidate WHERE channel='preview'"
+            ).fetchall()) == len(FAKE_CANDIDATE_AIDS)
+            assert conn.execute(
+                "SELECT 1 FROM up_blacklist WHERE mid=?", (FAKE_BLACKLIST_MID,)
+            ).fetchone() is not None
+        finally:
+            conn.close()
+
+        # 每个视频的累计指标单调不降
+        by_aid: dict[int, list[dict]] = {}
+        for row in rows:
+            if row["ok"] == 1:
+                by_aid.setdefault(int(row["aid"]), []).append(row)
+        for series in by_aid.values():
+            series.sort(key=lambda item: item["stat_date"])
+            for metric in METRICS:
+                values = [item[metric] for item in series if item[metric] is not None]
+                assert values == sorted(values), (metric, values)
+
+        config = json.loads(config_out.read_text(encoding="utf-8"))
+        assert config["telegram"]["enabled"] is False
+        assert config["bilibili"]["read"]["cookie"] == ""
+        assert config["bilibili"]["write"]["csrf"] == ""
+        assert config["web"]["port"] == 8081
+        assert config["storage"]["database_path"].endswith("app.db")
+
+        # 同 seed 生成结果一致
+        target2 = Path(tmp) / "preview2" / "app.db"
+        seed_preview(source, target2, days=7, seed=20261031, config_source=None)
+        rows2 = snapshot_rows(target2)
+        assert [
+            (row["stat_date"], row["aid"], row["ok"], row["view"], row["like"])
+            for row in rows
+        ] == [
+            (row["stat_date"], row["aid"], row["ok"], row["view"], row["like"])
+            for row in rows2
+        ]
+
+
+def test_first_day_profile_and_config_no_secrets():
+    with tempfile.TemporaryDirectory() as tmp:
+        source = build_source(tmp)
+        config_source = Path(tmp) / "config.json"
+        config_source.write_text(json.dumps({
+            "bilibili": {
+                "read": {"mid": 1, "csrf": "read-csrf", "cookie": "SESSDATA=fake-read-secret"},
+                "write": {"mid": 2, "csrf": "write-csrf", "cookie": "SESSDATA=secret2"},
+            },
+            "telegram": {"enabled": True, "bot_token": "secret-token", "chat_id": 1},
+            "folders": [{"year": 2026, "media_id": 3700023351, "title": "2026"}],
+            "track_fans": [37754047],
+            "heat_weights": {"view": 1.0},
+        }), encoding="utf-8")
+        target = Path(tmp) / "preview" / "app.db"
+        config_out = Path(tmp) / "preview" / "config.json"
+        summary = seed_preview(
+            source, target, days=7, profile="first-day",
+            config_source=config_source, config_out=config_out,
+        )
+        assert summary["snapshots_inserted"] == 0
+        assert len(snapshot_rows(target)) == 40
+        config = json.loads(config_out.read_text(encoding="utf-8"))
+        assert "secret" not in json.dumps(config)
+        assert config["telegram"]["enabled"] is False
+        assert config["heat_weights"] == {"view": 1.0}
+        assert config["folders"][0]["media_id"] == 3700023351
+
+        conn = sqlite3.connect(target)
+        try:
+            assert conn.execute(
+                "SELECT COUNT(*) FROM candidate WHERE channel='preview'"
+            ).fetchone()[0] == 0
+            assert conn.execute(
+                "SELECT COUNT(*) FROM up_blacklist WHERE reason='preview'"
+            ).fetchone()[0] == 0
+        finally:
+            conn.close()
+
+
+def main_test():
+    test_backfill_metrics()
+    test_spread_first_seen()
+    test_refuses_same_path_and_existing_target()
+    test_seed_preview_end_to_end()
+    test_first_day_profile_and_config_no_secrets()
+    print("OK: 预览库生成/单调/确定性/安全边界测试全部通过")
+
+
+if __name__ == "__main__":
+    main_test()