All JstPeriod callsites have moved to IGameCalendarService — delete the
static class and its tests. Updates the doc comments on Viewer.cs
(LastLoginBonusClaimedAt) and ViewerEventCounter.cs to reference the
new UTC / config-driven boundary instead of the old "JST 02:00".
GameConfigurationJsonb config-section snapshot test picks up "GameCalendar".
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Swaps all four DateTime.UtcNow.Date same-day checks in PackController for
_calendar.ResetReady() — daily-single open, /pack/info daily-single gate,
FreePacks /pack/info visibility, FreePacks /pack/open quota. Removes the
TODO(daily-reset) comments; ToDto promoted from static to instance so it
can see _calendar.
Also fills in GameCalendarConfig.ShippedDefaults() (required by the
[ConfigSection] discovery in SVSimDbContext.EnsureSeedDataAsync — bootstrap
throws without it).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
New config-driven service that replaces JstPeriod. All timestamps are UTC;
one config value (GameCalendarConfig.DailyResetUtcHour, default 0 = midnight
UTC) drives both ResetReady checks and DayKey/WeekKey/MonthKey bucket keys.
Because both derive from the same MostRecentBoundary helper, they can never
disagree.
Prod parity is a config flip: set DailyResetUtcHour=17 in GameConfigs to get
the 02:00 JST boundary. No JST-flavored math anywhere in the code.
JstPeriod callsite migration + deletion follow in subsequent commits.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
One row per (viewer, format) tracking accumulated Point + MasterPoint.
Unique index on (ViewerId, Format) enforces one row per format per viewer
(mirrors owned-collection precedent per project_owned_collection_unique_index).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Doubles the shipped defaults across all modes without touching per-mode
overrides. Curve interaction against classexp.csv (L1=50, L2=150):
- Win (200 XP): crosses both L1 + L2 → land at L3, Exp=0
- Loss (50 XP): exactly meets L1 → land at L2, Exp=0
Updates every affected test assertion (Practice, Rank, Free, TK2,
Colosseum, Story) for the new level/exp math.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
- StoryConfig.ClassXpPerClear retired; equivalent knob moves to
BattleXpConfig.StoryXpPerClear (falls back to XpPerWin when null).
- StoryService.FinishAsync now grants class XP via IBattleXpService with
BattleXpMode.Story on the firstClear && isPlayShape branch. Persists
post-grant Level/Exp to Viewer.Classes.
- Wire fields (get_class_experience, class_level) stringified as before.
- StoryConfig kept as an empty [ConfigSection] for future story-specific
knobs. StoryServiceTests updated to reflect the new default path
(XpPerWin=100 crosses classexp.csv L1=50 → L2 with 50 carry).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
- ArenaTwoPickService now delegates class-XP grants to IBattleXpService
with BattleXpMode.ArenaTwoPick. Inline GrantClassXp/ResolveClassLevel
helpers deleted.
- ClassXpPerBattle field removed from ArenaTwoPickConfig; TK2 XP now
configurable via ArenaTwoPickXpPerWin/XpPerLoss overrides or the global
BattleXpConfig defaults (100/25).
- Updated 5 test files (ArenaTwoPickService callers) for the new
constructor arg. Added loss-XP assertion to Finish tests.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Adds BattleXpConfig ([ConfigSection("BattleXp")]) with global
XpPerWin/XpPerLoss plus per-mode nullable overrides. Adds BattleXpMode
enum and IViewerRepository.LoadForBattleXpGrantAsync — focused tracked
load with viewer.Classes + Class nav ref included, for the service
about to be introduced.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Squashes 146 commits from battle-engine-extraction. Net: 2,045 files changed,
+11,896 / -158,687 lines. Ships engine passes 4-7 (dead-code cull, view-layer
stub, receive-path shrink) plus the Phase-5 AsyncLocal ambient deletion that
turns concurrent battles into a type-system property rather than a scope
contract.
## What landed
**Passes 4-7 (chunks 1-34):** Extended the Phase-4 const-false collapse into a
cascading cull across the skill graph, view layer, and receive-path periphery.
Six mode flags (IsWatchBattle/IsReplayBattle/IsAdmin/IsAdminWatch/IsPuzzleQuest/
IsAINetwork) became `const false`, every guarded block deleted. Field*.cs
subclass ctors + BackGroundBase + ObjectChecker culled to no-ops. Mulligan
family reworked to take a mgr param through IMulliganMgr.InitMulligan.
Emotion/Recovery/Resource clusters null-stubbed. Prediction/OperationSimulator/
skill filters converted from static ambient reads to per-mgr reads via
SkillPrm.ownerCard.SelfBattlePlayer.BattleMgr / ins.BattleMgr / this.BattleMgr.
**Phase-5 ambient rip (chunks 35-47):** Deleted BattleAmbient / BattleAmbient-
Context / TestBattleScope in full. Every per-battle mutable slot now lives on
the mgr instance itself:
mgr.InstanceIsForecast / InstanceIsRandomDraw / InstanceRecoveryInfo /
InstanceViewerId / InstanceNetworkAgent / GameMgr
BattleManagerBase.GetIns() returns null unconditionally; the residual static
flags + 3 façades (Certification.ViewerId, Data.BattleRecoveryInfo,
ToolboxGame.RealTimeNetworkAgent) are null-tolerant defaults kept for the
handful of engine-internal readers that still reference their types. Zero
BattleAmbient references anywhere in engine + node + tests.
Added pre-seeded GameMgr ctor overload threaded through the mgr chain
(BattleManagerBase → SingleBattleMgr / NetworkBattleManagerBase → NetworkStandard-
BattleMgr → HeadlessBattleMgr / HeadlessNetworkBattleMgr). Fixtures build a
GameMgr, seed it via HeadlessEngineEnv.SeedCharaIds/SeedNetUser, and pass it
to the mgr's ctor — no ambient reach.
Node side (SVSim.BattleNode/SessionBattleEngine): _ctx replaced with a plain
GameMgr field; 34 `using var _ambient = BattleAmbient.Enter(_ctx)` scope wraps
ripped from every accessor and mutator; EngineGlobalInit.WirePerSessionGameMgr
takes GameMgr as a param and runs from SessionBattleEngine.SetupInternal
BEFORE mgr construction.
Test side: TestBattleScope deleted; 18 fixture [SetUp]s migrated to
`HeadlessEngineEnv.EnsureProcessGlobals()`; MultiInstanceEngineTests rewritten
around per-mgr construction (GetIns() → null is the pinned invariant).
## Regression fixes
- **chunk-48** (MulliganCtrl): chunk-35's `= null` stubs on card lookups broke
the live receive-driven Deal path (BattlePlayerBase.DrawCard NRE'd downstream
of NetworkPlayerMulliganCtrl.StartMulliganVfx). Restored the three lookups
via `_battlePlayer.BattleMgr.GetBattleCardIdx`. Engine tests were satisfied
by the WireMulliganPhase seam; unit tests exposed the live-path gap.
## Ship state
- SVSim.BattleEngine.Tests: 56/56 pass, 2 skip
- SVSim.UnitTests: 1554/1554 pass (was 1523/31-fail before chunk 48)
- Solution build: 0 source warnings (40 pre-existing NU1902 MessagePack CVEs
in SVSim.EmulatedEntrypoint, unrelated)
- Sequential PVP smoke: verified live (two back-to-back battles, no regression
on cleanup/spinup)
- Concurrent PVP smoke: verified live
Adds tools/engine-port/ClosureAnalyzer/ — the Roslyn transitive-type-closure
analyzer needed to make future cascade cleanup safe (per feedback memory
"Engine cleanup needs closure tool" from the 2026-06-28 pass-3 failure).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Implements Task 9 from docs/superpowers/plans/2026-06-13-decomp-only-followups.md
— the pre-purchase commit endpoint the client fires from StarterClassSelectDialog
before opening a RotationStarterCardPack. Without this the per-class draw path
from yesterday's work is unreachable: the dialog blocks on the AJAX call and
never advances to /pack/open.
Wire surface:
- POST /pack/set_rotation_starter_class { pack_id, class_id } → EmptyData.
Validates 1..8, rejects unknown packs (404), non-RS packs (400), and second
commits (400 — one-shot per pack per spec).
- PackConfigDto carries selected_class_id (nullable, omitted via global
WhenWritingNull policy when unset). Placement verified against decomp at
PackInfoTask.cs:86 — it's on the parent PackConfig, NOT the child gacha as
the original plan text had it.
- /pack/open cross-checks request.class_id against the persisted choice for
RS packs; rejects 400 on missing-commit or class-mismatch so a tampered
request can't swap pools after the choice is locked.
Storage:
- ViewerPackStarterClass owned entity; (ViewerId, PackId) unique index via the
pattern from project_owned_collection_unique_index memory.
- Migration PackStarterClass — composite PK + FK cascade verified against
fresh Postgres bootstrap.
Tests (9 new in PackControllerSetRotationStarterClassTests):
- Round-trip set → /pack/info shows selected_class_id.
- Field absent before any commit.
- Invalid class (0/9/-1/100) → 400.
- Already-chosen rejection.
- Non-RS pack / unknown pack rejection.
- /pack/open rejects no-prior-commit AND class-mismatch; succeeds when class
matches the persisted choice.
Full suite: 1552 passed (1543 + 9).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Add a nullable ClassId axis to PackDrawCardWeightEntry. Null means the
row applies to any draw (the existing behavior for non-RS packs); 1..8
restricts the row to a specific class draw on a RotationStarterCardPack.
Slot rates are class-invariant (verified against per-klan captures), so
PackDrawSlotRateEntry is unchanged.
PackOpenService.Draw now takes an optional classId parameter. With a
class supplied, the card pool is filtered to (ClassId == classId ||
ClassId == null); without one, class-tagged rows are excluded so a
mistaken non-RS call against a mixed table can't leak class-specific
cards.
PackController.Open lifts the 501 guard on RotationStarterCardPack and
plumbs class_id through. Class_id is now required (and validated 1..8)
for RS packs and rejected as 400 BadRequest on non-RS packs. The skin
overload (target_card_id) remains 501 — out of scope for this work.
Seed JSON regenerated end-to-end: 32 RS packs now carry 30,160 class-
tagged weight rows; 247 non-RS packs keep their existing ~87k null-class
rows. Slot rate fix from the upstream extractor change replaces the
~99.99% conditional-rate sums on 97xxx/93xxx packs with the page's
authoritative 1.5/6/25/67.5 distribution.
Migration AddPackDrawCardWeightClassId. End-to-end bootstrap verified
against fresh Postgres; pack 93025 reads back exactly 122/119/120/120/
120/121/118/119 weights for classes 1-8.
Tests: 3 new PackOpenServiceTests (class filter happy path, disjoint
pools, defensive null-class isolation); existing class_id-on-non-RS
controller test updated 501 → 400 BadRequest; one stub-importer test
corrected to match extractor-authored gacha_detail.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
The captured shape ({"1":[], "3":[], "4":[]}) is the empty-period
wire echo; the client parser (LoadDetail.cs:553) only reads normal/total/
campaign. Table, entity, seed, and importer were never wired to a code path.
Replaced by config-section catalog + service in the follow-up commits.
Three fixups the Phase 1-3 work missed (would bite the first time an admin
flipped on a real event):
* IsDisplayTips + TipsId added to ColosseumSeasonConfig. Drives
Wizard/ColosseumEntryInfoTask.cs:129 (NeedsFirstTips = is_display_tips == 1).
The wire DTO already had the properties; the season config didn't carry
them, so admins had no way to gate the first-time tutorial tip popup.
* colosseum_id is now emitted from SeasonId. Wire field was always-absent
before (DTO had the property but builder never set it); client fell back
to 0 and couldn't disambiguate cups.
* ColosseumLobbyInfoBuilder.Build now takes ColosseumRoundsConfig + nowUtc
and derives the wire is_round_period / now_round / next_round / start_time
/ end_time from the rounds schedule against the current clock. Previous
hardcode (is_round_period: true, now_round: 1, event-window dates) would
have shown "Round 1 ends 2026-07-31" instead of the actual round's end.
Per the client parser at ColosseumEntryInfoTask.cs:111-126:
- inside a round → emit that round's window + is_round_period=true
- between rounds → emit next round's start as start_time + next_round
- pre-event → next_round=1, start_time=event start
- post-event → anchor to last round so the client doesn't NRE
- no rounds configured → fall back to event-level window
Why these were missed: I implemented ColosseumSeasonConfig from the plan's
explicit field list rather than cross-checking against the captured
colosseum.json wire shape; I added ColosseumId to the wire DTO without
wiring the builder side; and no test exercised a mid-event scenario with
realistic dates — Phase 2's E2E hand-edits run.RoundId via
PromoteRunToRoundAsync instead of letting the controller derive it.
* 10 new builder unit tests cover off-season / mid-round / between-rounds /
pre-event / post-event / empty-rounds-fallback + the new field surfacing.
Full suite: 1355/1355 (was 1345 + 10).
/mypage/index data.colosseum_info and /arena_colosseum/{top,get_fee_info}
now both project through ColosseumLobbyInfoBuilder.Build(ColosseumSeasonConfig).
One source of truth for the home-tab gating and the lobby reads — admins
flip the season on by writing to GameConfigs (per the runbook), no second
table to keep in sync.
* New ColosseumLobbyInfoBuilder — static projection from ColosseumSeasonConfig
to the wire ColosseumLobbyInfo. Replaces ArenaColosseumController's private
BuildColosseumInfo and MyPageController's BuildColosseumInfo(ColosseumConfig).
* MyPageController reads ColosseumSeasonConfig via IGameConfigService; the
IGlobalsRepository.GetCurrentColosseum() call goes away.
* MyPageIndexResponse.ColosseumInfo retyped to the new ColosseumLobbyInfo.
* ColosseumSalesPeriodInfo relocated into the ArenaColosseum namespace
alongside the rest of the family DTOs.
* Drops: Colosseums DB table (migration DropColosseumsTable), ColosseumConfig
entity, the captured-shape Models.Dtos.ColosseumInfo DTO, ColosseumSeed,
MyPageGlobalsImporter.ImportColosseumAsync + seed file colosseum.json,
IGlobalsRepository.GetCurrentColosseum + impl, the two related tests.
Net: home-screen Grand Prix tab now hidden by default because the season
ships with IsColosseumPeriod=false. Full suite: 1345/1345 (was 1347 — the
2 dropped tests covered the now-dead surface).
Closes the family. arena-colosseum 10/16 → 15/16 zero stubs (the 16th —
finish_load — is dead per project_dead_battle_endpoints).
* 2-Pick draft lift onto ArenaColosseumController:
- get_candidate_classes samples from ArenaTwoPickConfig.AllowedClassIds
and persists the slate onto the run.
- class_choose accepts class_id XOR chaos_id; both populate run.ClassId,
chaos branch stores ChaosId for replay.
- get_candidate_cards is the idempotent draft-resume snapshot.
- card_choose appends both cards from the picked pair, advances turn 1..15.
- Pool override: ArenaTwoPickCardPoolService gets a non-breaking
GeneratePickSetsForTurn(..., poolCardSetIds) overload; Colosseum routes
pass ColosseumSeasonConfig.PoolCardSetIds (falls back to challenge →
rotation when empty).
* Curated-deck schema: ColosseumHofDeck / ColosseumWindFallDeck /
ColosseumAvatarDeck — three identical tables (separate per per-pool
operational lifecycle), unique on DeckNo. Migration AddColosseumCuratedDecks.
* IColosseumCuratedDeck interface + a generic ColosseumCuratedDeckImporterBase<T>
with three concrete subclasses (HOF / WindFall / Avatar), registered in
Bootstrap. Seed files ship empty.
* 6 curated endpoints on ArenaColosseumController: get_{hof|windfall|avatar}_deck_list
return BARE-ARRAY data per spec; register_{hof|windfall|avatar}_deck share
one generic RegisterCuratedAsync<T> dispatcher. Cross-pool register
rejected via per-pool lookup. Curated register has no is_published flag
(constructed-only) — clears the run's flag for state consistency.
* Tests: 5 draft HTTP tests + 11 curated-deck HTTP tests (4 parameterized
3 ways across HOF/WindFall/Avatar + a cross-pool isolation test + an
is_published clear test). Existing TwoPick service tests updated for the
new pool overload. Full suite: 1347/1347.
Phase 3 ship gate met. Branch ready for merge.
Closes 4 more arena-colosseum spec shapes (6/16 → 10/16): the dual-prefix
battle URLs (colosseum_battle/* and colosseum_rank_battle/*) for both
do_matching and per-match finish, plus the bracket-end /finish + /retire
pair on /arena_colosseum.
* ColosseumProgressionService: pure-logic decisions for win-threshold
advancement, 3008 promotion trigger, and per-round reward bundles.
Extends ColosseumRoundsConfig with FinishRewards/RetireRewards per
round + a top-level ChampionRewards list.
* MatchContextBuilder.BuildForColosseumAsync: lifts the registered deck
via IDeckRepository (NOT viewer-graph) per project_ef_nav_include_pitfall.
* ArenaColosseumBattleController: dual route prefixes off one controller
via explicit absolute [HttpPost("/...")] routes — does NOT extend
SVSimController's [Route("[controller]")]. Same pattern as FreeBattle.
URL-vs-IsRankMatching mismatch returns 400 with a phase-mismatch error;
no-deck triggers the standard 3001 matchmaking-illegal state. 3008
flips run.IsRankMatching exactly once via the progression service.
* /arena_colosseum/finish + /retire share FinishResponse (rewards +
reward_list + colosseum_status). Champion finish appends ChampionRewards
and surfaces colosseum_status.is_champion = true. Both delete the run
row after granting via IInventoryTransaction.GrantAsync.
* DI: ModePolicy entries for "colosseum_battle" and "colosseum_rank_battle".
* Tests: 8 progression-service unit tests (pure logic), 5 battle-controller
HTTP tests, 5 bracket-terminate HTTP tests, 2 full-bracket E2E walks
(champion path + retire mid-round). Full suite: 1331/1331.
is_skip_gacha_effect, use_challenge_two_pick_premium_card, and
challenge_two_pick_sleeve_id all persist on ViewerInfo and round-trip
through /load/index.
The two challenge fields move from the global ChallengeConfig section to
ViewerInfo — they're viewer preferences, not server-wide policy. Premium
card defaults to 0; sleeve falls back to DefaultLoadoutConfig.SleeveId
(3000011) when unset. MatchContextBuilder's TK2 sleeve read switches to
the same viewer-scoped path.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
New table backs /replay/info; composite PK (ViewerId, BattleId), index on
(ViewerId, CreateTime) for the newest-first list query. 50-row per-viewer
retention enforced by BattleHistoryWriter (next commit).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Lays the persistence foundation for the /friend/* API surface. Three new
model classes with composite PKs / unique constraints / FK cascades registered
on SVSimDbContext; 4/4 persistence tests pass on SQLite in-memory.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Three new EF entities for /campaign/regist_serial_code: SerialCodeEntry (code, message,
window, enabled flag), SerialCodeRewardEntry (FK child, per-slot reward), and
ViewerSerialCodeRedemption (composite-PK redemption record). Registered in SVSimDbContext
with unique index on Code and cascade FK constraints. 3/3 persistence tests pass.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds the bool column (defaults false) to the [Owned] ViewerClassData
entity and a two-test persistence suite that verifies round-trip and
default-value behaviour via SQLite/EnsureCreated.
Adds BGType persistence (0=Deck/1=CustomBG/2=RandomBG) to Viewer via two scalar
columns and an owned collection keyed (ViewerId, Slot). Two persistence tests
confirm round-trip and zero-defaults on fresh viewers.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds the ViewerAcquireHistoryEntry model (8 fields: Id, ViewerId,
RewardType, RewardDetailId, RewardCount, AcquireType, Message,
AcquireTime), registers DbSet<ViewerAcquireHistoryEntry> on
SVSimDbContext, configures model (PK, FK cascade to Viewer, MaxLength
64 on Message, composite index on ViewerId/AcquireTime/Id), and adds a
DbSet round-trip integration test.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
The five tutorial gifts every fresh viewer is given at signup used to live as a
static C# array in SVSim.Database/SeedData/TutorialPresents.cs — outside the
seed-JSON pipeline used by all 40+ other globals tables. Editing a gift required
a code change + rebuild instead of a JSON edit + bootstrap re-run.
Now authored in SVSim.Bootstrap/Data/seeds/tutorial-presents.json and loaded into
a new TutorialPresentEntries table via TutorialPresentsImporter (clear-and-rewrite,
mirroring HomeDialogs). ViewerRepository.RegisterAnonymousViewer reads the table
at signup and projects each row into a ViewerPresent with Source="tutorial".
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
[Timestamp] byte[] doesn't work under SQLite (the test backend) — EF
expects the DB to populate it on insert, but SQLite has no equivalent
of Postgres's xmin. The WHERE Status = Unclaimed filter plus
IInventoryService's viewer-level concurrency is the practical defense;
RowVersion was only a backstop. Regenerated the migration without the
RowVersion column.
Wire reward_type on the gift endpoint uses a gift-specific scheme that
diverges from UserGoodsType for currencies: wire 1 = Crystal (enum=2),
wire 9 = Rupy (enum=9), wire 4 = Item (enum=4). A naked cast resolves
wire 1 to UserGoodsType.RedEther and silently grants the wrong wallet
— restored the explicit WireRewardTypeToUserGoodsType map from the old
tutorial controller.
Retrofits existing GiftControllerTests to call SeedTutorialPresentsAsync
on the new helper (RegisterViewer doesn't auto-seed; only the prod
signup path does). All 7 existing tests pass.
DDL-only per migrations-are-ddl-only convention. Seeded by
SVSim.Bootstrap MyPageGlobalsImporter (T5) — no HasData.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Replace bare `int RewardType` on 12 catalog/reward entities and GrantedReward
with the existing UserGoodsType enum. Verified against the decompiled client:
every wire reward_type decodes through the single Wizard.UserGoods.Type enum, so
one enum is correct across all endpoint families (item_type is a separate
Item.Type axis, left untouched). EF stores the enum as the same int column, so
there is no migration.
- Importers cast seed int -> UserGoodsType at the ingest boundary.
- New GrantedReward.ToRewardList() extension replaces 8 copy-pasted
GrantedReward -> RewardListEntry projections.
- Fix 3 .ToString() sites that would otherwise emit enum names ("Crystal")
instead of the int wire value ("2").
- Wire DTOs keep int; the enum is widened to int at the wire boundary only.
Build green; 962/962 tests pass.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Phase 3 shipped the AI rank battle bot pool as a hardcoded 8-entry list
inlined in SVSim.EmulatedEntrypoint/Matching/BotRoster.cs — editing meant
recompiling. Per PLAN.md 2026-06-02 item (d), move it to a Bootstrap
importer so the roster lives in seeds/bot-roster.json and the DB.
Shape mirrors PracticeOpponent end-to-end:
- BotRosterEntry (SVSim.Database/Models) — PK = AiId via the Id passthrough
pattern. DbSet<BotRosterEntry> BotRoster on SVSimDbContext.
- AddBotRoster migration (DDL only, per migrations-are-DDL-only rule).
- seeds/bot-roster.json — 8 rows preserving the current prod-verified
cosmetic ids (sleeve 704141010 / emblem 400001100 / degree 120027 /
field 5) and series-1 ai_ids from rm_ai_setting.csv (1111..1181).
- BotRosterSeed POCO + BotRosterImporter (idempotent upsert keyed by AiId,
leaves seed-missing rows intact). Wired into SVSim.Bootstrap/Program.cs
next to PracticeOpponentImporter.
- IGlobalsRepository.GetBotRoster() + impl.
IBotRoster.Pick → PickAsync because BotRoster now depends on the transient
IGlobalsRepository. RankBattleController awaits the new signature. The
deterministic hash-on-ctx invariant (same ctx → same bot, so /ai_<fmt>/start
retries pick the same opponent) is preserved.
DI: AddSingleton<IBotRoster> → AddTransient (matches IGlobalsRepository's
lifetime). Test fixture's SeedGlobalsAsync also runs the importer so
RankBattleControllerTests + the rewritten BotRosterTests both see seeded
rows.
Tests: 931 → 936 passing. Existing 3 BotRosterTests reshaped for the DB
backing + 1 new "throws on empty roster" guard; 4 new
BotRosterImporterTests mirror PracticeOpponentImporterTests
(round-trip / idempotent / seed-missing-row-intact / ai_id=0 skip).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Adds a [ConfigSection("Matching")] POCO carrying the
RankBattleAiFallbackThresholdSeconds tunable (default 15). Auto-picked
up by EnsureSeedDataAsync's reflection-based ConfigSection seeder.
Consumed by InProcessPairUp in a later task.
GameConfigurationJsonbTests.EnsureSeedData_writes_one_row_per_ConfigSection_with_ShippedDefaults_payload
updated to include the new Matching section in its expected keys.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Replaces the all-rows-granted reward model with per-group weighted
pick. Each ArenaTwoPickReward row now belongs to a RewardGroup with a
Weight; finish/retire groups the WinCount's rows by RewardGroup and
picks exactly one row per group, weighted by Weight (excluding
Weight==0). A RewardNum==0 outcome skips both the grant and the
rewards[] emission. Empty WinCount catalogs emit empty arrays.
Existing seed entries preserve deterministic behavior by living in
single-option groups (each with weight 1). Future seasons can expand
groups to multi-option for true randomized rewards (e.g. 200-280
rupies).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
User clarified: the 0..7 win reward tiers came from Shadowverse Worlds
Beyond (sequel), not the original game we're emulating. Original SV's
Take Two caps at 5 total battles played and has no loss limit (verified
on prod: queueing continues with 2+ losses).
- arena-two-pick-rewards.json: drop 6w + 7w tiers (12 rows remain).
- ArenaTwoPickConfig: remove MaxLosses property.
- ArenaTwoPickService: termination is now battlesPlayed >= maxBattles
(5 from MAX(reward.WinCount)). RecordBattleResult no longer flips
IsSelectCompleted on the 2nd loss.
- ResolveMaxBattleCountAsync empty-catalog default 7 → 5.
- Tests updated for the new counts (16 → 12 rows, max 7 → 5).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
