feat: add Girls' Frontline 2, from IOP Wiki

The best date material this project has after wiki.gg: every row states an
exact instant on both boundaries and names the zone, so nothing is converted
and nothing is assumed. `parseIsoClockRangeUtc` requires that `(UTC)` rather
than defaulting to it — a row that ever loses it should drop out instead of
landing hours off a boundary the reader is watching in-game.

The hazard here is the Server column. CN, EN and JP rows share one table and
the Chinese schedule runs about a year ahead, which is the akwiki CN-column
problem verbatim: publish EN, skip the rest. Betas are fenced off separately,
being dated exactly like everything else and playable by nobody, and the page
is an archive back to 2023, so currency is decided against ctx.now as in
bawiki. That leaves one live event today — thin, and true.

No resetOffsets: the EN boundaries land on three different clocks, which is a
patch window rather than a reset hour.

Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>

docs/DATA-MODEL.md

@@ -10,7 +10,7 @@ import { z } from "zod";
 
 export const GameId = z.enum([
   "genshin", "hsr", "zzz", "wuwa", "arknights", "endfield", "nte", "nikki", "p5x", "r1999",
-  "ba", "fgo", "holodori",
+  "ba", "fgo", "holodori", "gfl2",
 ]);
 
 export const EventType = z.enum([
@@ -308,8 +308,16 @@ nothing to add, which is the point worth carrying to the next game: an override
 the game moves nobody, while the same override added a year later re-labels every tick already
 logged under the old clock. Get it right at introduction or accept a migration.
 
-Infinity Nikki, P5X and Blue Archive carry **no `resetOffsets` entry**, so they take the regional
-default. That is an assumption, not a verified server map — none of the three sources states one.
+Infinity Nikki, P5X and Blue Archive carry **no `resetOffsets`
+entry**, so they take the regional default. That is an assumption, not a verified server map — none
+of those sources states one. The 2026-08-19 addition below is worth separating out, because it is
+silent for a different reason than the usual one:
+
+- **Girls' Frontline 2** states an exact UTC instant on every boundary, so the data is there and
+  still settles nothing: its EN events end at 22:59 (33 of them), 08:59 (11) and 02:59 (5). Arknights
+  and Reverse: 1999 each earned an override from a single boundary the whole page agreed on; three
+  of them is a patch window, not a reset hour.
+
 Blue Archive is the case where the assumption is most likely wrong and still the right entry: the
 game does run one worldwide server, but `bluearchive.wiki` states no time of day and no timezone
 anywhere on the page, so there is no offset to read off it. Confirm these against the games before

docs/INGESTION.md

@@ -54,6 +54,7 @@ Consequences worth internalising:
 | `fandom` | Fandom wikis via the MediaWiki `action=parse` API — two page templates: `Event \| Time Period \| Version` wikitables, and FGO's picture-fenced `ONGOING EVENTS` blocks | Reverse: 1999, Fate/Grand Order |
 | `bawiki` | bluearchive.wiki's rendered `/wiki/Events` — a JP/Global tabber over `Name (EN) \| Start date \| End date \| Notes` wikitables | Blue Archive |
 | `holodoriwiki` | holodori.wiki's rendered `/wiki/Events` — `Current Events` and `Past Events` wikitables over `Event \| Type \| Start Date \| End Date` | hololive Dreams |
+| `iopwiki` | iopwiki.com's `gf-table event-period` tables — `Title \| Period (start/end) \| Server \| Type \| Comment`, one table per event and one row per server | Girls' Frontline 2 |
 
 `wikigg` is the better shape by a distance: it emits ISO timestamps with one timer per server
 region, so its events carry exact precision and real `regionEnds`. Prefer a source like that over a
@@ -140,6 +141,7 @@ All live in `src/ingest/dates.ts`, each returning null rather than inferring any
 | `parseOrdinalDateTimeRange` | `November 9th, 05:00 - December 4th, 2023, 04:59 (UTC-5)` (ordinal days, stated offset) | Reverse: 1999 |
 | `parseIsoDay` | `2026-08-04` (one boundary per column, so nothing to split) | Blue Archive |
 | `parseSlashClockZone` | `08/17/2026 8:00PM (JST)` (one boundary per column, 12-hour clock, **named** zone) | hololive Dreams |
+| `parseIsoClockRangeUtc` | `2026-08-06 13:00 - 2026-08-26 22:59 (UTC)` (whole range in one cell, **zone required**, nothing converted) | Girls' Frontline 2 |
 | `parseOpenRange` | `Jul. 24, 2026 - End of 4.6`, `July 10, 2026 - Permanent` | Star Rail, Wuthering Waves |
 
 **A day-precision result is 00:00Z, and that is a placeholder rather than a time.** Every reader
@@ -338,14 +340,6 @@ start-date proximity check is the actual guard; the title threshold is deliberat
 that "Stygian Onslaught" and "Stygian Onslaught Event" collapse into one row rather than showing
 the user a duplicate.
 
-**Agreement raises confidence (+0.10) only across different `sourceId`s.** The same row seen twice
-in one document is not corroboration.
-
-**Disagreement is surfaced, never averaged.** Two sources whose `endsAt` differ by more than 24
-hours produce a `conflicts` entry; the pipeline routes those to quarantine. Splitting the difference
-between two dates would produce a value neither source asserts — the worst possible answer for a
-product whose promise is date accuracy.
-
 ## Stage 4 — validate
 
 Zod parse against `GachaEvent`, then calendar sanity rules. Anything failing a hard rule goes to

docs/PRD.md

@@ -59,7 +59,7 @@ have a source — this table is the launch scope, not a roster to keep in sync.
 
 Adding a game must require no change to `GachaEvent` — only a `GameId` entry, its `games.ts`
 metadata, and a source registration. That is the test of whether the data model is right, and it has
-held: the thirteen games here have cost the event schema nothing. Per-game *metadata* does occasionally
+held: the fourteen games here have cost the event schema nothing. Per-game *metadata* does occasionally
 grow (Reverse: 1999 needed `resetHourLocal` for a 05:00 reset), which is a different file and moves
 no stored key. A game may have several sources; see `docs/INGESTION.md` § Three layers.
 
@@ -90,7 +90,7 @@ default `"game"` — the board every existing reader already has). Merged, the o
 `endingSoonestFirst`, the same comparator behind the list's "Ending soonest" (F2), so the two views
 cannot mean different things by the same words and an unannounced end still sorts behind every dated
 one rather than claiming a place in the queue. Merged there is also no lane heading to say whose
-event a bar is, and hue alone cannot answer that once thirteen games share a stack — so each bar
+event a bar is, and hue alone cannot answer that once fourteen games share a stack — so each bar
 carries its game's short name, with the full name in its tooltip. A bar too narrow for a tag and a
 title both keeps the title: a chopped game name reads as a broken word, and the hue and the tooltip
 still answer it.