brackt/scripts/__tests__/backfill-canonical-layer.test.ts
Chris Parsons 8186dbb525
scripts: add backfill orchestrator for canonical layer
Populates canonical tournaments, participants, tournament_results, and
participant_surface_elos from per-window data for qualifying-points
sports. Skips already-linked rows, is idempotent, and supports dry-run
mode.

Critical invariants enforced by the implementation:
- qualifying_points_awarded is never copied to tournament_results
- season_participant_qualifying_totals is never touched
- conflicting surface-Elo values between windows raise a loud error
  (recorded in report.errors) rather than overwriting
2026-05-01 20:57:38 +00:00

450 lines
15 KiB
TypeScript

import { describe, it, expect, vi, beforeEach } from "vitest";
vi.mock("~/database/context", () => ({
database: vi.fn(),
}));
import { runBackfill } from "../backfill-canonical-layer";
import { database } from "~/database/context";
import * as schema from "~/database/schema";
/**
* Fixture rows used across multiple tests.
*/
const SPORT_ID = "sport-golf";
const SEASON_ID = "season-golf-2026";
const GOLF_SEASON = {
id: SEASON_ID,
sportId: SPORT_ID,
scoringPattern: "qualifying_points",
} as const;
function makeEvent(overrides: Partial<Record<string, unknown>> = {}) {
return {
id: "event-1",
sportsSeasonId: SEASON_ID,
tournamentId: null,
name: "Masters Tournament",
eventDate: "2026-04-09",
eventType: "tournament",
...overrides,
};
}
function makeSeasonParticipant(
overrides: Partial<Record<string, unknown>> = {},
) {
return {
id: "sp-1",
sportsSeasonId: SEASON_ID,
participantId: null,
name: "Scottie Scheffler",
...overrides,
};
}
/**
* Builds a fake drizzle db that routes select/insert/update calls based
* on the target table. The caller supplies per-table select-result
* arrays (one per call to that table's select()). Inserts return the
* value they were given, extended with a stub id. Updates are recorded
* but otherwise no-op.
*/
interface TableState {
/** Queue of results to return for successive select() calls on this table. */
selects?: unknown[][];
/** Rows that insert().values().returning() should yield. */
insertReturns?: unknown[];
/** Incremented on each update() call. */
updates?: { count: number };
}
interface FakeDbOptions {
tables: Map<unknown, TableState>;
/** Records every insert call: tableRef → rows seen. */
inserts?: Map<unknown, unknown[]>;
/** Records every update call: tableRef → count. */
updateCounts?: Map<unknown, number>;
}
function makeFakeDb(opts: FakeDbOptions) {
const { tables, inserts, updateCounts } = opts;
// Each select() starts a new chain that ultimately resolves to the
// next queued selects[] entry for the passed table. The chain is a
// thenable via .limit/.where resolution.
const db = {
select: vi.fn().mockImplementation(() => {
let boundTable: unknown;
const chain: Record<string, unknown> = {
from: vi.fn().mockImplementation((table: unknown) => {
boundTable = table;
return chain;
}),
where: vi.fn().mockImplementation(() => chain),
limit: vi.fn().mockImplementation(() => chain),
// Terminal: make chain thenable so `await chain` yields the queued result.
then: (resolve: (v: unknown) => unknown) => {
const state = tables.get(boundTable);
const queue = state?.selects ?? [];
const next = queue.shift() ?? [];
return Promise.resolve(next).then(resolve);
},
};
return chain;
}),
insert: vi.fn().mockImplementation((table: unknown) => {
return {
values: vi.fn().mockImplementation((row: unknown) => {
if (inserts) {
const seen = inserts.get(table) ?? [];
seen.push(row);
inserts.set(table, seen);
}
const state = tables.get(table);
const returnRows = state?.insertReturns ?? [
{ ...(row as object), id: `generated-${Math.random()}` },
];
const afterValues = {
returning: vi.fn().mockResolvedValue(returnRows),
// If the caller doesn't chain .returning(), make it awaitable anyway.
then: (resolve: (v: unknown) => unknown) =>
Promise.resolve(returnRows).then(resolve),
};
return afterValues;
}),
};
}),
update: vi.fn().mockImplementation((table: unknown) => {
if (updateCounts) {
updateCounts.set(table, (updateCounts.get(table) ?? 0) + 1);
}
return {
set: vi.fn().mockReturnValue({
where: vi.fn().mockResolvedValue(undefined),
}),
};
}),
};
return db;
}
beforeEach(() => {
vi.clearAllMocks();
});
// ─── 1. empty DB no-op ─────────────────────────────────────────────────────
describe("runBackfill - empty DB", () => {
it("returns zero counts when no qualifying-points seasons exist", async () => {
const tables = new Map<unknown, TableState>([
[schema.sportsSeasons, { selects: [[]] }],
]);
vi.mocked(database).mockReturnValue(makeFakeDb({ tables }) as never);
const report = await runBackfill({ dryRun: false });
expect(report).toMatchObject({
tournamentsCreated: 0,
tournamentsLinked: 0,
participantsCreated: 0,
participantsLinked: 0,
tournamentResultsCreated: 0,
surfaceElosCreated: 0,
errors: [],
});
});
});
// ─── 2. golf window with 4 events creates 4 tournaments ───────────────────
describe("runBackfill - golf tournaments", () => {
it("creates 4 canonical tournaments for a golf season with 4 events", async () => {
const events = [
makeEvent({ id: "ev-1", name: "Masters Tournament", eventDate: "2026-04-09" }),
makeEvent({ id: "ev-2", name: "PGA Championship", eventDate: "2026-05-14" }),
makeEvent({ id: "ev-3", name: "U.S. Open", eventDate: "2026-06-18" }),
makeEvent({ id: "ev-4", name: "The Open Championship", eventDate: "2026-07-16" }),
];
const inserts = new Map<unknown, unknown[]>();
// Build per-event insertReturns so each new tournament gets its id.
const tables = new Map<unknown, TableState>([
[schema.sportsSeasons, { selects: [[GOLF_SEASON]] }],
[
schema.scoringEvents,
{
// 1st select: unlinked events; 2nd: refetch for results loop.
selects: [events, events.map((e, i) => ({ ...e, tournamentId: `t-${i + 1}` }))],
},
],
[
schema.tournaments,
{
// One select-miss per event, all return [] (no existing row).
selects: [[], [], [], []],
// One insert per event; return sequential ids.
insertReturns: [{ id: "t-1" }],
},
],
[schema.seasonParticipants, { selects: [[]] }],
[schema.eventResults, { selects: [[], [], [], []] }],
[schema.seasonParticipantSurfaceElos, { selects: [[]] }],
]);
// Because our insertReturns is consumed once per test, re-queue per event
// by overriding insert behaviour via the fake db options.
const tournamentIds = ["t-1", "t-2", "t-3", "t-4"];
let insertIdx = 0;
const db: ReturnType<typeof makeFakeDb> = makeFakeDb({ tables, inserts });
// Override insert for tournaments specifically to return sequential ids.
db.insert.mockImplementation((table: unknown) => ({
values: vi.fn().mockImplementation((row: unknown) => {
const seen = inserts.get(table) ?? [];
seen.push(row);
inserts.set(table, seen);
let returnRows: unknown[];
if (table === schema.tournaments) {
returnRows = [{ ...(row as object), id: tournamentIds[insertIdx++] }];
} else {
returnRows = [{ ...(row as object), id: "x" }];
}
return {
returning: vi.fn().mockResolvedValue(returnRows),
then: (resolve: (v: unknown) => unknown) =>
Promise.resolve(returnRows).then(resolve),
};
}),
}));
vi.mocked(database).mockReturnValue(db as never);
const report = await runBackfill({ dryRun: false });
expect(report.tournamentsCreated).toBe(4);
expect(report.tournamentsLinked).toBe(4);
expect(inserts.get(schema.tournaments)).toHaveLength(4);
expect(report.errors).toEqual([]);
});
});
// ─── 3. re-running doesn't duplicate ──────────────────────────────────────
describe("runBackfill - idempotence", () => {
it("does not create new tournaments when they already exist", async () => {
const events = [
makeEvent({ id: "ev-1", name: "Masters Tournament", eventDate: "2026-04-09" }),
];
const existingTournament = {
id: "t-existing",
sportId: SPORT_ID,
name: "Masters Tournament",
year: 2026,
};
const inserts = new Map<unknown, unknown[]>();
const updateCounts = new Map<unknown, number>();
const tables = new Map<unknown, TableState>([
[schema.sportsSeasons, { selects: [[GOLF_SEASON]] }],
[
schema.scoringEvents,
{
selects: [events, [{ ...events[0], tournamentId: "t-existing" }]],
},
],
[schema.tournaments, { selects: [[existingTournament]] }],
[schema.seasonParticipants, { selects: [[]] }],
[schema.eventResults, { selects: [[]] }],
[schema.seasonParticipantSurfaceElos, { selects: [[]] }],
]);
vi.mocked(database).mockReturnValue(
makeFakeDb({ tables, inserts, updateCounts }) as never,
);
const report = await runBackfill({ dryRun: false });
expect(report.tournamentsCreated).toBe(0);
expect(report.tournamentsLinked).toBe(1);
expect(inserts.get(schema.tournaments)).toBeUndefined();
});
});
// ─── 4. participant backfill ───────────────────────────────────────────────
describe("runBackfill - participants", () => {
it("creates canonical participants and links season_participants", async () => {
const sps = [
makeSeasonParticipant({ id: "sp-1", name: "Scottie Scheffler" }),
makeSeasonParticipant({ id: "sp-2", name: "Rory McIlroy" }),
];
const inserts = new Map<unknown, unknown[]>();
const updateCounts = new Map<unknown, number>();
const pIds = ["p-1", "p-2"];
let pIdx = 0;
const tables = new Map<unknown, TableState>([
[schema.sportsSeasons, { selects: [[GOLF_SEASON]] }],
[schema.scoringEvents, { selects: [[], []] }],
[schema.seasonParticipants, { selects: [sps] }],
[schema.participants, { selects: [[], []] }],
[schema.eventResults, { selects: [] }],
[schema.seasonParticipantSurfaceElos, { selects: [[]] }],
]);
const db = makeFakeDb({ tables, inserts, updateCounts });
db.insert.mockImplementation((table: unknown) => ({
values: vi.fn().mockImplementation((row: unknown) => {
const seen = inserts.get(table) ?? [];
seen.push(row);
inserts.set(table, seen);
let returnRows: unknown[];
if (table === schema.participants) {
returnRows = [{ ...(row as object), id: pIds[pIdx++] }];
} else {
returnRows = [{ ...(row as object), id: "x" }];
}
return {
returning: vi.fn().mockResolvedValue(returnRows),
then: (resolve: (v: unknown) => unknown) =>
Promise.resolve(returnRows).then(resolve),
};
}),
}));
vi.mocked(database).mockReturnValue(db as never);
const report = await runBackfill({ dryRun: false });
expect(report.participantsCreated).toBe(2);
expect(report.participantsLinked).toBe(2);
expect(inserts.get(schema.participants)).toHaveLength(2);
expect(updateCounts.get(schema.seasonParticipants)).toBe(2);
});
});
// ─── 5. Masters 2026 round-trip ───────────────────────────────────────────
describe("runBackfill - tournament_results from event_results", () => {
it("copies placement and rawScore but NOT qualifying_points_awarded", async () => {
const events = [
makeEvent({
id: "ev-masters",
tournamentId: "t-masters",
name: "Masters Tournament",
eventDate: "2026-04-09",
}),
];
const seasonParticipant = {
id: "sp-1",
sportsSeasonId: SEASON_ID,
participantId: "p-scottie",
name: "Scottie Scheffler",
};
const eventResult = {
id: "er-1",
scoringEventId: "ev-masters",
seasonParticipantId: "sp-1",
placement: 1,
rawScore: "-12.00",
qualifyingPointsAwarded: "100.00",
};
const inserts = new Map<unknown, unknown[]>();
const tables = new Map<unknown, TableState>([
[schema.sportsSeasons, { selects: [[GOLF_SEASON]] }],
[
schema.scoringEvents,
{
// 1st: no unlinked events (tournamentId already set)
// 2nd: refetch — events linked to tournament
selects: [[], events],
},
],
[schema.tournaments, { selects: [] }],
[schema.seasonParticipants, { selects: [[], [seasonParticipant]] }],
[schema.participants, { selects: [] }],
[schema.eventResults, { selects: [[eventResult]] }],
[schema.tournamentResults, { selects: [[]] }],
[schema.seasonParticipantSurfaceElos, { selects: [[]] }],
]);
vi.mocked(database).mockReturnValue(
makeFakeDb({ tables, inserts }) as never,
);
const report = await runBackfill({ dryRun: false });
expect(report.tournamentResultsCreated).toBe(1);
const trInserts = inserts.get(schema.tournamentResults) ?? [];
expect(trInserts).toHaveLength(1);
const inserted = trInserts[0] as Record<string, unknown>;
expect(inserted.tournamentId).toBe("t-masters");
expect(inserted.participantId).toBe("p-scottie");
expect(inserted.placement).toBe(1);
expect(inserted.rawScore).toBe("-12.00");
// CRITICAL: qualifying_points_awarded must NOT be copied.
expect(inserted).not.toHaveProperty("qualifyingPointsAwarded");
});
});
// ─── 6. Surface elo conflict detection ────────────────────────────────────
describe("runBackfill - surface elo conflicts", () => {
it("records a conflict error when two windows disagree on eloHard", async () => {
const sp = {
id: "sp-1",
sportsSeasonId: SEASON_ID,
participantId: "p-tennis-1",
name: "Carlos Alcaraz",
};
const windowElo = {
id: "se-1",
participantId: "sp-1",
sportsSeasonId: SEASON_ID,
eloHard: 2050,
eloClay: 2100,
eloGrass: 2000,
worldRanking: 2,
};
const existingCanonicalElo = {
id: "cpe-1",
participantId: "p-tennis-1",
eloHard: 2000, // differs!
eloClay: 2100,
eloGrass: 2000,
worldRanking: 2,
};
const inserts = new Map<unknown, unknown[]>();
const tables = new Map<unknown, TableState>([
[schema.sportsSeasons, { selects: [[GOLF_SEASON]] }],
[schema.scoringEvents, { selects: [[], []] }],
[schema.seasonParticipants, { selects: [[], [sp]] }],
[schema.participants, { selects: [] }],
[schema.eventResults, { selects: [] }],
[schema.seasonParticipantSurfaceElos, { selects: [[windowElo]] }],
[schema.participantSurfaceElos, { selects: [[existingCanonicalElo]] }],
]);
vi.mocked(database).mockReturnValue(
makeFakeDb({ tables, inserts }) as never,
);
const report = await runBackfill({ dryRun: false });
expect(report.surfaceElosCreated).toBe(0);
expect(report.errors).toHaveLength(1);
expect(report.errors[0]).toMatch(/conflict for participant p-tennis-1/);
expect(report.errors[0]).toMatch(/eloHard/);
// Must not have inserted a conflicting row.
expect(inserts.get(schema.participantSurfaceElos)).toBeUndefined();
});
});