- Sites / accounts / feeds: registry view (/site-definitions), site CRUD with
enable/disable (history stays queryable, marked site_disabled), accounts and
feed URLs stored as secrets (only secret_ref / redacted urlDisplay exposed),
feed enable/disable/reset/fetch; one shared rate limiter + adapter per site.
- Tasks: ingestFeed (FeedFetchState/FeedCursor on feed_source, observation
feed+guid upsert with digest-change updates, source_release site+id
aggregation across feeds, auth failure -> account invalid + decision, no
retry, gap -> feed.gap_suspected -> budgeted backfill snapshot),
resolveObservation (parseRelease -> local recall via title_alias pg_trgm +
accepted identities -> TMDB -> resolveIdentity; auto binds entity /
provisional units / identities / aliases, otherwise identity_candidate pool
with evidence version and bounded review backoff; manual lock respected),
refreshMetadata (metadata_field + field_override, user locks win, paths
untouched), snapshotSite (site_snapshot coverage records).
- Events: feed.due, feed.gap_suspected, release.observed (self-consumed),
release.resolved, decision.created/updated; scheduler in worker start().
- Identity API: entity detail with redirects, manual mapping correction with
identity_revision, merge with identity_redirect, pool list/assign/dismiss/
retry, content unit numbering lock, aliases, external identities.
- Decisions: list/summary (by entity and kind)/detail/resolve; identity
options executed directly.
- Workbench: /sites/{id}/releases (snapshot first, coalesced live search with
backoff, origin + freshness) and /sites/{id}/releases/{rid} (volatile field
recheck times, refresh keeps data on failure).
- Capabilities: catalog.lookup, release.search, release.inspect,
identity.resolve.
- Migrations 0018-0020: site.adapter api, source_release.resolution_locked,
site_snapshot, identity_candidate, resolution_attempt, identity_redirect,
metadata_field, field_override.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
404 lines
18 KiB
TypeScript
404 lines
18 KiB
TypeScript
import { CapabilityRegistry } from "@umt/agent";
|
|
import { createMockTmdb, type MockTmdb, TMDB_FIXTURES, type TmdbFixture } from "@umt/testkit";
|
|
import { afterAll, beforeAll, describe, expect, it } from "vitest";
|
|
import { refreshMetadata } from "../../src/modules/sources-identity/service/metadata.ts";
|
|
import { resolveRelease } from "../../src/modules/sources-identity/service/resolve.ts";
|
|
import { createHarness, type Harness, json, pgAvailable, seedSite } from "./harness.ts";
|
|
|
|
const up = await pgAvailable();
|
|
|
|
const twin = (id: number): TmdbFixture => ({
|
|
type: "movie",
|
|
id,
|
|
original_title: "Twins",
|
|
original_language: "en",
|
|
origin_country: ["US"],
|
|
date: "1988-12-09",
|
|
titles: { "zh-CN": "龙兄鼠弟", "en-US": "Twins" },
|
|
popularity: 10 + id / 1000,
|
|
alternative_titles: [],
|
|
external_ids: {},
|
|
});
|
|
|
|
describe.skipIf(!up)("resolveObservation / 身份与元数据", () => {
|
|
let h: Harness;
|
|
let tmdb: MockTmdb;
|
|
let siteId: string;
|
|
let n = 0;
|
|
|
|
beforeAll(async () => {
|
|
tmdb = await createMockTmdb({ fixtures: [...TMDB_FIXTURES, twin(111), twin(112)] });
|
|
h = await createHarness("resolve", { tmdbUrl: tmdb.url });
|
|
siteId = (await seedSite(h, { baseUrl: "https://audiences.example/" })).site.id;
|
|
});
|
|
afterAll(async () => {
|
|
await h?.close();
|
|
await tmdb?.close();
|
|
});
|
|
|
|
const release = async (title: string, extra: { subtitle?: string; category?: string } = {}) =>
|
|
h.db
|
|
.insertInto("source_release")
|
|
.values({
|
|
site_id: siteId,
|
|
site_release_id: String(700000 + ++n),
|
|
title,
|
|
subtitle: extra.subtitle ?? null,
|
|
category_raw: extra.category ?? null,
|
|
})
|
|
.returningAll()
|
|
.executeTakeFirstOrThrow();
|
|
|
|
let oppId: string;
|
|
|
|
it("auto:按 TMDB 建本地作品 + accepted 外部身份 + 别名,发 release.resolved", async () => {
|
|
const r = await release("Oppenheimer.2023.2160p.UHD.BluRay.REMUX.HDR.HEVC-FGT");
|
|
const res = await resolveRelease(h.rt, { sourceReleaseId: r.id });
|
|
expect(res.outcome).toBe("auto");
|
|
oppId = res.mediaEntityId!;
|
|
const e = await h.db.selectFrom("media_entity").selectAll().where("id", "=", oppId).executeTakeFirstOrThrow();
|
|
expect(e).toMatchObject({ kind: "movie", year: 2023, status: "active" });
|
|
const ids = await h.db.selectFrom("external_identity").selectAll().where("media_entity_id", "=", oppId).execute();
|
|
expect(ids.map((i) => `${i.provider}:${i.entity_type}:${i.external_id}:${i.status}`).sort()).toEqual([
|
|
"imdb:movie:tt15398776:accepted",
|
|
"tmdb:movie:872585:accepted",
|
|
]);
|
|
const revs = await h.db.selectFrom("identity_revision").selectAll().execute();
|
|
expect(revs.length).toBeGreaterThanOrEqual(2);
|
|
const aliases = await h.db
|
|
.selectFrom("title_alias")
|
|
.select(["alias"])
|
|
.where("media_entity_id", "=", oppId)
|
|
.execute();
|
|
expect(aliases.map((a) => a.alias)).toEqual(expect.arrayContaining(["奥本海默", "Oppenheimer"]));
|
|
const rel = await h.db.selectFrom("source_release").selectAll().where("id", "=", r.id).executeTakeFirstOrThrow();
|
|
expect(rel).toMatchObject({ resolution_status: "resolved", media_entity_id: oppId, parser_version: "1" });
|
|
expect((rel.parsed as { year: number }).year).toBe(2023);
|
|
const ev = (await h.outbox("release.resolved")).find((x) => x.payload.sourceReleaseId === r.id);
|
|
expect(ev?.payload).toMatchObject({ mediaEntityId: oppId, evidenceLevel: "rule_confirmed", contentUnitIds: [] });
|
|
// 幂等:证据未变的重复识别直接短路
|
|
expect((await resolveRelease(h.rt, { sourceReleaseId: r.id })).outcome).toBe("skipped");
|
|
});
|
|
|
|
it("同一作品的其它资源经本地召回复用同一实体,不重复建档", async () => {
|
|
const r = await release("奥本海默.Oppenheimer.2023.1080p.WEB-DL.H264-GRP");
|
|
const res = await resolveRelease(h.rt, { sourceReleaseId: r.id });
|
|
expect(res).toMatchObject({ outcome: "auto", mediaEntityId: oppId });
|
|
const count = await h.db
|
|
.selectFrom("media_entity")
|
|
.select((eb) => eb.fn.countAll<number>().as("c"))
|
|
.executeTakeFirstOrThrow();
|
|
expect(Number(count.c)).toBe(1);
|
|
});
|
|
|
|
it("剧集:建 series 作品与临时季/集单元", async () => {
|
|
const r = await release("The.Bear.S03E02.2160p.DSNP.WEB-DL.DDP5.1.H.265-FLUX", { subtitle: "熊家餐馆 第三季" });
|
|
const res = await resolveRelease(h.rt, { sourceReleaseId: r.id });
|
|
expect(res.outcome).toBe("auto");
|
|
const units = await h.db
|
|
.selectFrom("content_unit")
|
|
.selectAll()
|
|
.where("media_entity_id", "=", res.mediaEntityId!)
|
|
.orderBy("kind")
|
|
.execute();
|
|
expect(units.map((u) => `${u.kind}:${u.season_number}:${u.episode_number}:${u.status}`)).toEqual([
|
|
"episode:3:2:provisional",
|
|
"season:3:null:provisional",
|
|
]);
|
|
expect(res.contentUnitIds).toEqual([units[0]!.id]);
|
|
const e = await h.db
|
|
.selectFrom("media_entity")
|
|
.select(["kind"])
|
|
.where("id", "=", res.mediaEntityId!)
|
|
.executeTakeFirstOrThrow();
|
|
expect(e.kind).toBe("series");
|
|
});
|
|
|
|
let twinsRelease: string;
|
|
it("ambiguous:进待识别池 + 待处理事项,不建正式作品", async () => {
|
|
const before = await h.db.selectFrom("media_entity").select("id").execute();
|
|
const r = await release("Twins.1988.1080p.BluRay.x264-GRP");
|
|
twinsRelease = r.id;
|
|
const res = await resolveRelease(h.rt, { sourceReleaseId: r.id });
|
|
expect(res.outcome).toBe("ambiguous");
|
|
expect((await h.db.selectFrom("media_entity").select("id").execute()).length).toBe(before.length);
|
|
const pool = await h.db
|
|
.selectFrom("identity_candidate")
|
|
.selectAll()
|
|
.where("source_release_id", "=", r.id)
|
|
.executeTakeFirstOrThrow();
|
|
expect(pool).toMatchObject({ status: "pending", reason: "ambiguous", attempts: 1 });
|
|
expect((pool.candidates as unknown[]).length).toBe(2);
|
|
expect(pool.next_review_at!.getTime()).toBeGreaterThan(Date.now());
|
|
const d = await h.db.selectFrom("decision").selectAll().where("id", "=", res.decisionId!).executeTakeFirstOrThrow();
|
|
expect(d).toMatchObject({ kind: "identity_ambiguous", status: "open", subject_id: r.id });
|
|
expect((d.options as { key: string }[]).map((o) => o.key)).toEqual(["c0", "c1", "dismiss"]);
|
|
expect((await h.outbox("decision.created")).some((e) => e.payload.decisionId === d.id)).toBe(true);
|
|
const rel = await h.db.selectFrom("source_release").selectAll().where("id", "=", r.id).executeTakeFirstOrThrow();
|
|
expect(rel.resolution_status).toBe("ambiguous");
|
|
expect(rel.media_entity_id).toBeNull();
|
|
const attempts = await h.db
|
|
.selectFrom("resolution_attempt")
|
|
.selectAll()
|
|
.where("source_release_id", "=", r.id)
|
|
.execute();
|
|
expect(attempts.map((a) => a.outcome)).toEqual(["ambiguous"]);
|
|
// 同一证据再次复查:计数递增(退避加长),不重复开待处理
|
|
await resolveRelease(h.rt, { sourceReleaseId: r.id, force: true });
|
|
const pool2 = await h.db
|
|
.selectFrom("identity_candidate")
|
|
.selectAll()
|
|
.where("source_release_id", "=", r.id)
|
|
.executeTakeFirstOrThrow();
|
|
expect(pool2.attempts).toBe(2);
|
|
expect(pool2.next_review_at!.getTime()).toBeGreaterThan(pool.next_review_at!.getTime());
|
|
expect(
|
|
await h.db.selectFrom("decision").select("id").where("dedupe_key", "=", `identity:${r.id}`).execute(),
|
|
).toHaveLength(1);
|
|
});
|
|
|
|
it("待识别池可按原始标题检索;经待处理中心选候选 = 人工指定并加锁", async () => {
|
|
const list = await json(await h.req("/identity-pool?q=Twins"));
|
|
expect(list.status).toBe(200);
|
|
expect(list.body.data.items.map((i: { sourceReleaseId: string }) => i.sourceReleaseId)).toContain(twinsRelease);
|
|
const pool = list.body.data.items.find((i: { sourceReleaseId: string }) => i.sourceReleaseId === twinsRelease);
|
|
const res = await json(await h.req(`/decisions/${pool.decisionId}/resolve`, { body: { optionKey: "c1" } }));
|
|
expect(res.status).toBe(200);
|
|
expect(res.body.data.status).toBe("resolved");
|
|
const rel = await h.db
|
|
.selectFrom("source_release")
|
|
.selectAll()
|
|
.where("id", "=", twinsRelease)
|
|
.executeTakeFirstOrThrow();
|
|
expect(rel.resolution_locked).toBe(true);
|
|
const chosen = pool.candidates[1];
|
|
const ident = await h.db
|
|
.selectFrom("external_identity")
|
|
.selectAll()
|
|
.where("media_entity_id", "=", rel.media_entity_id!)
|
|
.where("provider", "=", "tmdb")
|
|
.executeTakeFirstOrThrow();
|
|
expect(ident).toMatchObject({
|
|
external_id: chosen.externalId,
|
|
evidence_level: "manual",
|
|
locked: true,
|
|
status: "accepted",
|
|
});
|
|
const p = await h.db
|
|
.selectFrom("identity_candidate")
|
|
.selectAll()
|
|
.where("source_release_id", "=", twinsRelease)
|
|
.executeTakeFirstOrThrow();
|
|
expect(p.status).toBe("resolved");
|
|
const ev = (await h.outbox("release.resolved")).find((x) => x.payload.sourceReleaseId === twinsRelease);
|
|
expect(ev?.payload.evidenceLevel).toBe("manual");
|
|
// 人工锁:自动识别不再改写
|
|
const again = await resolveRelease(h.rt, { sourceReleaseId: twinsRelease, force: true });
|
|
expect(again.outcome).toBe("skipped_locked");
|
|
const after = await h.db
|
|
.selectFrom("source_release")
|
|
.select(["media_entity_id"])
|
|
.where("id", "=", twinsRelease)
|
|
.executeTakeFirstOrThrow();
|
|
expect(after.media_entity_id).toBe(rel.media_entity_id);
|
|
});
|
|
|
|
it("no_match 只进池不开待处理;TMDB 失败记为 lookup_failed 而不是无结果", async () => {
|
|
const r1 = await release("Zzqxv.Nonexistent.2019.1080p.WEB-DL-NOPE");
|
|
const a = await resolveRelease(h.rt, { sourceReleaseId: r1.id });
|
|
expect(a.outcome).toBe("no_match");
|
|
expect(a.decisionId).toBeNull();
|
|
const rel = await h.db
|
|
.selectFrom("source_release")
|
|
.select(["resolution_status"])
|
|
.where("id", "=", r1.id)
|
|
.executeTakeFirstOrThrow();
|
|
expect(rel.resolution_status).toBe("unidentifiable");
|
|
|
|
tmdb.fault(/\/search\//, { status: 500, times: 20 });
|
|
const r2 = await release("Qwerty.Unknown.2020.1080p.WEB-DL-NOPE");
|
|
const b = await resolveRelease(h.rt, { sourceReleaseId: r2.id });
|
|
expect(b.outcome).toBe("lookup_failed");
|
|
const pool = await h.db
|
|
.selectFrom("identity_candidate")
|
|
.selectAll()
|
|
.where("source_release_id", "=", r2.id)
|
|
.executeTakeFirstOrThrow();
|
|
expect(pool.reason).toBe("lookup_failed");
|
|
// 查询失败的复查间隔短于无结果
|
|
expect(pool.next_review_at!.getTime() - Date.now()).toBeLessThan(20 * 60_000);
|
|
tmdb.clearFaults();
|
|
});
|
|
|
|
it("自动化关闭:证据足够也只提议(proposed),进池待确认", async () => {
|
|
await h.db
|
|
.insertInto("settings")
|
|
.values({ key: "automation", value: JSON.stringify({ enabled: false }) } as never)
|
|
.execute();
|
|
const r = await release("Dune.Part.Two.2024.2160p.UHD.BluRay.REMUX-FraMeSToR", { subtitle: "沙丘2" });
|
|
const res = await resolveRelease(h.rt, { sourceReleaseId: r.id });
|
|
expect(res.outcome).toBe("proposed");
|
|
expect(res.decisionId).not.toBeNull();
|
|
await h.db.deleteFrom("settings").where("key", "=", "automation").execute();
|
|
});
|
|
|
|
it("refreshMetadata:字段级刷新;用户锁定字段不被覆盖,解锁后回到来源值", async () => {
|
|
const r1 = await refreshMetadata(h.rt, oppId);
|
|
expect(r1.status).toBe("refreshed");
|
|
const fields = await h.db.selectFrom("metadata_field").selectAll().where("media_entity_id", "=", oppId).execute();
|
|
expect(fields.map((f) => f.field)).toEqual(expect.arrayContaining(["title", "overview", "year"]));
|
|
let e = await h.db.selectFrom("media_entity").selectAll().where("id", "=", oppId).executeTakeFirstOrThrow();
|
|
expect(e.title).toBe("奥本海默");
|
|
expect((e.metadata as Record<string, { source: string }>).overview?.source).toBe("tmdb");
|
|
|
|
const put = await json(
|
|
await h.req(`/media-entities/${oppId}/fields/title`, {
|
|
method: "PUT",
|
|
body: { value: "奥本海默(IMAX)", reason: "偏好" },
|
|
}),
|
|
);
|
|
expect(put.status).toBe(200);
|
|
expect(put.body.data.lockedFields).toContain("title");
|
|
const r2 = await refreshMetadata(h.rt, oppId);
|
|
expect(r2.lockedFields).toContain("title");
|
|
e = await h.db.selectFrom("media_entity").selectAll().where("id", "=", oppId).executeTakeFirstOrThrow();
|
|
expect(e.title).toBe("奥本海默(IMAX)");
|
|
expect(e.field_locks).toContain("title");
|
|
expect((e.metadata as Record<string, { source: string; locked: boolean }>).title).toMatchObject({
|
|
source: "user",
|
|
locked: true,
|
|
});
|
|
|
|
const del = await json(await h.req(`/media-entities/${oppId}/fields/title`, { method: "DELETE" }));
|
|
expect(del.status).toBe(200);
|
|
e = await h.db.selectFrom("media_entity").selectAll().where("id", "=", oppId).executeTakeFirstOrThrow();
|
|
expect(e.title).toBe("奥本海默");
|
|
expect(e.field_locks).not.toContain("title");
|
|
});
|
|
|
|
it("人工纠错映射:旧映射退役留修订,新映射 accepted + 锁;自动匹配不改回", async () => {
|
|
const res = await json(
|
|
await h.req(`/media-entities/${oppId}/identities`, {
|
|
body: { provider: "tmdb", entityType: "movie", externalId: "693134", reason: "误匹配" },
|
|
}),
|
|
);
|
|
expect(res.status).toBe(200);
|
|
expect(res.body.data.changed).toBe(true);
|
|
const ids = await h.db
|
|
.selectFrom("external_identity")
|
|
.selectAll()
|
|
.where("media_entity_id", "=", oppId)
|
|
.where("provider", "=", "tmdb")
|
|
.execute();
|
|
const old = ids.find((i) => i.external_id === "872585")!;
|
|
expect(old).toMatchObject({ status: "retired", locked: true });
|
|
expect(ids.find((i) => i.external_id === "693134")).toMatchObject({
|
|
status: "accepted",
|
|
locked: true,
|
|
evidence_level: "manual",
|
|
});
|
|
const revs = await json(await h.req(`/external-identities/${old.id}/revisions`));
|
|
expect(revs.body.data.items.map((r: { action: string }) => r.action)).toEqual(["created", "status_changed"]);
|
|
|
|
// 新来一条奥本海默资源:TMDB 872585 不再自动绑回该作品
|
|
const r = await release("Oppenheimer.2023.1080p.BluRay.x264-OTHER");
|
|
await resolveRelease(h.rt, { sourceReleaseId: r.id });
|
|
const accepted = await h.db
|
|
.selectFrom("external_identity")
|
|
.select(["media_entity_id"])
|
|
.where("provider", "=", "tmdb")
|
|
.where("external_id", "=", "872585")
|
|
.where("status", "=", "accepted")
|
|
.execute();
|
|
expect(accepted.every((a) => a.media_entity_id !== oppId)).toBe(true);
|
|
const still = await h.db
|
|
.selectFrom("external_identity")
|
|
.select(["external_id"])
|
|
.where("media_entity_id", "=", oppId)
|
|
.where("provider", "=", "tmdb")
|
|
.where("status", "=", "accepted")
|
|
.executeTakeFirstOrThrow();
|
|
expect(still.external_id).toBe("693134");
|
|
|
|
// 目标身份已属于其它作品 → 409
|
|
const bear = await h.db
|
|
.selectFrom("external_identity")
|
|
.select(["media_entity_id"])
|
|
.where("external_id", "=", "136315")
|
|
.executeTakeFirstOrThrow();
|
|
const conflict = await json(
|
|
await h.req(`/media-entities/${oppId}/identities`, {
|
|
body: { provider: "tmdb", entityType: "tv", externalId: "136315" },
|
|
}),
|
|
);
|
|
expect(conflict.status).toBe(409);
|
|
expect(bear.media_entity_id).not.toBe(oppId);
|
|
});
|
|
|
|
it("合并重复作品:旧 ID 跳转,来源资源与别名迁到目标", async () => {
|
|
const dup = await json(
|
|
await h.req("/media-entities", { body: { kind: "movie", title: "奥本海默 重复", year: 2023 } }),
|
|
);
|
|
expect(dup.status).toBe(200);
|
|
const dupId = dup.body.data.id;
|
|
const r = await release("Dup.Title.2023.1080p-X");
|
|
await h.db.updateTable("source_release").set({ media_entity_id: dupId }).where("id", "=", r.id).execute();
|
|
const m = await json(await h.req(`/media-entities/${dupId}/merge`, { body: { intoId: oppId, reason: "重复" } }));
|
|
expect(m.status).toBe(200);
|
|
expect(m.body.data.id).toBe(oppId);
|
|
const detail = await json(await h.req(`/media-entities/${dupId}/detail`));
|
|
expect(detail.body.data.entity.id).toBe(oppId);
|
|
expect(detail.body.data.redirectedFrom).toBe(dupId);
|
|
expect(detail.body.data.aliases.map((a: { alias: string }) => a.alias)).toContain("奥本海默 重复");
|
|
const moved = await h.db
|
|
.selectFrom("source_release")
|
|
.select(["media_entity_id"])
|
|
.where("id", "=", r.id)
|
|
.executeTakeFirstOrThrow();
|
|
expect(moved.media_entity_id).toBe(oppId);
|
|
const redirect = await h.db
|
|
.selectFrom("identity_redirect")
|
|
.selectAll()
|
|
.where("from_entity_id", "=", dupId)
|
|
.executeTakeFirstOrThrow();
|
|
expect(redirect.to_entity_id).toBe(oppId);
|
|
// 再合并一次 → 409
|
|
expect((await h.req(`/media-entities/${dupId}/merge`, { body: { intoId: oppId } })).status).toBe(409);
|
|
});
|
|
|
|
it("待处理中心聚合与列表", async () => {
|
|
const s = await json(await h.req("/decisions/summary"));
|
|
expect(s.status).toBe(200);
|
|
expect(s.body.data.total).toBeGreaterThanOrEqual(1);
|
|
expect(s.body.data.groups.some((g: { kind: string }) => g.kind === "identity_ambiguous")).toBe(true);
|
|
const l = await json(await h.req("/decisions?kind=identity_ambiguous"));
|
|
expect(
|
|
l.body.data.items.every((d: { status: string }) => ["open", "investigating", "awaiting_user"].includes(d.status)),
|
|
).toBe(true);
|
|
// 已关闭的事项不可再处理
|
|
const closed = await h.db
|
|
.selectFrom("decision")
|
|
.select(["id"])
|
|
.where("status", "=", "resolved")
|
|
.executeTakeFirstOrThrow();
|
|
expect((await h.req(`/decisions/${closed.id}/resolve`, { body: { dismiss: true } })).status).toBe(409);
|
|
});
|
|
|
|
it("能力:catalog.lookup / release.search 可调用;identity.resolve(propose)未授权被拒", async () => {
|
|
const reg = new CapabilityRegistry({});
|
|
h.module.registerCapabilities!(reg, h.ctx);
|
|
const c = { actorId: "t", taskId: "t", scopeIds: [], signal: new AbortController().signal };
|
|
const look = (await reg.invoke("catalog.lookup", { query: "奥本海默", includeRemote: true }, c)) as {
|
|
local: { entity: { id: string } }[];
|
|
remote: { externalId: string }[];
|
|
};
|
|
expect(look.local.map((l) => l.entity.id)).toContain(oppId);
|
|
expect(look.remote.map((r) => r.externalId)).toContain("872585");
|
|
const search = (await reg.invoke("release.search", { query: "Oppenheimer", allowLiveSearch: false }, c)) as {
|
|
items: unknown[];
|
|
};
|
|
expect(search.items.length).toBeGreaterThan(0);
|
|
await expect(reg.invoke("identity.resolve", { sourceReleaseId: twinsRelease }, c)).rejects.toThrow();
|
|
});
|
|
});
|