feat(scrape): 入库后自动刮削缺省开启,没排上时在整理记录上记「未刮削（原因）」

- scrape_record 状态加 skipped(迁移 0074);目标不在媒体库、库根只登记、作品没有 TMDB id 也没有本地元数据、排队失败时留一条挂在整理记录上的「未刮削」记录(同原因只记一次),设置关闭不记; - 缺省开启 scrapeAfterImport,只对有 TMDB id 或本地元数据的作品排; - 已结束记录重放与收尾同一口径补排刮削(按整理记录幂等); - 整理详情「整理后刮削」显示未刮削原因,并列出失败与非预期跳过(缺元数据、认不出集号)的文件。 Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_01Cxwgnn2pbzQ7rXv3rq7fei
2026-10-03 23:19:28 +08:00
1 parent 8c4c7d028a
commit cdf62dbf02
@@ -82,10 +82,8 @@ export async function executeRecord(
    await settleSamePathReplacement(d, rec.plan as unknown as FrozenPlan, recordId);
    await cleanupReplacement(d, rec.plan as unknown as FrozenPlan, recordId);
    const frozen = rec.plan as unknown as FrozenPlan;
-    if (
-      (frozen.metadataMode === "local" || frozen.scrapeAfterImport !== undefined) &&
-      ["succeeded", "partial"].includes(rec.status)
-    ) {
+    // 与 finalizeRecord 同一口径(按设置 / 本次选项);排队按整理记录幂等,重放不会重复刮
+    if (["succeeded", "partial"].includes(rec.status)) {
      await scrapeAfterImport(d, frozen.entityId, recordId, {
        force: frozen.metadataMode === "local",
        enabled: frozen.scrapeAfterImport,
@@ -422,7 +422,55 @@ export async function scheduleEntityRescrape(d: QueueDeps, entityId: string): Pr
  return out.length;
 }

-/** 整理入库成功后(设置打开或本次要求刮削):照 MP 对作品目录刮削,集只刮本次整理的文件。 */
+/**
+ * 入库后自动刮削没排上:写一条 skipped 刮削记录挂在整理记录上(整理详情「整理后刮削」显示「未刮削（原因）」)。
+ * 同一整理记录同一原因只记一次(崩溃重放会再调)。失败只告警。
+ */
+async function recordScrapeSkipped(
+  d: AoDeps,
+  recordId: string,
+  reason: string,
+  at: { root?: { id: string; local_path: string } | undefined; path?: string; match?: ScrapeMatch | null },
+): Promise<void> {
+  try {
+    const dup = await d.ctx.db
+      .selectFrom("scrape_record")
+      .select("id")
+      .where("organization_record_id", "=", recordId)
+      .where("status", "=", "skipped")
+      .where("reason", "=", reason)
+      .executeTakeFirst();
+    if (dup) return;
+    const now = d.ports.now();
+    await d.ctx.db
+      .insertInto("scrape_record")
+      .values({
+        root_id: at.root ? `lib:${at.root.id}` : "",
+        root_path: at.root?.local_path ?? "",
+        library_root_id: at.root?.id ?? null,
+        path: at.path ?? "",
+        media_entity_id: at.match?.mediaEntityId ?? null,
+        match: at.match ? jsonb(at.match) : null,
+        status: "skipped",
+        trigger: "auto",
+        mode: (await scrapeSettings(d)).scrapeMode,
+        organization_record_id: recordId,
+        actor_type: "system",
+        reason,
+        requested_at: now,
+        finished_at: now,
+      })
+      .execute();
+  } catch (err) {
+    d.ctx.logger.warn({ recordId, reason, err: { message: errMessage(err) } }, "记录「未刮削」失败");
+  }
+}
+
+/**
+ * 整理入库成功后(设置打开或本次要求刮削):照 MP 对作品目录刮削,集只刮本次整理的文件。
+ * 只对有 TMDB id 或本地元数据(Local)的作品排;没排上(目标不在媒体库、库根只登记、作品没有元数据、排队失败)时
+ * 在整理记录上记「未刮削（原因）」;设置关闭是预期,不记。
+ */
 export async function scrapeAfterImport(
  d: AoDeps,
  entityId: string | null,
@@ -431,23 +479,11 @@ export async function scrapeAfterImport(
 ): Promise<void> {
  if (!entityId) return;
  if (opts.enabled === false) return;
+  let root: { id: string; local_path: string; management_mode: string } | undefined;
+  let match: ScrapeMatch | null = null;
  try {
    const s = await scrapeSettings(d);
    if (!s.scrapeAfterImport && !opts.force && opts.enabled !== true) return;
-    const organization = await d.ctx.db
-      .selectFrom("organization_record")
-      .select(["target_library_root_id"])
-      .where("id", "=", recordId)
-      .executeTakeFirst();
-    const match = await entityScrapeMatch(d, entityId, "library");
-    const root = organization?.target_library_root_id
-      ? await d.ctx.db
-          .selectFrom("library_root")
-          .selectAll()
-          .where("id", "=", organization.target_library_root_id)
-          .executeTakeFirst()
-      : undefined;
-    if (!match || !root || root.management_mode !== "managed") return;
    const files = await d.ctx.db
      .selectFrom("organization_file_result")
      .select("target_path")
@@ -456,6 +492,35 @@ export async function scrapeAfterImport(
      .where("status", "=", "succeeded")
      .orderBy("ordinal")
      .execute();
+    // 本次没有新入库的视频(都已在库等):没什么可刮
+    if (!files.some((f) => f.target_path)) return;
+    const organization = await d.ctx.db
+      .selectFrom("organization_record")
+      .select(["target_library_root_id"])
+      .where("id", "=", recordId)
+      .executeTakeFirst();
+    root = organization?.target_library_root_id
+      ? await d.ctx.db
+          .selectFrom("library_root")
+          .select(["id", "local_path", "management_mode"])
+          .where("id", "=", organization.target_library_root_id)
+          .executeTakeFirst()
+      : undefined;
+    match = await entityScrapeMatch(d, entityId, "library");
+    if (!root) return await recordScrapeSkipped(d, recordId, "scrape_not_library", { match });
+    if (root.management_mode !== "managed")
+      return await recordScrapeSkipped(d, recordId, "scrape_root_read_only", { root, match });
+    if (!match) return await recordScrapeSkipped(d, recordId, "scrape_no_metadata", { root });
+    if (!match.externalId) {
+      // 没有 TMDB id:只有本地元数据(Local 模式补全的字段)时才刮
+      const e = await d.ctx.db
+        .selectFrom("media_entity")
+        .select("metadata")
+        .where("id", "=", entityId)
+        .executeTakeFirst();
+      if (!Object.keys((e?.metadata ?? {}) as object).length)
+        return await recordScrapeSkipped(d, recordId, "scrape_no_metadata", { root, match });
+    }
    // 照 MP 整理后刮削:按作品目录分组,对作品目录刮(作品 / 季的 NFO 与图片照常初始化),集只刮本次整理的文件;
    // 视频直接在库根(推不出作品目录)时对文件本身刮。
    const byDir = new Map<string, string[]>();
@@ -464,20 +529,35 @@ export async function scrapeAfterImport(
      const dir = inferWorkDir(match.kind, f.target_path) ?? f.target_path;
      byDir.set(dir, [...(byDir.get(dir) ?? []), f.target_path]);
    }
-    for (const [dir, videos] of byDir)
-      await queueScrape(d, {
-        rootId: `lib:${root.id}`,
-        rootPath: root.local_path,
-        libraryRootId: root.id,
-        path: dir,
-        match,
-        trigger: "auto",
-        organizationRecordId: recordId,
-        idempotencyKey: `${SCRAPE_TASK}:record:${recordId}:${dir}`,
-        ...(dir === videos[0] ? {} : { videos }),
-      });
+    for (const [dir, videos] of byDir) {
+      try {
+        await queueScrape(d, {
+          rootId: `lib:${root.id}`,
+          rootPath: root.local_path,
+          libraryRootId: root.id,
+          path: dir,
+          match,
+          trigger: "auto",
+          organizationRecordId: recordId,
+          idempotencyKey: `${SCRAPE_TASK}:record:${recordId}:${dir}`,
+          ...(dir === videos[0] ? {} : { videos }),
+        });
+      } catch (err) {
+        // 新记录入队失败时 queueScrape 已把它标成 task_enqueue_failed(整理详情可见);这里再兜一条,防止没留下记录
+        d.ctx.logger.warn({ entityId, recordId, dir, err: { message: errMessage(err) } }, "入库后自动刮削没排上");
+        const marked = await d.ctx.db
+          .selectFrom("scrape_record")
+          .select("id")
+          .where("organization_record_id", "=", recordId)
+          .where("path", "=", dir)
+          .where("reason", "=", "task_enqueue_failed")
+          .executeTakeFirst();
+        if (!marked) await recordScrapeSkipped(d, recordId, "task_enqueue_failed", { root, path: dir, match });
+      }
+    }
  } catch (err) {
    d.ctx.logger.warn({ entityId, recordId, err: { message: errMessage(err) } }, "入库后自动刮削没排上");
+    await recordScrapeSkipped(d, recordId, "scrape_internal_error", { root, match });
  }
 }