fix: 合并服务器未推送改动并保留手动视频标签

2026-10-03 22:09:09 +08:00
parent 6f0e38054b b9af979582
commit d3896522ec
29 changed files with 1459 additions and 532 deletions
@@ -0,0 +1 @@
+export { GET } from "../../stt/videos/route";
@@ -31,6 +31,7 @@ export async function GET() {
      author: {
        nickname: v.author.nickname,
      },
+      content_tags: v.manual_content_tags,
      transcription: transcriptionState(v.transcript, v.transcriptionTask),
    }));

@@ -0,0 +1 @@
+export { GET, POST } from "../stt/route";
@@ -0,0 +1,121 @@
+export const runtime = "nodejs";
+
+import { type NextRequest, NextResponse } from "next/server";
+import { isVideoContentTag } from "@/lib/video-tags";
+import { prisma } from "@/lib/prisma";
+
+type UpdateMode = "set" | "add";
+
+function uniqueValidTags(tags: unknown[]) {
+  const validTags: string[] = [];
+  for (const tag of tags) {
+    if (typeof tag !== "string" || !isVideoContentTag(tag)) {
+      throw new Error("Invalid tag");
+    }
+    if (!validTags.includes(tag)) validTags.push(tag);
+  }
+  return validTags;
+}
+
+export async function POST(req: NextRequest) {
+  let body: unknown;
+  try {
+    body = await req.json();
+  } catch {
+    return NextResponse.json({ error: "Invalid JSON body" }, { status: 400 });
+  }
+
+  if (!body || typeof body !== "object") {
+    return NextResponse.json(
+      { error: "Invalid request body" },
+      { status: 400 },
+    );
+  }
+
+  const raw = body as {
+    awemeId?: unknown;
+    awemeIds?: unknown;
+    tag?: unknown;
+    contentTags?: unknown;
+    mode?: unknown;
+  };
+  const awemeIds = Array.isArray(raw.awemeIds)
+    ? raw.awemeIds
+    : typeof raw.awemeId === "string"
+      ? [raw.awemeId]
+      : [];
+  const normalizedAwemeIds = [...new Set(awemeIds)]
+    .filter(
+      (id): id is string => typeof id === "string" && id.trim().length > 0,
+    )
+    .map((id) => id.trim());
+
+  if (normalizedAwemeIds.length === 0 || normalizedAwemeIds.length > 200) {
+    return NextResponse.json(
+      { error: "awemeIds must contain 1 to 200 ids" },
+      { status: 400 },
+    );
+  }
+
+  const mode: UpdateMode = raw.mode === "add" ? "add" : "set";
+  const rawTags = Array.isArray(raw.contentTags)
+    ? raw.contentTags
+    : typeof raw.tag === "string"
+      ? [raw.tag]
+      : [];
+
+  let contentTags: string[];
+  try {
+    contentTags = uniqueValidTags(rawTags);
+  } catch {
+    return NextResponse.json({ error: "Invalid tag" }, { status: 400 });
+  }
+
+  if (mode === "add" && contentTags.length === 0) {
+    return NextResponse.json(
+      { error: "add mode requires at least one tag" },
+      { status: 400 },
+    );
+  }
+
+  const videos = await prisma.video.findMany({
+    where: { aweme_id: { in: normalizedAwemeIds } },
+    select: {
+      aweme_id: true,
+      manual_content_tags: true,
+    },
+  });
+
+  if (videos.length === 0) {
+    return NextResponse.json({ error: "No videos found" }, { status: 404 });
+  }
+
+  const updates = videos.map((video) => {
+    const baseTags = video.manual_content_tags;
+    const manualTags =
+      mode === "add"
+        ? [...new Set([...baseTags, ...contentTags])]
+        : contentTags;
+    return { awemeId: video.aweme_id, manualTags };
+  });
+
+  await prisma.$transaction(
+    updates.map((update) =>
+      prisma.video.update({
+        where: { aweme_id: update.awemeId },
+        data: { manual_content_tags: update.manualTags },
+      }),
+    ),
+  );
+
+  return NextResponse.json({
+    updated: updates.map((update) => ({
+      aweme_id: update.awemeId,
+      manual_content_tags: update.manualTags,
+      content_tags: update.manualTags,
+    })),
+    missing: normalizedAwemeIds.filter(
+      (id) => !videos.some((video) => video.aweme_id === id),
+    ),
+  });
+}
@@ -1,7 +1,8 @@
-import { NextRequest, NextResponse } from "next/server";
-import { prisma } from "@/lib/prisma";
+import { type NextRequest, NextResponse } from "next/server";
 import type { FeedItem, FeedResponse } from "@/app/types/feed";
+import { isVideoContentTag } from "@/lib/video-tags";
 import { getFileUrl } from "@/lib/minio";
+import { prisma } from "@/lib/prisma";

 // Contract
 // Inputs: search params { before?: ISOString, limit?: number }
@@ -11,24 +12,45 @@ export async function GET(req: NextRequest) {
  const { searchParams } = new URL(req.url);
  const limitParam = searchParams.get("limit");
  const beforeParam = searchParams.get("before");
+  const rawTag = (searchParams.get("tag") || "").trim();
+  const selectedTag = rawTag && isVideoContentTag(rawTag) ? rawTag : null;
+
+  if (rawTag && !selectedTag) {
+    return NextResponse.json({ error: "Invalid tag" }, { status: 400 });
+  }

  const limit = Math.min(Math.max(Number(limitParam ?? "24"), 1), 60); // 1..60
  const before = beforeParam ? new Date(beforeParam) : null;
+  const videoWhere = {
+    ...(before ? { created_at: { lt: before } } : {}),
+    ...(selectedTag
+      ? {
+          manual_content_tags: { has: selectedTag },
+        }
+      : {}),
+  };

  // fetch chunk from both tables
  const [videos, posts] = await Promise.all([
    prisma.video.findMany({
-      where: before ? { created_at: { lt: before } } : undefined,
+      where: videoWhere,
      orderBy: { created_at: "desc" },
      take: limit,
-      include: { author: true },
-    }),
-    prisma.imagePost.findMany({
-      where: before ? { created_at: { lt: before } } : undefined,
-      orderBy: { created_at: "desc" },
-      take: limit,
-      include: { author: true, images: { orderBy: { order: "asc" }, take: 1 } },
+      include: {
+        author: true,
+      },
    }),
+    selectedTag
+      ? Promise.resolve([])
+      : prisma.imagePost.findMany({
+          where: before ? { created_at: { lt: before } } : undefined,
+          orderBy: { created_at: "desc" },
+          take: limit,
+          include: {
+            author: true,
+            images: { orderBy: { order: "asc" }, take: 1 },
+          },
+        }),
  ]);

  const merged: FeedItem[] = [
@@ -47,29 +69,33 @@ export async function GET(req: NextRequest) {
        sec_uid: v.author.sec_uid,
      },
      likes: Number(v.digg_count),
+      content_tags: v.manual_content_tags,
    })),
-    ...posts.map((p) => ({
-      type: "image" as const,
-      aweme_id: p.aweme_id,
-      created_at: p.created_at,
-      desc: p.desc,
-      cover_url: getFileUrl(p.images?.[0]?.url ?? null),
-      width: p.images?.[0]?.width ?? null,
-      height: p.images?.[0]?.height ?? null,
-      author: {
-        nickname: p.author.nickname,
-        avatar_url: getFileUrl(p.author.avatar_url ?? ""),
-        sec_uid: p.author.sec_uid,
-      },
-      likes: Number(p.digg_count),
-    })),
+    ...posts.map((p) => {
+      const coverPath = p.images?.[0]?.url;
+      return {
+        type: "image" as const,
+        aweme_id: p.aweme_id,
+        created_at: p.created_at,
+        desc: p.desc,
+        cover_url: coverPath ? getFileUrl(coverPath) : null,
+        width: p.images?.[0]?.width ?? null,
+        height: p.images?.[0]?.height ?? null,
+        author: {
+          nickname: p.author.nickname,
+          avatar_url: getFileUrl(p.author.avatar_url ?? ""),
+          sec_uid: p.author.sec_uid,
+        },
+        likes: Number(p.digg_count),
+      };
+    }),
  ]
    .sort((a, b) => +new Date(b.created_at) - +new Date(a.created_at))
    .slice(0, limit);

  const nextCursor =
    merged.length > 0
-      ? new Date(merged[merged.length - 1].created_at as any).toISOString()
+      ? new Date(merged[merged.length - 1].created_at).toISOString()
      : null;
  const payload: FeedResponse = { items: merged, nextCursor };
  return NextResponse.json(payload);
@@ -1,11 +1,32 @@
+import { NextResponse } from "next/server";
+import { isVideoContentTag } from "@/lib/video-tags";
 import { json } from "@/lib/json";
 import { getFileUrl } from "@/lib/minio";
-import { prisma } from "@/lib/prisma"; // 你的 Prisma 客户端实例
-import { NextResponse } from "next/server";
+import { prisma } from "@/lib/prisma";
+
+type SearchRow = {
+  id: string;
+  awemeId: string;
+  type: "video" | "image";
+  rank: number;
+  snippet: string;
+  contentTags: string[];
+};
+
+function sanitizeSnippetHtml(snippet: string) {
+  return snippet
+    .replaceAll("&", "&amp;")
+    .replaceAll("<", "&lt;")
+    .replaceAll(">", "&gt;")
+    .replaceAll("&lt;mark&gt;", "<mark>")
+    .replaceAll("&lt;/mark&gt;", "</mark>");
+}

 export async function GET(req: Request) {
  const { searchParams } = new URL(req.url);
  const q = (searchParams.get("q") || "").trim();
+  const rawTag = (searchParams.get("tag") || "").trim();
+  const contentTag = rawTag && isVideoContentTag(rawTag) ? rawTag : null;
  const page = Math.max(1, Number(searchParams.get("page") || 1));
  const limit = Math.min(
    50,
@@ -13,7 +34,11 @@ export async function GET(req: Request) {
  );
  const offset = (page - 1) * limit;

-  if (!q) {
+  if (rawTag && !contentTag) {
+    return NextResponse.json({ error: "Invalid tag" }, { status: 400 });
+  }
+
+  if (!q && !contentTag) {
    return NextResponse.json({ results: [], total: 0 });
  }

@@ -23,90 +48,117 @@ export async function GET(req: Request) {
    // - 合并搜索来源：VideoTranscript.transcript_zh_tsv、Video.desc 与 ImagePost.desc
    // - 使用 UNION ALL 合并视频和图文的搜索结果
    // - rank 取两者匹配得分的最大值，snippet 优先展示转写匹配，否则展示 desc 匹配
-    const rows = await prisma.$queryRaw<
-      {
-        id: string;
-        awemeId: string;
-        type: "video" | "image";
-        rank: number;
-        snippet: string;
-      }[]
-    >`
-      WITH tsq AS (
-        SELECT websearch_to_tsquery('zhcfg', ${q}) AS query
-      )
-      -- 视频搜索结果
-      SELECT
-        COALESCE(vt.id, 'desc:' || v.aweme_id) AS id,
-        v.aweme_id AS "awemeId",
-        'video'::text AS type,
-        GREATEST(
-          COALESCE(ts_rank(vt.transcript_zh_tsv, tsq.query), 0),
-          COALESCE(ts_rank(to_tsvector('zhcfg', v.desc), tsq.query), 0)
-        ) AS rank,
-        CASE
-          WHEN vt.transcript_zh_tsv @@ tsq.query THEN
+    const rows = q
+      ? await prisma.$queryRaw<SearchRow[]>`
+          WITH tsq AS (
+            SELECT websearch_to_tsquery('zhcfg', ${q}) AS query
+          )
+          -- 视频搜索结果
+          SELECT
+            COALESCE(vt.id, 'desc:' || v.aweme_id) AS id,
+            v.aweme_id AS "awemeId",
+            'video'::text AS type,
+            GREATEST(
+              COALESCE(ts_rank(vt.transcript_zh_tsv, tsq.query), 0),
+              COALESCE(ts_rank(to_tsvector('zhcfg', v.desc), tsq.query), 0)
+            ) AS rank,
+            CASE
+              WHEN vt.transcript_zh_tsv @@ tsq.query THEN
+                ts_headline(
+                  'zhcfg',
+                  array_to_string(vt.transcript, ' '),
+                  tsq.query,
+                  'StartSel=<mark>,StopSel=</mark>,MaxFragments=2,MinWords=2,MaxWords=20'
+                )
+              ELSE
+                ts_headline(
+                  'zhcfg',
+                  v.desc,
+                  tsq.query,
+                  'StartSel=<mark>,StopSel=</mark>,MaxFragments=2,MinWords=2,MaxWords=20'
+                )
+            END AS snippet,
+            v.manual_content_tags AS "contentTags"
+          FROM "Video" v
+          LEFT JOIN "VideoTranscript" vt ON vt."videoId" = v.aweme_id,
+               tsq
+          WHERE
+            (vt.transcript_zh_tsv @@ tsq.query OR to_tsvector('zhcfg', v.desc) @@ tsq.query)
+            AND CASE
+              WHEN ${contentTag}::text IS NULL THEN TRUE
+              ELSE v.manual_content_tags @> ARRAY[${contentTag}]::text[]
+            END
+
+          UNION ALL
+
+          -- 图文搜索结果
+          SELECT
+            'img_desc:' || ip.aweme_id AS id,
+            ip.aweme_id AS "awemeId",
+            'image'::text AS type,
+            ts_rank(to_tsvector('zhcfg', ip.desc), tsq.query) AS rank,
            ts_headline(
              'zhcfg',
-              array_to_string(vt.transcript, ' '),
+              ip.desc,
              tsq.query,
              'StartSel=<mark>,StopSel=</mark>,MaxFragments=2,MinWords=2,MaxWords=20'
-            )
-          ELSE
-            ts_headline(
-              'zhcfg',
+            ) AS snippet,
+            ARRAY[]::text[] AS "contentTags"
+          FROM "ImagePost" ip, tsq
+          WHERE to_tsvector('zhcfg', ip.desc) @@ tsq.query
+            AND ${contentTag}::text IS NULL
+
+          ORDER BY rank DESC
+          OFFSET ${offset}
+          LIMIT ${limit};
+        `
+      : await prisma.$queryRaw<SearchRow[]>`
+          SELECT
+            COALESCE(vt.id, 'tag:' || v.aweme_id) AS id,
+            v.aweme_id AS "awemeId",
+            'video'::text AS type,
+            1.0::double precision AS rank,
+            COALESCE(
+              NULLIF(array_to_string(vt.transcript[1:2], ' '), ''),
              v.desc,
-              tsq.query,
-              'StartSel=<mark>,StopSel=</mark>,MaxFragments=2,MinWords=2,MaxWords=20'
-            )
-        END AS snippet
-      FROM "Video" v
-      LEFT JOIN "VideoTranscript" vt ON vt."videoId" = v.aweme_id,
-           tsq
-      WHERE
-        (vt.transcript_zh_tsv @@ tsq.query OR to_tsvector('zhcfg', v.desc) @@ tsq.query)
-      
-      UNION ALL
-      
-      -- 图文搜索结果
-      SELECT
-        'img_desc:' || ip.aweme_id AS id,
-        ip.aweme_id AS "awemeId",
-        'image'::text AS type,
-        ts_rank(to_tsvector('zhcfg', ip.desc), tsq.query) AS rank,
-        ts_headline(
-          'zhcfg',
-          ip.desc,
-          tsq.query,
-          'StartSel=<mark>,StopSel=</mark>,MaxFragments=2,MinWords=2,MaxWords=20'
-        ) AS snippet
-      FROM "ImagePost" ip, tsq
-      WHERE to_tsvector('zhcfg', ip.desc) @@ tsq.query
-      
-      ORDER BY rank DESC
-      OFFSET ${offset}
-      LIMIT ${limit};
-    `;
+              '标签匹配'
+            ) AS snippet,
+            v.manual_content_tags AS "contentTags"
+          FROM "Video" v
+          LEFT JOIN "VideoTranscript" vt ON vt."videoId" = v.aweme_id
+          WHERE v.manual_content_tags @> ARRAY[${contentTag}]::text[]
+          ORDER BY v.created_at DESC
+          OFFSET ${offset}
+          LIMIT ${limit};
+        `;

    // 查询总数（参数化，避免注入）
-    const totalRows = await prisma.$queryRaw<
-      {
-        count: number;
-      }[]
-    >`
-      WITH tsq AS (
-        SELECT websearch_to_tsquery('zhcfg', ${q}) AS query
-      )
-      SELECT (
-        (SELECT COUNT(*)::int FROM "Video" v
-         LEFT JOIN "VideoTranscript" vt ON vt."videoId" = v.aweme_id
-         WHERE (vt.transcript_zh_tsv @@ tsq.query OR to_tsvector('zhcfg', v.desc) @@ tsq.query))
-        +
-        (SELECT COUNT(*)::int FROM "ImagePost" ip
-         WHERE to_tsvector('zhcfg', ip.desc) @@ tsq.query)
-      ) AS count
-      FROM tsq;
-    `;
+    const totalRows = q
+      ? await prisma.$queryRaw<{ count: number }[]>`
+          WITH tsq AS (
+            SELECT websearch_to_tsquery('zhcfg', ${q}) AS query
+          )
+          SELECT (
+            (SELECT COUNT(*)::int FROM "Video" v
+             LEFT JOIN "VideoTranscript" vt ON vt."videoId" = v.aweme_id
+             WHERE (vt.transcript_zh_tsv @@ tsq.query OR to_tsvector('zhcfg', v.desc) @@ tsq.query)
+               AND CASE
+                 WHEN ${contentTag}::text IS NULL THEN TRUE
+                 ELSE v.manual_content_tags @> ARRAY[${contentTag}]::text[]
+               END)
+            +
+            (SELECT COUNT(*)::int FROM "ImagePost" ip
+             WHERE to_tsvector('zhcfg', ip.desc) @@ tsq.query
+               AND ${contentTag}::text IS NULL)
+          ) AS count
+          FROM tsq;
+        `
+      : await prisma.$queryRaw<{ count: number }[]>`
+          SELECT COUNT(*)::int AS count
+          FROM "Video" v
+          LEFT JOIN "VideoTranscript" vt ON vt."videoId" = v.aweme_id
+          WHERE v.manual_content_tags @> ARRAY[${contentTag}]::text[];
+        `;

    // 分离视频和图文ID
    const videoIds = rows
@@ -176,6 +228,7 @@ export async function GET(req: Request) {
    return json({
      results: rows.map((r) => ({
        ...r,
+        snippet: sanitizeSnippetHtml(r.snippet),
        video:
          r.type === "video"
            ? videos.find((v) => v.aweme_id === r.awemeId)
				`@@ -0,0 +1 @@`
				`export { GET } from "../../stt/videos/route";`