语音转文本接口，bgm路径修复

app/api/stt/index.ts

@@ -2,6 +2,8 @@ import fs from "fs";
 import OpenAI from "openai";
 import prompt from "./prompt.md";
 import { prisma } from "@/lib/prisma";
+import { downloadFile } from "@/lib/minio";
+import { extractAudio } from "../media";
 
 const client = new OpenAI({
   apiKey: process.env.OPENAI_API_KEY,
@@ -13,6 +15,7 @@ async function transcriptAudio(audio: Buffer | string) {
     audio = fs.readFileSync(audio);
   }
   const base64Audio = Buffer.from(audio).toString("base64");
+
   const response = await client.chat.completions.create({
     model: "gemini-2.5-flash-lite",
     messages: [
@@ -27,7 +30,7 @@ async function transcriptAudio(audio: Buffer | string) {
             type: "input_audio",
             input_audio: {
               data: base64Audio,
-              format: "wav",
+              format: "mp3",
             },
           },
         ],
@@ -36,14 +39,24 @@ async function transcriptAudio(audio: Buffer | string) {
   });
 
   console.log(response.choices[0].message.content);
+  return response.choices[0].message.content;
 }
 
-async function transcriptAweme(awemeId: string) {
+export async function transcriptAweme(awemeId: string) {
   const aweme = await prisma.video.findUnique({
     where: { aweme_id: awemeId },
   });
   if (!aweme) {
     throw new Error("Aweme not found or aweme is not a video post");
   }
-  
+  const vPath = aweme.video_url
+  const buffer = await downloadFile(vPath);
+
+  const audioDat = await extractAudio(buffer, { format: "mp3", bitrateKbps: 96 });
+
+  if (!audioDat || !audioDat.buffer) {
+    throw new Error("Failed to extract audio from video");
+  }
+
+  return await transcriptAudio(audioDat.buffer);
 }

app/api/stt/route.ts

@@ -0,0 +1,19 @@
+import { NextRequest, NextResponse } from 'next/server';
+import { prisma } from '@/lib/prisma';
+import type { FeedItem, FeedResponse } from '@/app/types/feed';
+import { getFileUrl } from '@/lib/minio';
+import { transcriptAweme } from '.';
+
+// Contract
+// Inputs: search params { before?: ISOString, limit?: number }
+// Output: { items: FeedItem[], nextCursor: ISOString | null }
+
+export async function GET(req: NextRequest) {
+  const { searchParams } = new URL(req.url)
+  const awemeId = searchParams.get('awemeId');
+  if (!awemeId) {
+    return NextResponse.json({ error: "Missing awemeId parameter" }, { status: 400 });
+  }
+  const script = await transcriptAweme(awemeId);
+  return NextResponse.json(script);
+}