From 7193dd8bb710659461cec2b20b3d11b81ab59314 Mon Sep 17 00:00:00 2001
From: "Satsuki Akiba (aider)" <anosatsuk124@gmail.com>
Date: Wed, 4 Jun 2025 11:26:55 +0900
Subject: [PATCH] feat: use VOICEVOX API for TTS

---
 services/tts.ts | 70 +++++++++++++++++++++++++++++++++++--------------
 1 file changed, 51 insertions(+), 19 deletions(-)

diff --git a/services/tts.ts b/services/tts.ts
index 1fdd0ee..d7d53c2 100644
--- a/services/tts.ts
+++ b/services/tts.ts
@@ -1,37 +1,69 @@
 import fs from "fs";
 import path from "path";
-import { PollyClient, SynthesizeSpeechCommand } from "@aws-sdk/client-polly";
+import fetch from "node-fetch";
 
-const polly = new PollyClient({ region: "ap-northeast-1" });
+// VOICEVOX APIの設定
+const VOICEVOX_HOST = "http://localhost:50021";
+
+interface VoiceStyle {
+  speakerId: number;
+  styleId: number;
+}
+
+// 仮の声設定（例: サイドM = 3）
+const defaultVoiceStyle: VoiceStyle = {
+  speakerId: 3,
+  styleId: 2,
+};
 
 export async function generateTTS(
   itemId: string,
   scriptText: string,
 ): Promise<string> {
-  const params = {
-    OutputFormat: "mp3",
-    Text: scriptText,
-    VoiceId: "Mizuki",
-    LanguageCode: "ja-JP",
-  };
-  const command = new SynthesizeSpeechCommand(params);
-  const response = await polly.send(command);
+  // 音声合成クエリの生成
+  const queryResponse = await fetch(`${VOICEVOX_HOST}/audio_query`, {
+    method: "POST",
+    headers: {
+      "Content-Type": "application/json",
+    },
+    body: JSON.stringify({
+      text: scriptText,
+      speaker: defaultVoiceStyle.speakerId,
+    }),
+  });
 
-  if (!response.AudioStream) {
-    throw new Error("TTSのAudioStreamが空です");
+  if (!queryResponse.ok) {
+    throw new Error("VOICEVOX 音声合成クエリ生成に失敗しました");
   }
 
+  const audioQuery = await queryResponse.json();
+
+  // 音声合成
+  const audioResponse = await fetch(`${VOICEVOX_HOST}/synthesis`, {
+    method: "POST",
+    headers: {
+      "Content-Type": "application/json",
+    },
+    body: JSON.stringify({
+      audio_query: audioQuery,
+      speaker: defaultVoiceStyle.speakerId,
+    }),
+  });
+
+  if (!audioResponse.ok) {
+    throw new Error("VOICEVOX 音声合成に失敗しました");
+  }
+
+  const audioBuffer = await audioResponse.buffer();
+
+  // 出力ディレクトリの準備
   const outputDir = path.join(__dirname, "../public/podcast_audio");
   if (!fs.existsSync(outputDir)) {
     fs.mkdirSync(outputDir, { recursive: true });
   }
-  const filePath = path.resolve(outputDir, `${itemId}.mp3`);
 
-  const chunks: Uint8Array[] = [];
-  for await (const chunk of response.AudioStream as any) {
-    chunks.push(chunk);
-  }
-  const buffer = Buffer.concat(chunks);
-  fs.writeFileSync(filePath, buffer);
+  const filePath = path.resolve(outputDir, `${itemId}.mp3`);
+  fs.writeFileSync(filePath, audioBuffer);
+
   return filePath;
 }