feat: 支持 MiniMax 音色克隆和 2.8 语音模型

2026-06-17 02:13:21 +08:00
parent 02ba5d3cdd
commit c4341335d7
20 changed files with 1645 additions and 10 deletions
@@ -0,0 +1,135 @@
+#!/usr/bin/env node
+
+const baseURL = (process.env.GATEWAY_BASE_URL || 'http://localhost:8080').replace(/\/+$/, '');
+const apiKey = process.env.GATEWAY_API_KEY || process.env.EASYAI_GATEWAY_API_KEY;
+const cloneModel = process.env.GATEWAY_VOICE_CLONE_MODEL || 'MiniMax-Voice-Clone';
+const speechModel = process.env.GATEWAY_TTS_MODEL || 'speech-2.6-turbo';
+const voiceId =
+  process.env.VOICE_CLONE_ID || `voice_clone_${Date.now().toString(36)}`;
+const audioURL =
+  process.env.VOICE_CLONE_AUDIO_URL ||
+  `${baseURL}/static/simulation/audio.wav`;
+const marker = `voice-clone-e2e-${Date.now().toString(36)}`;
+
+if (!apiKey) {
+  throw new Error('Set GATEWAY_API_KEY or EASYAI_GATEWAY_API_KEY');
+}
+
+function assert(condition, message) {
+  if (!condition) throw new Error(message);
+}
+
+async function request(path, init = {}) {
+  const res = await fetch(`${baseURL}${path}`, {
+    ...init,
+    headers: {
+      Authorization: `Bearer ${apiKey}`,
+      'Content-Type': 'application/json',
+      ...(init.headers || {}),
+    },
+  });
+  const text = await res.text();
+  const body = text ? JSON.parse(text) : {};
+  if (!res.ok) {
+    throw new Error(`${init.method || 'GET'} ${path} failed ${res.status}: ${text}`);
+  }
+  return body;
+}
+
+async function postAsyncTask(path, body) {
+  const accepted = await request(path, {
+    method: 'POST',
+    headers: { 'X-Async': 'true' },
+    body: JSON.stringify(body),
+  });
+  const taskId = accepted.taskId || accepted.task?.id;
+  assert(taskId, `Expected async task id from ${path}`);
+  return pollTask(taskId);
+}
+
+async function pollTask(taskId, timeoutMs = 120000) {
+  const started = Date.now();
+  while (Date.now() - started < timeoutMs) {
+    const task = await request(`/api/v1/tasks/${taskId}`);
+    if (task.status === 'succeeded') return task;
+    if (task.status === 'failed') {
+      throw new Error(`Task ${taskId} failed: ${task.errorMessage || task.error || JSON.stringify(task)}`);
+    }
+    await new Promise((resolve) => setTimeout(resolve, 1000));
+  }
+  throw new Error(`Timed out waiting for task ${taskId}`);
+}
+
+const cloneTask = await postAsyncTask('/v1/voice_clone', {
+  model: cloneModel,
+  voice_id: voiceId,
+  audio_url: audioURL,
+  text: 'hello voice clone preview',
+  preview_model: process.env.VOICE_CLONE_PREVIEW_MODEL || 'speech-2.8-hd',
+  runMode: 'simulation',
+  simulation: true,
+  integrationTestMarker: `${marker}-clone`,
+});
+
+const cloneResult = cloneTask.result || {};
+const clonedVoice = cloneResult.cloned_voice || cloneResult.clonedVoice;
+assert(cloneResult.status === 'success', `Unexpected clone result: ${JSON.stringify(cloneResult)}`);
+assert((cloneResult.voice_id || clonedVoice?.voiceId || clonedVoice?.voice_id) === voiceId, 'Clone voice_id mismatch');
+assert(clonedVoice?.platformId || clonedVoice?.platform_id, 'Clone result missing platform binding');
+
+const listResult = await request('/v1/voice_clone/voices');
+const voices = listResult.items || listResult.data || [];
+const listedVoice = voices.find((item) => item.voiceId === voiceId || item.voice_id === voiceId);
+assert(listedVoice, 'Cloned voice is missing from voice list');
+assert(
+  (listedVoice.platformId || listedVoice.platform_id) ===
+    (clonedVoice.platformId || clonedVoice.platform_id),
+  'Listed voice platform binding mismatch',
+);
+
+const speechTask = await postAsyncTask('/v1/speech/generations', {
+  model: speechModel,
+  text: 'hello from cloned voice',
+  cloned_voice_id: clonedVoice.id,
+  runMode: 'simulation',
+  simulation: true,
+  integrationTestMarker: `${marker}-speech`,
+});
+const speechResult = speechTask.result || {};
+assert(speechResult.status === 'success', `Unexpected speech result: ${JSON.stringify(speechResult)}`);
+const speechAttemptPlatformId = speechTask.attempts?.[0]?.platformId;
+assert(speechAttemptPlatformId, 'Speech task is missing attempt platformId');
+assert(
+  speechAttemptPlatformId === (clonedVoice.platformId || clonedVoice.platform_id),
+  `Speech used ${speechAttemptPlatformId}, expected cloned voice platform ${clonedVoice.platformId || clonedVoice.platform_id}`,
+);
+
+if (process.env.GATEWAY_CROSS_PLATFORM_TTS_MODEL) {
+  try {
+    await postAsyncTask('/v1/speech/generations', {
+      model: process.env.GATEWAY_CROSS_PLATFORM_TTS_MODEL,
+      text: 'this should not cross platform',
+      cloned_voice_id: clonedVoice.id,
+      runMode: 'simulation',
+      simulation: true,
+    });
+    throw new Error('Cross-platform TTS request unexpectedly succeeded');
+  } catch (error) {
+    if (String(error?.message || '').includes('unexpectedly succeeded')) throw error;
+  }
+}
+
+console.log(
+  JSON.stringify(
+    {
+      ok: true,
+      voiceId,
+      clonedVoiceId: clonedVoice.id,
+      platformId: clonedVoice.platformId || clonedVoice.platform_id,
+      cloneTaskId: cloneTask.id,
+      speechTaskId: speechTask.id,
+    },
+    null,
+    2,
+  ),
+);