Merge pull request #237 from Qixingchen/main

feat: add emotion, text-normalization and latex-read flags to speech synthesize

Author: RyanLee-Dev

src/commands/speech/synthesize.ts

@@ -26,6 +26,9 @@ export default defineCommand({
     { flag: '--speed <n>',               description: 'Speech speed multiplier', type: 'number' },
     { flag: '--volume <n>',              description: 'Volume level', type: 'number' },
     { flag: '--pitch <n>',               description: 'Pitch adjustment', type: 'number' },
+    { flag: '--emotion <emotion>',       description: 'Emotion: happy, sad, angry, fearful, disgusted, surprised, calm, fluent, whisper (passed through as-is; see API Reference below for new values)' },
+    { flag: '--text-normalization',      description: 'Enable Chinese/English text normalization' },
+    { flag: '--latex-read',              description: 'Read LaTeX formulas ($$...$$, Chinese only)' },
     { flag: '--format <fmt>',            description: `Audio format: ${formatList(T2A_FORMATS)} (default: mp3)` },
     { flag: '--sample-rate <hz>',        description: 'Sample rate (default: 32000)', type: 'number' },
     { flag: '--bitrate <bps>',           description: 'Bitrate (default: 128000)', type: 'number' },
@@ -77,6 +80,9 @@ export default defineCommand({
         speed: (flags.speed as number) ?? undefined,
         vol: (flags.volume as number) ?? undefined,
         pitch: (flags.pitch as number) ?? undefined,
+        emotion: (flags.emotion as string) || undefined,
+        text_normalization: flags.textNormalization === true ? true : undefined,
+        latex_read: flags.latexRead === true ? true : undefined,
       },
       audio_setting: {
         format: (flags.format as string) || 'mp3',

src/types/api.ts

@@ -97,6 +97,11 @@ export interface SpeechRequest {
     speed?: number;
     vol?: number;
     pitch?: number;
+    // Emotion hint (e.g. 'happy', 'sad', 'calm'); passed through as-is — the
+    // API validates the value, so no local enum to keep in sync.
+    emotion?: string;
+    text_normalization?: boolean;
+    latex_read?: boolean;
   };
   audio_setting?: {
     format?: string;

test/commands/speech/synthesize.test.ts

@@ -397,4 +397,70 @@ describe('speech synthesize format validation', () => {
       console.log = originalLog;
     }
   });
-});
+});
+
+describe('speech synthesize voice_setting flags', () => {
+  const config = {
+    apiKey: 'test-key',
+    region: 'global' as const,
+    baseUrl: 'https://api.mmx.io',
+    output: 'json' as const,
+    timeout: 10,
+    verbose: false,
+    quiet: false,
+    noColor: true,
+    yes: false,
+    dryRun: true,
+    nonInteractive: true,
+    async: false,
+  };
+
+  const flags = {
+    text: 'Hello',
+    quiet: false,
+    verbose: false,
+    noColor: true,
+    yes: false,
+    dryRun: true,
+    help: false,
+    nonInteractive: true,
+    async: false,
+  };
+
+  async function dryRunRequest(extraFlags: Record<string, unknown>) {
+    const originalLog = console.log;
+    let output = '';
+    console.log = (msg: string) => { output += msg; };
+    try {
+      await synthesizeCommand.execute(config, { ...flags, ...extraFlags });
+      return JSON.parse(output).request;
+    } finally {
+      console.log = originalLog;
+    }
+  }
+
+  it('--emotion sets voice_setting.emotion', async () => {
+    const request = await dryRunRequest({ emotion: 'happy' });
+    expect(request.voice_setting.emotion).toBe('happy');
+  });
+
+  it('--emotion passes through arbitrary values without local validation', async () => {
+    const request = await dryRunRequest({ emotion: 'bored' });
+    expect(request.voice_setting.emotion).toBe('bored');
+  });
+
+  it('omits voice_setting.emotion when --emotion is not provided', async () => {
+    const request = await dryRunRequest({});
+    expect(request.voice_setting.emotion).toBeUndefined();
+  });
+
+  it('--text-normalization sets voice_setting.text_normalization', async () => {
+    const request = await dryRunRequest({ textNormalization: true });
+    expect(request.voice_setting.text_normalization).toBe(true);
+  });
+
+  it('--latex-read sets voice_setting.latex_read', async () => {
+    const request = await dryRunRequest({ latexRead: true });
+    expect(request.voice_setting.latex_read).toBe(true);
+  });
+});