feat: centralize audio format validation for TTS and music commands

[omghante]

Description

This PR centralizes audio format validation across the MiniMax CLI for speech synthesize, music generate, and music cover commands. Previously, invalid --format values were silently passed to the API, resulting in confusing server-side errors. Now the CLI validates early and fails fast with a clear message.

Changes Included:

• Created a centralized src/utils/audio-formats.ts with type-safe format constants and validation
• Added early CLI-side --format validation across speech and music commands (quick-fail UX)
• Updated help text to dynamically list supported formats (mp3, wav, flac, pcm) instead of hardcoded strings
• Added unit test for invalid format rejection

Note: Only API-supported formats (mp3, wav, flac, pcm) are included. When backend opus/ogg support lands, it's a one-line addition to AUDIO_FORMATS.

Fixes: #111

@stayif Thanks for raising this the CLI-side validation infrastructure is now in place. Once the backend adds opus/ogg support, enabling it here is literally adding two words to an array. Let me know if you have any feedback!

@RyanLee-Dev Updated per your review removed opus/ogg, kept only the 4 API-validated formats. The centralized validation and dynamic help text are ready to go. Thanks for the thorough testing!

[raylanlin]

Thanks for this PR, @omghante! The code quality and structure are solid — here are some notes from a detailed review:

✅ What's great:

• Centralized audio-formats.ts utility with type-safe as const — clean pattern
• Dynamic help text (AUDIO_FORMATS_DISPLAY) is much better than hardcoded strings in each command
• Early validation before API call — great UX for quick-fail
• Well-written tests covering opus, ogg, and invalid format rejection

⚠️ Issue with opus/ogg support:
The MiniMax TTS API (v1/t2a_v2) does not actually support opus or ogg as output formats. I tested all 6 formats against the live API:

Format	API response	Supported?
mp3	0 (success)	✅
wav	0 (success)	✅
flac	0 (success)	✅
pcm	0 (success)	✅
opus	2013 (invalid params)	❌
ogg	2013 (invalid params)	❌

With this PR as-is, if a user passes --format opus, the CLI will accept it (passes validateAudioFormat) but then the API returns 2013 - invalid params: audio_setting format, giving a confusing error message.

Suggested fix:
Remove opus and ogg from AUDIO_FORMATS in src/utils/audio-formats.ts, keeping only the 4 API-supported formats:

export const AUDIO_FORMATS = ['mp3', 'wav', 'flac', 'pcm'] as const;

After this change, the rest of this PR (centralized validation, dynamic help text, test infrastructure) is all meaningful and ready to merge.

Happy to follow up if you want to adjust the supported formats — the refactoring work here is definitely worth landing.

[stayif]

效率真高啊，厉害！

[RyanLee-Dev]

The Opus format will be supported this week.

[omghante]

Thanks for testing against the live API! Dropped opus/ogg and force-pushed easy one-liner to re-add when backend support lands. Let me know if anything else needs.