feat: add MiniMax TTS integration

[octo-patch]

Summary

This PR adds a NovaSR/minimax_tts.py module that integrates MiniMax TTS with NovaSR's audio super-resolution pipeline — fulfilling the use case already described in the README:

"Enhancing models: NovaSR can enhance TTS model quality considerably with nearly 0 computational cost."

What's added

• NovaSR/minimax_tts.py — Two public functions:

  • generate_speech(text, ...) — streams the MiniMax TTS API (speech-2.8-hd / speech-2.8-turbo) and returns MP3 bytes; reads MINIMAX_API_KEY from the environment
  • generate_and_enhance(text, upsampler, ...) — one-call pipeline: MiniMax TTS → FastSR → 48 kHz enhanced audio

• tests/test_minimax_tts.py — 13 unit tests (fully mocked / offline, all passing)

• README.md — MiniMax TTS integration section with copy-paste examples

Usage

import os
from NovaSR import FastSR
from NovaSR.minimax_tts import generate_and_enhance
import torchaudio

os.environ["MINIMAX_API_KEY"] = "your_minimax_api_key"

upsampler = FastSR()
enhanced = generate_and_enhance("Hello, world!", upsampler=upsampler)
torchaudio.save("enhanced.wav", enhanced.cpu(), 48000)

API references

• Chat (OpenAI Compatible): https://platform.minimax.io/docs/api-reference/text-openai-api
• TTS: https://platform.minimax.io/docs/api-reference/speech-t2a-http