feat: Add ModelsLab TTS provider

[adhikjoshi]

Summary

Adds ModelsLab as a new TTS backend for VideoLingo, following the exact same per-file pattern as the existing SiliconFlow, OpenAI, Azure, and Edge TTS integrations.

Changes

File	Change
core/tts_backend/modelslab_tts.py	New provider module
core/tts_backend/tts_main.py	Import + elif TTS_METHOD == 'modelslab_tts' dispatch
config.yaml	modelslab_tts: section with api_key and voice
core/st_utils/sidebar_setting.py	Streamlit API key input + voice picker dropdown

Features

• 10 voices: Bella, Antoni, Elli, Josh, Arnold, Adam, Sam, Rachel, Domi, Gigi
• Async polling: handles status: "processing" responses (polls every 5s, up to 5 min)
• Character limit: 2,500 chars enforced automatically
• Retry logic: 3 attempts via @except_handler (same pattern as other providers)
• WAV output: compatible with VideoLingo's audio pipeline

Setup

1. Get your API key from modelslab.com
2. Set modelslab_tts.api_key in config.yaml (or via Streamlit sidebar)
3. Select modelslab_tts as the TTS Method and choose a voice

modelslab_tts:
  api_key: 'YOUR_MODELSLAB_API_KEY'
  voice: 'Bella'   # Bella, Antoni, Elli, Josh, Arnold, Adam, Sam, Rachel, Domi, Gigi

About ModelsLab

ModelsLab offers affordable AI inference including TTS, image generation, and video generation. API docs: https://docs.modelslab.com