fix: prevent STT and refinement model settings from being randomly overwritten

[yixin0829]

Summary

• Fixed bug where STT model and Refinement Model settings were randomly overwritten in the config GUI
• Root cause: provider-specific caching mechanism was not properly syncing with GUI state
• Added sync helper methods to keep cache in sync regardless of whether values are typed or selected
• Used list membership to correctly determine which provider's cache to update when switching providers

Changes

• Added _sync_stt_model_to_cache() and _sync_refinement_model_to_cache() helper methods
• Simplified _on_stt_model_changed() and _on_refinement_model_changed() to use sync helpers
• Fixed _update_stt_model_options() and _update_refinement_model_options() to:
  • Save to PREVIOUS provider's cache using list membership (since provider var is already updated)
  • Restore from cache without list validation (preserves custom/non-listed models)
• Updated set_values() to call sync methods after setting model values

Test plan

• All 242 existing tests pass
• Verified test_stt_model_preserved_when_switching_providers passes
• Manual test: Type custom model name, click provider dropdown - model should be preserved
• Manual test: Load config with non-listed model - model should be preserved
• Manual test: Switch providers back and forth - models should be remembered per provider

🤖 Generated with Claude Code

[codecov-commenter]

⚠️ Please install the to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

❌ Patch coverage is 68.06723% with 76 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
src/push_to_talk.py	59.21%	31 Missing ⚠️
src/text_refiner_gemini.py	35.71%	27 Missing ⚠️
src/audio_recorder.py	59.25%	11 Missing ⚠️
src/hotkey_service.py	75.00%	5 Missing ⚠️
src/text_refiner_cerebras.py	77.77%	2 Missing ⚠️

📢 Thoughts on this report? Let us know!