feat: persist composer model and reasoning effort selections

[Dimillian]

Motivation

• Remember the last model and reasoning effort used in the composer so new threads default to the last selections and users don't lose their configured reasoning level when switching chats.

Description

• Added lastComposerModelId and lastComposerReasoningEffort to the frontend AppSettings (src/types.ts) and to the backend settings struct (src-tauri/src/types.rs) with default handling and serde names.
• Initialized the default settings in useAppSettings (src/features/settings/hooks/useAppSettings.ts) to include the new fields.
• Updated useModels (src/features/models/hooks/useModels.ts) to accept preferredModelId and preferredEffort and to seed/retain selections using existing, preferred, or default values when refreshing the model list.
• Wired the saved preferences into the app root (src/App.tsx) by passing appSettings.lastComposerModelId and appSettings.lastComposerReasoningEffort into useModels and adding an effect to persist changes via queueSaveSettings when the composer selection changes.

Testing

• Ran npm run lint, which produced an informational warning about the TypeScript version used by @typescript-eslint but completed without blocking changes.
• Ran npm run typecheck (tsc --noEmit) and it completed successfully.

---

Codex Task

[Dimillian]

@codex review

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 70d8d14e26

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".