- Add Gemma4Service with local inference capabilities using flutter_gemma package
- Implement model download and management functionality for Gemma models
- Create SetupModelScreen allowing users to choose and download Gemma models
- Add ModelDownloadTile widget for displaying and managing available models
- Update README to document Gemma 4 integration and local inference features
- Modify RequestDetailScreen to use Gemma4Service instead of GeminiService
- Update AppBarWidget to include model setup button
- Add flutter_gemma dependency to pubspec.yaml
- Remove old Gemini API integration and related code
- Update .gitignore with proper formatting

This integration provides local AI inference using Gemma models with privacy-focused on-device processing, replacing the previous cloud-based Gemini API implementation.

- Added flutter_gemma plugin to macOS project.
- Updated GeneratedPluginRegistrant.swift to register FlutterGemmaPlugin.
- Modified Podfile to include setup script for LiteRT-LM Desktop.
- Updated Podfile.lock to include flutter_gemma dependency.
- Adjusted project.pbxproj to include FlutterGeneratedPluginSwiftPackage.
- Added shell script phase for LiteRT-LM Desktop setup in Xcode project.
- Updated Runner.xcscheme to run Flutter framework preparation script.
- Modified entitlements files to disable library validation for security.
- Updated widget test to reference the main application entry point.