Skip to content

GEMINI.md

Latest commit

 

History

History
44 lines (32 loc) · 4.47 KB

GEMINI.md

File metadata and controls

44 lines (32 loc) · 4.47 KB
Raw

AI Development Guidelines for "Bloc de notas" in GitHub Codespaces

These guidelines define the operational principles and capabilities of the AI agent interacting with the Bloc de notas codebase. The environment is shifted from Firebase to a web-based GitHub Codespaces cloud container optimized for Android and Flutter development.

1. Environment & Context Awareness

The AI operates within a GitHub Codespaces container, utilizing terminal-based tools, headless Android tools, and web previews.

  • Project Structure & Architecture: Standard Flutter layout applies. The architecture must explicitly separate UI from business logic using Feature-first Structure or Layered Architecture (Presentation, Domain, Data).
  • DRY Principle: The AI must rigorously respect the DRY (Don't Repeat Yourself) philosophy. Before writing any utility, widget, or helper, the AI will inspect existing files to reuse logic and avoid code duplication.
  • Localization & Internationalization: Every user-facing string must support multi-language translations. Hardcoded strings in the UI are strictly prohibited. The AI will utilize Flutter's localization files (e.g., ARB files or localization context) to add translations for any new feature.
  • Environment Limitations: Since development occurs via web-browser containers, local physical device access isn't available. The AI will leverage web previews or headless testing configurations inside the Codespace environment.

2. Material Design 3 UI/UX Specifications

The application interface must strictly adhere to Material Design 3 (Material You) guidelines.

Iconography Rule

  • Outlined Icons Only: The AI must always use the outlined variant of Material Icons (Icons.xxxx_outlined) whenever available. Filled, rounded, or sharp variations should never be generated unless an outlined equivalent does not exist.

Dynamic & Harmonic Colors

  • The application must utilize ColorScheme.fromSeed to ensure cohesive, modern, and accessible color palettes across light and dark modes.
  • Visual effects like multi-layered drop shadows, soft deep elevation for lifted cards, and elegant colored glows on interactive states should be incorporated to create a premium, tactile feel.

3. Code Modification & State Management

  • Core Code Base: The main entry point is lib/main.dart. The application relies on Provider and ChangeNotifier for app-wide state management, dependency injection, and theme toggling (Light/Dark/System).
  • Local UI Reactivity: For localized, single-value state changes, the AI will use ValueNotifier and ValueListenableBuilder to maximize performance and minimize unnecessary widget rebuilds.
  • Dependency Control: External packages must be stable and added via standard Flutter tooling:
  • Automated Clean Code: The AI will automatically trigger formatting and basic fixes to keep the code clean and compile-ready:

4. Local Architecture & Storage (No Firebase)

Since Firebase is no longer used, all persistence, cloud-sync references, and AI architectures are shifted to modern standalone local methods or user-controlled client-side strategies.

  • Local Storage: High-performance local storage (like Hive, Isar, or Sqflite) is preferred for caching notes, metadata, and user categories.
  • Client-Side Cloud Integration: Any external backup systems (e.g., Google Drive sync) must be implemented strictly Client-Side via OAuth2 and official REST APIs, maintaining privacy and decoupled architecture.
  • Local Logging: Application logs must rely exclusively on the structured dart:developer library.

5. Iterative Blueprint Management

Every cycle of change follows a strict documentation and verification flow to ensure consistency across separate IDE sessions:

  1. Blueprint Synchronization: Before changing anything, the AI will create or update a blueprint.md file in the root directory. This file is the single source of truth and contains:
    • Project Overview: Current state, styling parameters, and established features.
    • Change History: Log of changes from initial versions up to version latest and upcoming milestones.
    • Current Action Plan: Step-by-step roadmap of the current prompt.
  2. Verification Flow: After changes are applied, the AI simulates or triggers flutter analyze and flutter test to ensure that no existing features are broken, and formatting conforms exactly to local specifications.