ArchitectPro

ArchitectPro is a professional-grade desktop environment designed for Software Architects and Senior Developers. It serves as a "Context Engineering Cockpit", bridging the gap between complex local Python codebases and Large Language Models (LLMs).

By integrating high-level decision-making strategies (Hexagonal Architecture, Black Box Systems) with the repository-packing power of Repomix, ArchitectPro allows you to generate highly specific, secure, and architecturally sound contexts for AI analysis.

---

🚀 Features

1. Python Context Engineering

• Smart Packing: Uses repomix to consolidate local Python projects into AI-friendly XML or Markdown formats.
• Abstraction Control: Choose between Application (Detailed), Solution (Mid-Level), or Enterprise (Abstract) levels. This automatically tunes the repomix filters to ignore boilerplate, tests, or migrations based on your needs.
• Selective Injection: Browse and check specific .py files to append their raw content directly to your prompt for targeted code reviews.

2. Architectural Strategies (Prompt Library)

Built-in, Jinja2-templated prompt engineering strategies designed to force LLMs into structural thinking:

• Hexagonal Architecture: Guides the AI to refactor code into strict Domain, Ports, and Adapters layers.
• Steenberg Black Box: Implements modularity principles focusing on strict I/O boundaries and replaceability.
• Strategic Reasoning: One-click access to Graph-of-Thoughts (GoT), Chain-of-Verification (CoVe), and Architectural Debt Mapping.

3. Professional Suite

• Modern Interface: High-DPI responsive GUI built with CustomTkinter.
• System Logging: Real-time console output for tracking Repomix operations and system status.
• Token Estimation: Built-in character-to-token calculator to help stay within LLM context limits (Claude, GPT-4, Gemini).

---

🛠 Prerequisites

1. Python 3.10+
2. Node.js & NPM (Required for the packing engine)
3. Repomix:

   npm install -g repomix

---

📥 Installation

1. Clone the repository:

   git clone https://github.com/bbsoft0/ArchitectPro.git
   cd ArchitectPro

2. Install Python dependencies:

   pip install customtkinter jinja2 python-dotenv

3. API Configuration (Optional): Create a .env file in the root directory if using advanced AI features:

   GEMINI_API_KEY=your_key_here

---

🖥 Usage Workflow

Step 1: Setup & Source

1. Target Directory: Select your Python project root.
2. Context Focus: Define your goal (e.g., "Enterprise" for high-level refactoring).
3. Smart Pack: Click "Py Smart Pack (XML)". This triggers Repomix to map your entire project structure.

Step 2: Context & Files

• Select specific files from the generated list that require immediate attention.
• Toggle "Lazy Mode" if you only want the AI to return the specific lines that need changing (Find/Replace format).
• Observe the Token Counter to ensure your project fits in the LLM window.

Step 3: Strategy & Run

1. Navigate the tabs (Black Box, Hexagonal, Reasoning).
2. Click a strategy (e.g., "Strategic Black-Box Planning").
3. ArchitectPro copies the combined Instructions + Project Map + Source Code to your clipboard.
4. Paste into your preferred LLM (Claude, ChatGPT, etc.) for an architect-level response.

---

🧠 Philosophy: The "Five Lines" Rule

ArchitectPro follows the philosophy that it is faster to write five lines of clean code today than to edit one line of spaghetti later.

The tool enforces this by:

1. Defining Boundaries: Forcing the AI to see modules as Black Boxes.
2. Isolating Side Effects: Encouraging Hexagonal adapters for database and API calls.
3. Structural Context: Providing the AI with the whole repository map so it understands cross-file dependencies.

---

🔧 Troubleshooting

Repomix Command Fails

• Verify npm install -g repomix was successful.
• On Windows, ensure the npm global bin folder is in your System PATH.

Files not appearing in list

• Click "Browse" again or refresh the path. The app specifically filters for .py files in professional mode.

Log Files

• See ArchitectPro.log for a full history of subprocess commands and internal errors.