Mastering AI Driven Development - Lesson 4

⚠️ RECORDED: February 2025 Models and features evolve quickly, but these core techniques remain my daily workflow. Join my community for latest updates: https://www.skool.com/prompt-to-prod-9369 GitHub Repository for this lesson: https://github.com/jayminwest/tf-ai-development-lesson04 Lesson 4 - Enhancing Code Consistency with Aider 📖 Project Overview Welcome to Lesson 4 of the TandemFlow AI Development Course. This lesson focuses on ensuring consistency when using Aider across multiple files in a project. The complexity of projects increases from this point forward, making consistency, structure, and developer guidelines essential. You'll learn how to: - Establish a Developer Guide to maintain AI-assisted coding standards. - Use linting and testing to improve project reliability. - Manage multi-file projects efficiently using Aider. 📌 Lesson Breakdown 1️⃣ Introduction to Code Consistency (00:00:00 - 00:00:20:20) - Lessons will start building on each other, reinforcing previous concepts. - AI-generated code must be consistent across multiple files. - Aider needs guidance to match your coding standards and expectations. 2️⃣ Understanding the Developer Guide (00:01:03 - 00:02:22:17) - Similar to Scratchpad.md, but more detailed. - Acts as an instruction manual for AI when writing code. - A template Developer Guide is available in the GitHub repo. - Key sections include: - Project Overview & Goals - Coding Style & Preferences - Project Architecture - Implementation Patterns & Naming Conventions - Error Handling & Security 3️⃣ Defining a Clear Project Architecture (00:02:22 - 00:05:09:18) - Large-scale projects require a structured file system. - Example project refactor: - wikipedia_fetcher.py → Handles API requests. - article_analysis.py → Processes text data. - article_visualization.py → Generates visual outputs. - Aider can map the project using: /map 4️⃣ Implementing Linting & Testing (00:05:09:18 - 00:12:02:13) - Linting ensures clean, error-free code before execution. - Aider automatically detects and fixes linting issues. - Unit testing verifies that different parts of the project function correctly. - Example test case for wikipedia_fetcher.py: /add test_wikipedia_fetcher.py - Uses pytest for structured test execution. - Automated tests run after each significant change. 5️⃣ Creating and Configuring Aider's Developer Guide (00:12:02:13 - 00:18:18:16) - Example Base Prompt: # Base Prompt: A Python tool for data analysis on Wikipedia articles. - Coding Preferences: - Language: Python - Package Manager: pip - Testing Framework: pytest - Documentation Style: Google-style docstrings - Implementation Patterns: - Private variables should start with _ - Function names should use snake_case - Use type annotations to improve readability. - Error Handling Strategy: - Use try-except blocks for external API calls. - Ensure consistent logging messages. 6️⃣ Using the Aider Config File for Automation (00:18:18:16 - 00:31:12:22) - Configuration allows for custom model selection, linting, testing, and workflow enhancements. - Key settings in aider.config.yaml: model: "claude-3.5-sonnet" auto_commit: true auto_test: true test_command: "pytest" read_only: - "developer_guide.md" - Auto-committing ensures incremental saves to Git. - Auto-testing triggers tests after every change. - Read-only files prevent unwanted modifications to key documentation. 7️⃣ Executing Multi-File Refactoring with Aider (00:31:12:22 - 00:33:56:20) - Command to break down wikipedia_fetcher.py: /refactor wikipedia_fetcher.py to article_analysis.py, article_visualization.py - File breakdown: - wikipedia_fetcher.py → Fetches Wikipedia data. - article_analysis.py → Processes and tokenizes text. - article_visualization.py → Generates word clouds and bar charts. - Code is restructured automatically using Aider with minimal manual intervention. 8️⃣ Running & Validating the New Project Structure (00:33:56:20 - 00:34:33:07) - Run the updated script: python wikipedia_fetcher.py "Seattle" - Outputs checked: - Markdown file with Seattle's Wikipedia article. - Word cloud & bar chart visualizations. - Command-line statistics (word count, unique words, sentence length). 🚀 Next Steps - Refine developer guide for larger projects. - Experiment with multi-file modifications using Aider. - Explore advanced linting and testing frameworks. - Prepare for Lesson 5, which covers Aider's Architect Mode for large-scale refactoring. See you in Lesson 05! 🚀