B2 Refactoring & Testing

💡 One-liner: Use AI to identify code smells, refactor safely, and automatically generate test cases.

📝 Course Notes

Key takeaways from this lesson:

---

What You'll Be Able to Do

• Let AI identify "code smells" in your code
• Safely refactor code
• Automatically generate unit tests
• Analyze edge cases and supplement tests

---

Your Current Struggles

• Code gets messier over time, but you're afraid to change it
• Want to refactor but lack test coverage
• Too few test cases, can't improve coverage

---

When to Use This Technique

• When you need to: Improve code quality and test coverage
• And don't want to: Manually write numerous test cases

---

🎒 Before You Start

Make sure you've completed the following:

• [ ] Completed B1 Daily Development
• [ ] Have a code file that needs refactoring or testing

---

Core Concepts

Safe Refactoring in Three Steps

1. Tests first → 2. Small refactoring steps → 3. Tests pass

Code Smells Overview

Code Smell	Symptoms (Experience Signals)	Refactoring Direction
Long Function	Can't see the whole thing at a glance, mixing multiple concerns	Extract function / Split responsibilities
Too Many Parameters	Callers struggle to remember what each parameter means	Introduce Parameter Object / Split function
Duplicated Code	Similar logic appears in multiple places	Extract common function
Deep Nesting	Reading code requires repeatedly tracing back through conditions	Early return / Extract conditional logic
Unclear Naming	Can't tell purpose from the name	Rename

---

Follow Along

Step 1: Identify Code Smells

Why Diagnose first, then treat.

Switch to Plan Agent:

@src/utils/data.ts Please analyze the code quality of this file:
1. List discovered "code smells"
2. Severity of each issue (High/Medium/Low)
3. Recommended refactoring approaches
4. Priority suggestions for refactoring

Step 2: Generate Tests First

Why Tests enable safe refactoring.

Switch to Build Agent:

@src/utils/data.ts Generate unit tests for this file:
1. Use Vitest framework
2. Cover all exported functions
3. Include normal cases and edge cases
4. Save as src/utils/data.test.ts

Step 3: Safe Refactoring

Why Small steps, verify each step.

@src/utils/data.ts Please refactor the parseUserData function:
- Issue: Function is too long (50 lines), single responsibility violation
- Requirements: Split into 3 smaller functions
- Keep the public interface unchanged
- Run tests after refactoring to confirm

Step 4: Supplement Edge Case Tests

Why Edge cases are often bug-prone areas.

@src/utils/data.ts @src/utils/data.test.ts

Analyze the code's edge cases and supplement the following tests:
1. Empty inputs (null, undefined, empty arrays)
2. Extreme values (very large numbers, very long strings)
3. Type errors (passing wrong types)
4. Concurrency scenarios (if async operations exist)

Step 5: Run Tests to Verify

Why Confirm refactoring didn't break functionality.

In OpenCode TUI, messages starting with ! execute shell commands and bring output into the conversation: https://opencode.ai/docs/tui

!npm test src/utils/data.test.ts

You should see: All tests pass

---

📋 Magic Prompts

✅ Code Review

Expected outcome: Comprehensive analysis of code quality issues with improvement suggestions

## Role
You are a Tech Lead responsible for code quality. Your review style is strict but constructive.

## Task
Perform a comprehensive review of submitted code, identify issues, and provide improvement suggestions.

## Input
### Required
- Programming Language: [Language]
- Code: @[File Path] or [Paste Code]

### Optional
- Review Focus: [What aspects to focus on?]
- Project Context: [What is the code's context?]

## Review Dimensions
1. **Code Quality**: Naming conventions, code structure, readability, duplicated code
2. **Potential Issues**: Bug risks, edge cases, error handling, performance issues
3. **Security Concerns**: Input validation, sensitive information, injection risks
4. **Best Practices**: Design patterns, SOLID principles, test coverage

## Output Format
### Review Summary
- **Code Rating**: A(Excellent)/B(Good)/C(Average)/D(Needs Refactoring)
- **One-liner Comment**
- **Must Fix Before Merge**: Yes/No

### Issues List
- 🔴 **Critical Issues**: `[File:Line]` Issue + Risk + Suggestion
- 🟡 **Suggested Improvements**: Issues recommended to fix
- 🟢 **Strengths**: Things worth keeping

## Constraints
- ✅ Issues should be specific to line numbers
- ✅ Every issue should have a fix suggestion
- ✅ Acknowledge what's done well
- ❌ Avoid replacing objective standards with subjective preferences
- ❌ Avoid criticism without solutions

🧪 Test Generation

Expected outcome: Generate comprehensive unit tests covering edge cases

## Role
You are a QA Engineer skilled at designing comprehensive test cases covering normal flows and edge cases.

## Task
Generate unit tests for the function or module provided by the user.

## Input
### Required
- Programming Language: [Language]
- Code: @[File Path] or [Paste Code]

### Optional
- Test Framework: [Jest/Vitest/Pytest/JUnit?] (Default: auto-detect)
- Test Style: [BDD describe/it style?]

## Output Format
Use AAA Pattern (Arrange-Act-Assert), covering:
1. **Happy Path** (Normal flow)
2. **Edge Cases** (Null values, extremes, boundary points)
3. **Exception Cases** (Invalid input, thrown exceptions)
4. **Async Operations** (If Promise/async exists)

Each test case should include:
- Clear naming (describing test intent)
- Necessary comments (explaining test scenario)

## Constraints
- ✅ Test case names should clearly express test intent
- ✅ Cover main edge cases
- ✅ Test code should be directly runnable
- ❌ Avoid testing implementation details
- ❌ Avoid dependencies between test cases

🔧 Refactoring Suggestions

Expected outcome: Provide safe refactoring plan to reduce change risk

## Role
You are a Refactoring Expert skilled at identifying code smells and providing safe refactoring plans.

## Task
Analyze code issues and provide safe refactoring steps and suggestions.

## Input
### Required
- Code: @[File Path] or [Paste Code]

### Optional
- Known Issues: [What problems have you noticed?]
- Refactoring Goal: [What effect do you want to achieve?]

## Output Format
1. **Current Problem Analysis**
   - Discovered code smells (Long Function/Too Many Parameters/Deep Nesting/Duplicated Code/Unclear Naming)
   - Severity of each issue (High/Medium/Low)

2. **Recommended Refactoring Steps** (in order)
   - What to do in each step
   - Risk assessment for each step
   - How to verify step success

3. **Tests to Add**
   - Tests that should exist before refactoring
   - Tests to add after refactoring

4. **Expected Results After Refactoring**
   - Code quality improvements
   - Maintainability improvements

## Constraints
- ✅ Small refactoring steps, each verifiable
- ✅ Keep public interface unchanged
- ✅ Tests before refactoring
- ❌ Avoid changing too much at once
- ❌ Avoid changing business logic

---

Checklist ✅

All items must pass to continue

• [ ] Identified code smells using AI
• [ ] Generated unit tests
• [ ] Completed safe refactoring
• [ ] All tests pass

---

Common Pitfalls

Symptom	Cause	Solution
Functionality broken after refactoring	No tests written first	Generate tests before refactoring
Low test coverage	Only testing normal cases	Add edge case and exception tests
Inaccurate smell detection	Code snippet too small	Give AI complete file context

---

Lesson Summary

You learned:

1. Using AI to identify code smells
2. Safe refactoring flow with tests first
3. Automatically generating unit tests
4. Analyzing and supplementing edge case tests

---

Next Lesson Preview

Next lesson we'll learn about documentation automation and Git collaboration techniques.

---

📚 More Complete Templates: Prompt Template Library