LLM4TDD: TypeScript Example

This project demonstrates Test-Driven Development (TDD) using Large Language Models (LLMs). The workflow consists of three main steps:

1. Generate Test Skeleton: Create a basic structure for tests
2. Generate Complete Tests: Flesh out the test skeleton with actual test cases
3. Generate Function Implementation: Create a function that passes all the tests

Workflow

graph TD
    A[Test skeleton created] --> B[Complete tests generated]
    B --> C[Function implementation generated]
    C --> D[Tests passing]

Loading

Usage

Run the complete TDD workflow:

npm run tdd

This will:

1. Generate a test skeleton for a random function
2. Generate a complete test file from the skeleton
3. Generate a function implementation that passes all the tests

Individual Steps

You can also run each step individually:

# Generate a test skeleton
npx tsx generateTestSkeleton.ts

# Generate a complete test from a skeleton
npx tsx generateTestFileFromSkeleton.ts

# Generate a function implementation from a test file
npx tsx generateFunctionFromSpec.ts

Utilities

The project includes several utility functions in the utils directory:

• generateTestSkeleton.ts: Creates a test skeleton using OpenAI
• generateTestFromSkeleton.ts: Generates complete tests from a skeleton
• generateFunctionFromSpec.ts: Creates a function implementation that passes tests
• chat.ts: Utility for interacting with OpenAI
• readFileContent.ts: Utility for reading file content
• writeFileContent.ts: Utility for writing file content
• runTests.ts: Utility for running tests

Requirements

• Node.js
• npm
• OpenAI API key (set as OPENAI_API_KEY environment variable)