AI-TestGen

AI-TestGen is a CLI tool that leverages AI to automatically generate unit, service, integration, and end-to-end tests for service-oriented codebases (Node.js, TypeScript, etc.).

Purpose

• Save developer time and effort
• Improve test coverage and quality
• Reduce bugs and regressions
• Standardize testing practices across teams

Features

• Code analysis and test suggestion
• Test code generation (unit, service, integration, E2E)
• Customizable test templates
• Review and edit interface
• Integration with code editors and CI/CD
• Flexible output directory and filename options
• Overwrite protection and auto-directory creation
• Dry run mode: Preview what would be generated without writing files
• Optional LLM/AI-powered test generation: Use GPT (OpenAI) to generate tests if desired, with support for custom endpoints/gateways

Quickstart

1. Install dependencies:

   npm install

2. Generate tests for a TypeScript service:

   node src/cli/index.js generate --all src/userService.ts --output tests/generated

3. Preview what would be generated (dry run):

   node src/cli/index.js generate --all src/userService.ts --output tests/generated --dry-run

4. Specify custom output directories and filenames:

   node src/cli/index.js generate --all src/userService.ts \
     --unit-output tests/unit --unit-filename userService.unit.test.ts \
     --service-output tests/service --service-filename userService.service.test.ts \
     --integration-output tests/integration --integration-filename userService.integration.test.ts \
     --e2e-output tests/e2e --e2e-filename userService.e2e.test.ts

5. Overwrite existing files:

   node src/cli/index.js generate --all src/userService.ts --output tests/generated --force

LLM/AI-Powered Test Generation (Optional)

If you want to use OpenAI GPT to generate tests:

1. Set your API key:

   export OPENAI_API_KEY=sk-...

2. (Optional) Set a custom OpenAI-compatible base URL (for gateways, Azure, etc.):

   export OPENAI_BASE_URL=https://your-gateway.example.com/v1
   # or use the CLI flag:
   # --openai-base-url https://your-gateway.example.com/v1

3. Run with the --ai flag:

   node src/cli/index.js generate --all src/userService.ts --output tests/generated --ai
   # or with a custom base URL:
   node src/cli/index.js generate --all src/userService.ts --output tests/generated --ai --openai-base-url https://your-gateway.example.com/v1

• If no API key is set, or --ai is not used, the tool falls back to static/template-based generation.
• Each developer must use their own API key. Never share your key.

CLI Usage

Usage: ai-testgen generate [type] <target> [options]

Arguments:
  [type]         Test type: unit | service | integration | e2e (optional if --all is used)
  <target>       Target file or directory

Options:
  -o, --output <dir>           Default output directory (default: "examples")
  --all                        Generate all test types for the target file (omit [type] argument)
  --unit-output <dir>          Unit test output directory
  --unit-filename <name>       Unit test filename
  --service-output <dir>       Service test output directory
  --service-filename <name>    Service test filename
  --integration-output <dir>   Integration test output directory
  --integration-filename <name> Integration test filename
  --e2e-output <dir>           E2E test output directory
  --e2e-filename <name>        E2E test filename
  --force                      Overwrite existing test files
  --dry-run                    Show what would be generated, but do not write files
  --ai                         Use LLM-powered test generation if OPENAI_API_KEY is set
  --openai-base-url <url>      Custom OpenAI-compatible base URL (for gateways, Azure, etc.)

Real-World Examples

Preview all test types with custom locations (dry run, no [type] needed):

node src/cli/index.js generate --all src/userService.ts \
  --unit-output tests/unit --unit-filename userService.unit.test.ts \
  --service-output tests/service --service-filename userService.service.test.ts \
  --integration-output tests/integration --integration-filename userService.integration.test.ts \
  --e2e-output tests/e2e --e2e-filename userService.e2e.test.ts --dry-run

Generate all test types with OpenAI GPT (if API key is set):

export OPENAI_API_KEY=sk-...
node src/cli/index.js generate --all src/userService.ts --output tests/generated --ai

Generate all test types with a custom OpenAI-compatible endpoint:

export OPENAI_API_KEY=sk-...
export OPENAI_BASE_URL=https://your-gateway.example.com/v1
node src/cli/index.js generate --all src/userService.ts --output tests/generated --ai
# or
node src/cli/index.js generate --all src/userService.ts --output tests/generated --ai --openai-base-url https://your-gateway.example.com/v1

Generate only a unit test in a custom location:

node src/cli/index.js generate unit src/userService.ts --unit-output tests/unit --unit-filename userService.unit.test.ts

Overwrite existing files:

node src/cli/index.js generate --all src/userService.ts --output tests/generated --force

Preview (dry run) before generating:

node src/cli/index.js generate --all src/userService.ts --output tests/generated --dry-run

License

MIT