ai_rspec_writer is an AI-powered tool for generating RSpec test cases in Ruby on Rails applications. It supports ChatGPT and Gemini AI to create model, controller, and service tests automatically, improving test coverage and development efficiency.
- 📝 Automatically generates RSpec tests for models, controllers, and services.
- 📌 Extracts database schema for validation and association tests.
- 🛠 Supports both ChatGPT and Gemini AI.
- 🔍 Customizable test instructions using extra comments.
- 🏆 Ensures high code coverage by generating edge-case scenarios.
- 🔑 Secure API authentication using OpenAI & Google Gemini keys.
group :development, :test do
gem 'ai_rspec_writer'
gem 'factory_bot_rails' # Generates dummy data
gem 'rails-controller-testing' # Required for controller specs
gem 'faker' # Fake data generation
gem 'shoulda-matchers', '~> 5.0' # One-liner matchers for RSpec
gem 'simplecov', require: false # Code coverage analysis
endbundle installCreate a .env file in your project root and add:
CHATGPT_API_KEY=your-openai-api-key
GEMINI_API_KEY=your-google-gemini-api-key
DEFAULT_AI=chatgpt
CHATGPT_MODEL=your-model
GEMINI_MODEL=your-model
Or export them in your terminal:
export CHATGPT_API_KEY="your-openai-api-key"
export GEMINI_API_KEY="your-google-gemini-api-key"
export DEFAULT_AI="chatgpt"
export CHATGPT_MODEL="your-model"
export GEMINI_MODEL="your-model"Default Setting:
GEMINI_MODEL=o3-mini
CHATGPT_MODEL=gemini-2.0-flash
DEFAULT_AI=chatgpt
bundle exec ai_rspec_writer -f <file_path> -t <table_names> -e "<extra_comment>" -a <ai_model>bundle exec ai_rspec_writer -f app/controllers/obento/obento_zaikos_controller.rb,app/models/store_notification.rb,app/models/common_setting.rb -t store_notifications,common_settings -e "use devise admin and ensure 100% test coverage" -a geminibundle exec ai_rspec_writer -f app/controllers/obento/obento_zaikos_controller.rb,app/models/store_notification.rb,app/models/common_setting.rb -t store_notifications,common_settings -e "use devise admin and ensure 100% test coverage" -a chatgptIf no AI model (-a) is passed, ai_rspec_writer will default to DEFAULT_AI from .env or use ChatGPT if not explicitly set.
bundle exec ai_rspec_writer -f app/controllers/obento/obento_zaikos_controller.rb,app/models/store_notification.rb,app/models/common_setting.rb -t store_notifications,common_settings -e "use devise admin and ensure 100% test coverage"| Option | Description | Example |
|---|---|---|
-f, --files |
Files to generate tests for (comma-separated paths). | -f app/models/user.rb,app/controllers/users_controller.rb |
-t, --table_name |
Database tables to extract schema info for tests. | -t users,orders |
-e, --ec |
Extra comments to customize test behavior. | -e "use Devise authentication" |
-a, --ai |
AI model to use (chatgpt or gemini). |
-a gemini |
After running:
bundle exec ai_rspec_writer -f app/models/store_notification.rb -t store_notifications -e "use FactoryBot"It generates:
# spec/models/store_notification_spec.rb
require 'rails_helper'
RSpec.describe StoreNotification, type: :model do
let(:store_notification) { create(:store_notification) }
describe "Validations" do
it { should validate_presence_of(:title) }
it { should validate_presence_of(:message) }
end
describe "Associations" do
it { should belong_to(:store) }
end
describe "Methods" do
it "sends notification successfully" do
expect(store_notification.send_notification).to be_truthy
end
end
endSolution: Set the API key in .env:
CHATGPT_API_KEY=your-api-key
Or export it:
export CHATGPT_API_KEY="your-api-key"Solution: Set the Gemini API key in .env:
GEMINI_API_KEY=your-api-key
Solution: Ensure you pass the -f option with valid file paths:
bundle exec ai_rspec_writer -f app/models/user.rbSolution: Try providing additional extra comments:
bundle exec ai_rspec_writer -f app/models/user.rb -e "Generate edge case tests"This project is licensed under the MIT License.
- @aniruddhami 🎉
- Open for Contributions! Feel free to fork & improve.
- 📌 RSpec test generation improvements
- 📌 Support for Minitest
- 📌 AI model selection enhancements
- 📌 More detailed error handling