I want to create an AI model that generates text. Specifically, BDD Gherkin cucumber scenarios and step definitions based of input of a user story.

User story with BDD Gherkin cucumber example

For example.

User story (Input): I want to add products to my shopping basket on an e-commerce website to make a purchase.

Output: Automatically create the test case scenarios and step definitions
Test case scenarios:

• Scenario 1: Validate that user can add one item to cart
• Scenario 2: Validate that user can remove one item from cart

Test case scenario 1:

• Given the user launch and login e-commerce application with and
• Then the user navigates to items page.
• And the user selects and click on a .
• And the user clicks the “Add to cart” button.
• Then the user should navigate to the Shopping Cart page.
• And user should validate in Shopping Cart Page has been successfully added.

Test case scenario 2:

• Given the user launch and login e-commerce application with and
• Then the user should navigate to the Shopping Cart page.
• And the user finds in shopping cart and click “Remove from cart” button.
• Then user should validate that in shopping has been successfully removed.

I have created a sample dataset that contains user stories mapped to scenarios and step definitions.

Dataset

From my current understand, the logic is that: I want to train a model based of the dataset of existing user stories and scenarios. After the model has been trained, I want to input a user story, and the model should come up with a suitable scenario with step definitions.

I am new to machine learning and has only done some form of supervised learning, regression. From some research, I would need to use some NLP techniques to process the dataset. From then on, I am pretty lost. I’ve seen some people talking about using ChatGPT to train of the dataset or something.

What would be a good way to do this project.