Cross Browser AI Authoring (CBAA) Demo

This repository demonstrates how to use BrowserStack's Cross Browser AI Authoring (CBAA) with Python and Selenium. It executes Gherkin-based feature scenarios on BrowserStack's infrastructure, leveraging AI to interpret and execute natural language steps.

Prerequisites

• Python 3.x: Ensure Python 3 is installed on your system.
• BrowserStack Account: You need a BrowserStack username and access key. Sign up for a free trial if you don't have one, and connect with your Account Manager to enable Cross Browser AI Authoring.

Setup

1. Clone the repository:

   git clone <repository-url>
   cd cbaa-demo

2. Create and activate a virtual environment (optional but recommended):

   python3 -m venv venv
   source venv/bin/activate  # On Windows use: venv\Scripts\activate

3. Install dependencies:

   This project requires selenium, gherkin-official, and python-dotenv.

   pip install -r requirements.txt
   pip install python-dotenv

4. Configure Environment Variables:

   Create a .env file in the root directory of the project and add your BrowserStack credentials:

   BROWSERSTACK_USERNAME=your_username
   BROWSERSTACK_ACCESS_KEY=your_access_key

Project Structure

• cbaa.py: The main entry point script. It initializes the Selenium driver with BrowserStack options, parses feature files, and executes scenarios using a thread pool.
• features/: Contains Gherkin .feature files defining the test scenarios.
  • purchase.feature: A sample e-commerce purchase scenario.
• utils/: Helper utilities.
  • parse_feature.py: Parses Gherkin feature files and "hydrates" them with examples (replaces placeholders with actual data).
• browserstack.yml: Configuration file for BrowserStack SDK (reference).
• requirements.txt: List of Python dependencies.

Running the Tests

To run the AI-authored tests, execute the cbaa.py script:

python cbaa.py

What happens during execution:

1. The script reads features/purchase.feature.
2. It parses the scenarios and replaces placeholders (e.g., <Category>, <Email>) with the values from the Examples table.
3. For each scenario iteration, it spawns a thread.
4. Each thread starts a Selenium session on BrowserStack.
5. Steps are executed using browserstack_executor with the ai action, allowing the AI to interpret natural language commands like "Click on the 'Men' category".
6. Results are reported back to BrowserStack.

Notes

• Ensure your BrowserStack plan supports the number of parallel threads you intend to run.
• The aiAuthoring: "true" capability is critical for enabling the AI features.