This Flask API generates Behavior-Driven Development (BDD) scenarios in Gherkin syntax for cloud infrastructure security. The API leverages Gemini 1.5 Flash for natural language generation and a pre-built vector index for document querying.

• Generates Gherkin-based BDD scenarios for Terraform configurations.
• Accepts cloud provider and service as inputs.
• Uses a pre-persisted vector index and Gemini 1.5 Flash LLM for responses.

• Python 3.8+
• A GOOGLE_API_KEY environment variable with access to Gemini 1.5 Flash.
• Pre-built index stored in the ./storage directory.

Ensure you have the following installed:

• Python
• Pip
• Flask
• Other Python dependencies listed in the requirements.txt.

1. Clone the Repository:

   git clone https://github.com/your-repository/terraform-bdd-api.git
   cd terraform-bdd-api

2. Install Dependencies:

   pip install -r requirements.txt

3. Set the Environment Variable: Create a .env file in the root directory:

   echo "GOOGLE_API_KEY=your_google_api_key" > .env

   Replace your_google_api_key with your valid API key.

4. Ensure the Persisted Index: Place the pre-built index files in the ./storage directory.

This API relies on a vector index created using llama-index. Make sure your ./storage directory contains the persisted index files (docstore.json, index_store.json, etc.). Use StorageContext.from_defaults(persist_dir="./storage") during indexing.

Start the Flask app:

python app.py

The API will run locally at http://127.0.0.1:5000.

Generate BDD scenarios for a specified cloud provider and service.

• Request:

  {
    "provider": "GCP",
    "service": "Storage Bucket"
  }

• Response: Returns a string containing the Gherkin-based BDD scenario.

  Example:

  Scenario: Mitigating public access vulnerability in Storage Bucket
      Given a configuration for Storage Bucket on GCP
      When permissions are set to allow public access
      Then restrict bucket access to authorized users only to prevent unauthorized access

curl -X POST http://127.0.0.1:5000/generate_scenario      -H "Content-Type: application/json"      -d '{"provider": "AWS", "service": "S3 Bucket"}'

├── app.py                 # Flask API implementation
├── storage/               # Persisted index files (pre-built)
├── requirements.txt       # Python dependencies
├── README.md              # Project documentation
└── .env                   # Environment variables

• Flask
• llama-index
• huggingface/transformers
• dotenv

• GOOGLE_API_KEY: Required for Gemini 1.5 Flash API access.

1. Missing API Key: Ensure the .env file contains the correct GOOGLE_API_KEY.

2. Index Not Found: Verify the ./storage directory contains the persisted index files.

3. Flask Errors:

   • Ensure all dependencies are installed.
   • Check for typos in the request payload.

This project is licensed under the MIT License.