Skip to content

Getting Started

This Getting Started guide targets software developers who would like to generate persona-based questions for an application.

There are several steps.

  1. Get your Python virtual environment set up and add the necessary Python libraries. We use Conda in this example.
  2. Get your OpenAI API Key and put it in your .zshrc file so it is set up when you start a shell
  3. Test your API Key
  4. For a new application, generate the personas
  5. For each Persona, generate the questions

Conda Setup

We recommend you set up a named virtual environment with Conda or a Python VENV.

Here are the steps with Conda

1
2
3
conda create -n "copilot-metadata" python=3
conda activate copilot-metadata
pip install requests openai mkdocs mkdocs-material

You will need an OpenAI API Key

Store the key in your ~/.zshrc file 

1
2
OPENAI_API_KEY=sk...
export OPENAI_API_KEY

Test Your Environment Variable

1
echo $OPENAI_API_KEY

Run the following program Python program in the src directory that tests your OpenAI API Key works.

1
python src/01-openai-test.py

Generate Personas from the Application Description

Edit the src/05-get-personas-from-app-desc.py Python program to add a description. Use the following Python program:

1
python 05-get-personas-from-app-desc.py

That will generate a list of personas. You will use this list for the next section. Make sure you review the list and remove any personas that are not appropriate such as security and performance roles.

Generate Questions for Each Persona

Take the description and list of personas and copy them into the appropriate areas of the 06-get-questions-for-personas.py program.

1
python 06-get-questions-for-personas.py

TODO

  1. Add a feature to take the application description as an input parameter in a file or the section within a markdown documents after the ## Application Description section header.
  2. Take the output of the first Python program and use it as the input to the second.