A feather weight mock data generator API

Extremely simple python API to generate mock data for your app using chatgpt openai model.

Features

• Mock Data generation for especific fields and context
• Data caching and saving to file

Installation

Use poetry to create a virtual environment and install the dependencies.

Move into /mockary/mockary folder and run:

poetry install

Quick Start

Mockary is very simple to use. It only takes around 5 minutes to get it up and running (including the time of reading this guide).

Add OpenAI API Key

Modify the .env.example with the OpenAI Api Key and rename it to .env.

Start the server

Inside the /mockary/mockary directory run:

poetry run uvicorn main:app

Define a Mock

There are some examples in the /mockary/mockary/config.ini file. To add a new one use this base template:

[MockName]
fields=field1,field2,field3
message="Message to give the AI model context about the data"
cache=true

Call the API

curl -X 'GET' \
  'http://localhost:8000/[MockName]?samples=[number of samples]'

The result will be a JSON with the generated data.

{
    "data": [
        {
            ... generated data ...
        },
    ]
}

Settings

Global Settings

Option	Description
cache	Can be used in CONFIG or individual mock. If true, the generated data will cached in memory.
max_tokens	Maximum tokens to generate. Default is the maximum - 4096.
model	Especific model to use. Default is gpt-3.5-turbo. To use other models use the name as described in the openai documentation
temperature	Temperature to use in the model. Default is 1.

Mock Settings

Option	Description
save	Save the generated data to a file.
save_path	Path to save the generated data. The default is ./mock_name
fields	Fields to generate data.
message	Message to give the AI model context about the data.