This script extends mongosh with a helper function:

await db.chatgpt(keyword, count = 5, arrays = false)

It uses the OpenAI API to generate realistic JSON documents based on a keyword and inserts them into a MongoDB collection.

• Generate synthetic data using ChatGPT (via OpenAI API)
• Creates a new MongoDB collection (based on keyword)
• Supports nested array fields for richer document structure
• Integrates natively into mongosh

• MongoDB shell (mongosh) v1.10 or higher (must support fetch)
• OpenAI API key
• Internet access (to reach OpenAI)

1. Clone this repository:

   git clone https://github.com/olaekdahl/mongodb-helpers.git

2. In chatgpt-helper.js, replace <your-openai-api-key>).

Start mongosh and run:

load('/path/to/chatgpt-helper.js')
db.chatgpt("techstartups", 10, true)

You’ll get a new collection techstartups populated with AI-generated documents that include arrays.

To use this helper every time you start mongosh:

1. Open or create your mongosh RC file:

   nano ~/.mongoshrc.js

2. Add the following line:

   load('/path/to/chatgpt-helper.js')

3. Save and restart mongosh. You can now use:

   db.chatgpt("techstartups", 10, true)

await db.chatgpt(keyword, count = 5, arrays = false)

• keyword (string): Topic or type of documents to generate.
• count (number, optional): Number of documents to generate (default: 5).
• arrays (boolean, optional): If true, includes nested array fields in documents (default: false).

Avoid committing your OpenAI API key to version control. You can instead read it from a config file and inject into .mongoshrc.js dynamically if needed.