format-gpt2

Sanitize ChatGPT output and format it the way you want!

FormatGPT is a robust library designed to sanitize output provided by ChatGPT. It provides a consistent and well defined method for retrieving the data from ChatGPT by acting as a middle man between your code and ChatGPT API. It supports various formats such as:

• text
• json
• json-list
• csv
• html
• markdown
• arrays
• tables
• xml
• yaml

Additionally each output format can be easily customized to suit your needs.

Table of Contents

1. Installation
2. Usage
3. API Reference
4. Building
5. Tests
6. License

Installation

To install the library using npm run the command:

npm install format-gpt

To install the library using yarn run the command:

yarn add format-gpt

Usage

Here's a quick example to get you started:

import formatGPT from "format-gpt";

formatGpt.format({
	format: "json",
	options: {
		attributes: [
			{ name: "name", type: "string" },
			{ name: "description", type: "string", maxLength: 200 },
			{ name: "released", type: "date" },
			{ name: "price", type: "decimal" }
		]
	},
	language: "en"
});

API Reference

The default export is formatChatGptPrompt.

formatChatGptPrompt

formatChatGptPrompt({
  prompt: string,
  format: Format,
  options?: FormatOptions,
  language?: string
}): string;

props

• prompt (string) - content of the prompt for chat-gpt
• format (Format) - format in which to return output from chat-gpt
• options (FormatOptions)
  • attributes: (IAttribute[])- array with attribute definitions
  • columnSeparator (string)- column separator (for CSV format, default: ",")
  • rowSeparator (string)- row separator (for CSV format. default: \n)
• language (string) - language code (example: en, pl)

formatChatGptMessages

formatChatGptMessages({
  messages: ChatCompletionRequestMessage[],
  format: Format,
  options?: FormatOptions,
  language?: string
}): ChatCompletionRequestMessage[];

props

• messages (ChatCompletionRequestMessage[]) - messages sent to chat-gpt api
• format (Format) - format in which to return output from chat-gpt
• options (FormatOptions)
  • attributes: (IAttribute[])- array with attribute definitions
  • columnSeparator (string)- column separator (for CSV format, default: ",")
  • rowSeparator (string)- row separator (for CSV format. default: \n)
• language (string) - language code (example: en, pl)

formatChatGptRequest

formatChatGptRequest({
  request: CreateChatCompletionRequest,
  format: Format,
  options?: FormatOptions,
  language?: string
}): CreateChatCompletionRequest;

props

• request (CreateChatCompletionRequest) - request config for chat-gpt api
• format (Format) - format in which to return output from chat-gpt
• options (FormatOptions)
  • attributes: (IAttribute[])- array with attribute definitions
  • columnSeparator (string)- column separator (for CSV format, default: ",")
  • rowSeparator (string)- row separator (for CSV format. default: \n)
• language (string) - language code (example: en, pl)
  

Format

type Format = "array" | "csv" | "html" | "json" | "json-list" | "markdown" | "table" | "text" | "xml" | "yaml";

IAttribute

interface IAttribute {
    name: string;
    type: "string" | "number" | "boolean" | "date" | "integer" | "decimal" | Array<IAttribute>;
    minLength?: number;
    maxLength?: number;
    custom?: string;
}

For more information on ChatCompletionRequestMessage and CreateChatCompletionRequest please refer OpenAI documentation.

Building

To build format-gpt follow these steps:

1. Clone this repository
2. Install dependencies using npm install or yarn install command
3. If using yarn with Visual Studio Code also run yarn dlx @yarnpkg/sdks vscode
4. Run npm build or yarn build command

Tests

To run the test suite, first install the dependencies, then execute tests.

Using npm:

npm install
npm test

Using yarn:

yarn install
yarn test

License

This project is licensed under the MIT License.