ChatGPT Structured Output Demo

This project demonstrates how to leverage the Strategy Pattern along with Spring Boot class mapping to easily implement structured output from ChatGPT. The approach allows you to generate and parse JSON responses from ChatGPT in a type-safe manner using custom DTOs and message strategies.

Features

• Structured Output via ChatGPT:
  Generate responses in a predefined JSON structure by instructing ChatGPT with system messages. The response is automatically mapped to your defined DTOs (e.g., HadithOfTheDayDTO, IslamicEventDetailDTO, QuranOfTheDayDTO) using the strategy pattern.

• Strategy Pattern for Flexibility:
  Each ChatGPT request is encapsulated in a strategy class (e.g., HadithOfTheDayStrategy, IslamicEvent, QuranOfTheDay). This makes it easy to extend the system with additional structured outputs without changing the core logic.

• Spring Boot Integration:
  The application is built with Spring Boot, making it easy to manage dependency injection, configuration, and RESTful endpoints.

• Official OpenAI Java Library:
  This project uses the official OpenAI Java library to communicate with the ChatGPT API.

Getting Started

Prerequisites

• Java 11 or higher
• Maven or Gradle
• Spring Boot 2.5+

Setup

1. Clone the Repository:

   git clone https://github.com/sharifrahim/your-repo.git
   cd your-repo

2. Generate your own ChatGPT API Key:

   Sign up or log in to OpenAI Platform and create your API key.

3. Configure the API Key:

   In your application's application.properties or application.yml, add the following property:

   openai.api.key=YOUR_OPENAI_API_KEY

4. Build and Run the Application:

   Using Maven:

   mvn spring-boot:run

   Or using Gradle:

   ./gradlew bootRun

How It Works

ChatGPT Structured Output

The core of this implementation is the ChatGptServiceImpl class. It:

• Generates an expected JSON format using an empty instance of the response DTO.
• Constructs the API request with a system message instructing ChatGPT to output a JSON in the specified format.
• Calls the ChatGPT API using the official OpenAI Java client.
• Processes and deserializes the JSON response into the respective DTO.

This approach allows for:

• Strongly Typed Responses:
  By using Spring Boot’s class mapping and DTOs, responses from ChatGPT are automatically converted into Java objects.
• Easy Extensibility:
  By leveraging the MessageStrategy interface, adding a new structured output is as simple as creating a new strategy implementation.

Example Strategy Implementations

• Hadith of the Day:
  HadithOfTheDayStrategy generates a user message like "Give me random islamic hadith" and expects a HadithOfTheDayDTO as the response.

• Islamic Event Details:
  IslamicEvent takes an event name as input and constructs a message requesting details about the event's origin and sunnah, formatted as a paragraph with reference links. The response is mapped to an IslamicEventDetailDTO.

• Quran of the Day:
  QuranOfTheDay requests a random Quran verse and maps the output to a QuranOfTheDayDTO.

Technologies Used

• Spring Boot: For rapid application development and dependency management.
• Official OpenAI Java Library: To interact with ChatGPT via the OpenAI API.
  OpenAI Java Library
• Lombok: To reduce boilerplate code for DTOs and services.
• Jackson: For JSON processing and deserialization.

Contributing

Contributions are welcome! Please open issues or submit pull requests for any improvements or bug fixes.

---

Author: sharifrahim

Generate your own ChatGPT API key at OpenAI Platform.

Enjoy building structured outputs with ease!